Merge remote-tracking branch 'origin/dev' into halo2

Resolves #154

.clang-format

@@ -0,0 +1,38 @@
+Language: Cpp
+AlignAfterOpenBracket: AlwaysBreak
+AlignConsecutiveMacros: true
+AlignTrailingComments: true
+AllowAllParametersOfDeclarationOnNextLine: true
+AllowShortBlocksOnASingleLine: true
+AllowShortCaseLabelsOnASingleLine: false
+AllowShortFunctionsOnASingleLine: All
+AllowShortIfStatementsOnASingleLine: true
+AlwaysBreakTemplateDeclarations: true
+BinPackArguments: true
+BinPackParameters: false
+BreakBeforeBraces: Custom
+BraceWrapping:
+  AfterClass: true
+  AfterFunction: true
+BreakBeforeBinaryOperators: false
+BreakBeforeTernaryOperators: true
+ColumnLimit: 120
+ContinuationIndentWidth: 2
+Cpp11BracedListStyle: true
+DisableFormat: false
+IndentFunctionDeclarationAfterType: false
+IndentWidth: 2
+KeepEmptyLinesAtTheStartOfBlocks: false
+MaxEmptyLinesToKeep: 1
+NamespaceIndentation: All
+PointerAlignment: Left
+SpaceBeforeAssignmentOperators: true
+SpaceBeforeParens: ControlStatements
+SpaceInEmptyParentheses: false
+SpacesBeforeTrailingComments: 1
+SpacesInAngles: false
+SpacesInContainerLiterals: false
+SpacesInCStyleCastParentheses: false
+SpacesInParentheses: false
+Standard: c++17
+UseTab: Never

.github/workflows/build-dev.yml

@@ -0,0 +1,40 @@
+name: Build Dev
+
+on:
+  pull_request:
+    branches:
+      - dev
+    paths:
+      - icicle/**
+      - src/**
+      - Cargo.toml
+      - build.rs
+
+env:
+  CARGO_TERM_COLOR: always
+  ARCH_TYPE: sm_70
+
+jobs:
+  build-linux:
+    runs-on: [self-hosted, Linux, X64, icicle]
+    steps:
+    - name: Checkout Repo
+      uses: actions/checkout@v3
+    - name: Build
+      run: cargo build --release --verbose
+      
+  
+  build-windows:
+    runs-on: windows-2022
+    steps:     
+    - name: Checkout Repo
+      uses: actions/checkout@v3
+    - name: Download and Install Cuda
+      uses: Jimver/cuda-toolkit@v0.2.11
+      with:
+        cuda: '12.0.0'
+        method: 'network'
+        # https://docs.nvidia.com/cuda/archive/12.0.0/cuda-installation-guide-microsoft-windows/index.html
+        sub-packages: '["cudart", "nvcc", "thrust"]'
+    - name: Build
+      run: cargo build --release --verbose

.github/workflows/build.yml

@@ -7,16 +7,6 @@ on:
       - opened
     branches:
       - main
-      - dev
-    paths:
-      - icicle/**
-      - src/**
-      - Cargo.toml
-      - build.rs
-  push:
-    branches-ignore:
-      - main
-      - dev
     paths:
       - icicle/**
       - src/**
@@ -29,32 +19,25 @@ env:
 
 jobs:
   build-linux:
-    runs-on: ubuntu-latest
-
+    runs-on: [self-hosted, Linux, X64, icicle]
     steps:
-    # Checkout code
-    - uses: actions/checkout@v3
-    # Download (or from cache) and install CUDA Toolkit 12.1.0
-    - uses: Jimver/cuda-toolkit@v0.2.9
-      id: cuda-toolkit
-      with:
-        cuda: '12.1.0'
-        use-github-cache: true
-      # Build from cargo - Rust utils are preinstalled on latest images
-      # https://github.com/actions/runner-images/blob/main/images/linux/Ubuntu2204-Readme.md#rust-tools
+    - name: Checkout Repo
+      uses: actions/checkout@v3
     - name: Build
       run: cargo build --release --verbose
       
   
   build-windows:
-    runs-on: windows-latest
-
+    runs-on: windows-2022
     steps:     
-    - uses: actions/checkout@v3
-    - uses: Jimver/cuda-toolkit@v0.2.9
-      id: cuda-toolkit
+    - name: Checkout Repo
+      uses: actions/checkout@v3
+    - name: Download and Install Cuda
+      uses: Jimver/cuda-toolkit@v0.2.11
       with:
-        cuda: '12.1.0'
-        use-github-cache: true
+        cuda: '12.0.0'
+        method: 'network'
+        # https://docs.nvidia.com/cuda/archive/12.0.0/cuda-installation-guide-microsoft-windows/index.html
+        sub-packages: '["cudart", "nvcc", "thrust"]'
     - name: Build
       run: cargo build --release --verbose

.gitignore

@@ -5,6 +5,9 @@
 *.cubin
 *.bin
 *.fatbin
+*.so
+*.nsys-rep
+*.ncu-rep
 **/target
 **/.vscode
 **/.*lock*csv#

.rustfmt.toml

@@ -0,0 +1,10 @@
+# https://github.com/rust-lang/rustfmt/blob/master/Configurations.md
+
+# Stable Configs
+chain_width = 0
+max_width = 120
+merge_derives = true
+use_field_init_shorthand = true
+use_try_shorthand = true
+
+# Unstable Configs

README.md

@@ -6,9 +6,11 @@
 
 
 <div align="center">
-![Build status](https://github.com/ingonyama-zk/icicle/actions/workflows/buil.yml/badge.svg)
+
+![Build status](https://github.com/ingonyama-zk/icicle/actions/workflows/build.yml/badge.svg)
 ![Discord server](https://img.shields.io/discord/1063033227788423299?label=Discord&logo=Discord&logoColor=%23&style=plastic)
 ![Follow us on twitter](https://img.shields.io/twitter/follow/Ingo_zk?style=social)
+
 </div>
 
 ## Background
@@ -161,6 +163,16 @@ Testing the new curve could be done by running the tests in ``tests_curve_name``
 
 Join our [Discord Server][DISCORD] and find us on the icicle channel. We will be happy to work together to support your use case and talk features, bugs and design.
 
+### Development Contributions
+
+If you are changing code, please make sure to change your [git hooks path][HOOKS_DOCS] to the repo's [hooks directory][HOOKS_PATH] by running the following command:
+
+```sh
+git config core.hooksPath ./scripts/hooks
+```
+
+This will ensure our custom hooks are run and will make it easier to follow our coding guidelines.
+
 ### Hall of Fame
 
 - [Robik](https://github.com/robik75), for his on-going support and mentorship 
@@ -183,5 +195,7 @@ See [LICENSE-MIT][LMIT] for details.
 [LMIT]: ./LICENSE
 [DISCORD]: https://discord.gg/Y4SkbDf2Ff
 [googletest]: https://github.com/google/googletest/
+[HOOKS_DOCS]: https://git-scm.com/docs/githooks
+[HOOKS_PATH]: ./scripts/hooks/
 
 <!-- End Links -->

benches/msm.rs

@@ -34,16 +34,20 @@ fn bench_msm(c: &mut Criterion) {
             #[cfg(feature = "g2")]
             let mut d_g2_points = DeviceBuffer::from_slice(&g2_batch_points[..]).unwrap();
 
-            group.sample_size(30).bench_function(
-                &format!("MSM of size 2^{} in batch {}", log_msm_size, batch_size),
-                |b| b.iter(|| commit_batch_bls12_381(&mut d_points, &mut d_scalars, batch_size)),
-            );
+            group
+                .sample_size(30)
+                .bench_function(
+                    &format!("MSM of size 2^{} in batch {}", log_msm_size, batch_size),
+                    |b| b.iter(|| commit_batch_bls12_381(&mut d_points, &mut d_scalars, batch_size)),
+                );
 
             #[cfg(feature = "g2")]
-            group.sample_size(10).bench_function(
-                &format!("G2 MSM of size 2^{} in batch {}", log_msm_size, batch_size),
-                |b| b.iter(|| commit_batch_g2(&mut d_g2_points, &mut d_scalars, batch_size))
-            );
+            group
+                .sample_size(10)
+                .bench_function(
+                    &format!("G2 MSM of size 2^{} in batch {}", log_msm_size, batch_size),
+                    |b| b.iter(|| commit_batch_g2(&mut d_g2_points, &mut d_scalars, batch_size)),
+                );
         }
     }
 }

benches/ntt.rs

@@ -2,33 +2,84 @@ extern crate criterion;
 
 use criterion::{criterion_group, criterion_main, Criterion};
 
-use icicle_utils::test_bls12_381::{interpolate_scalars_batch_bls12_381, interpolate_points_batch_bls12_381, set_up_scalars_bls12_381, set_up_points_bls12_381};
+use icicle_utils::test_bls12_381::*;
 
-
-const LOG_NTT_SIZES: [usize; 1] = [15];
-const BATCH_SIZES: [usize; 2] = [8, 16];
+const LOG_NTT_SIZES: [usize; 3] = [20, 9, 10];
+const BATCH_SIZES: [usize; 3] = [1, 512, 1024];
 
 fn bench_ntt(c: &mut Criterion) {
     let mut group = c.benchmark_group("NTT");
     for log_ntt_size in LOG_NTT_SIZES {
         for batch_size in BATCH_SIZES {
             let ntt_size = 1 << log_ntt_size;
+
+            if ntt_size * batch_size > 1 << 25 {
+                continue;
+            }
+
+            let scalar_samples = 20;
+
             let (_, mut d_evals, mut d_domain) = set_up_scalars_bls12_381(ntt_size * batch_size, log_ntt_size, true);
-            let (_, mut d_points_evals, _) = set_up_points_bls12_381(ntt_size * batch_size, log_ntt_size, true);
 
-            group.sample_size(100).bench_function(
-                &format!("Scalar NTT of size 2^{} in batch {}", log_ntt_size, batch_size),
-                |b| b.iter(|| interpolate_scalars_batch_bls12_381(&mut d_evals, &mut d_domain, batch_size))
-            );
+            group
+                .sample_size(scalar_samples)
+                .bench_function(
+                    &format!("Scalar NTT of size 2^{} in batch {}", log_ntt_size, batch_size),
+                    |b| b.iter(|| evaluate_scalars_batch_bls12_381(&mut d_evals, &mut d_domain, batch_size)),
+                );
 
-            group.sample_size(10).bench_function(
-                &format!("EC NTT of size 2^{} in batch {}", log_ntt_size, batch_size),
-                |b| b.iter(|| interpolate_points_batch_bls12_381(&mut d_points_evals, &mut d_domain, batch_size))
-            );
+            group
+                .sample_size(scalar_samples)
+                .bench_function(
+                    &format!("Scalar iNTT of size 2^{} in batch {}", log_ntt_size, batch_size),
+                    |b| b.iter(|| interpolate_scalars_batch_bls12_381(&mut d_evals, &mut d_domain, batch_size)),
+                );
+
+            group
+                .sample_size(scalar_samples)
+                .bench_function(
+                    &format!("Scalar inplace NTT of size 2^{} in batch {}", log_ntt_size, batch_size),
+                    |b| b.iter(|| ntt_inplace_batch_bls12_381(&mut d_evals, &mut d_domain, batch_size, false, 0)),
+                );
+
+            group
+                .sample_size(scalar_samples)
+                .bench_function(
+                    &format!("Scalar inplace iNTT of size 2^{} in batch {}", log_ntt_size, batch_size),
+                    |b| b.iter(|| ntt_inplace_batch_bls12_381(&mut d_evals, &mut d_domain, batch_size, true, 0)),
+                );
+
+            drop(d_evals);
+            drop(d_domain);
+
+            if ntt_size * batch_size > 1 << 18 {
+                continue;
+            }
+
+            let point_samples = 10;
+
+            let (_, mut d_points_evals, mut d_domain) =
+                set_up_points_bls12_381(ntt_size * batch_size, log_ntt_size, true);
+
+            group
+                .sample_size(point_samples)
+                .bench_function(
+                    &format!("EC NTT of size 2^{} in batch {}", log_ntt_size, batch_size),
+                    |b| b.iter(|| interpolate_points_batch_bls12_381(&mut d_points_evals, &mut d_domain, batch_size)),
+                );
+
+            group
+                .sample_size(point_samples)
+                .bench_function(
+                    &format!("EC iNTT of size 2^{} in batch {}", log_ntt_size, batch_size),
+                    |b| b.iter(|| evaluate_points_batch_bls12_381(&mut d_points_evals, &mut d_domain, batch_size)),
+                );
+
+            drop(d_points_evals);
+            drop(d_domain);
         }
     }
 }
 
 criterion_group!(ntt_benches, bench_ntt);
 criterion_main!(ntt_benches);
-

build.rs

@@ -26,8 +26,6 @@ fn main() {
     nvcc.debug(false);
     nvcc.flag(&arch);
     nvcc.flag(&stream);
-    nvcc.files([
-        "./icicle/curves/index.cu",
-    ]);
+    nvcc.files(["./icicle/curves/index.cu"]);
     nvcc.compile("ingo_icicle"); //TODO: extension??
 }

curve_parameters/bn254.json

@@ -3,11 +3,11 @@
     "modulus_p" : 21888242871839275222246405745257275088548364400416034343698204186575808495617,
     "bit_count_p" : 254,
     "limb_p" :  8,
-    "ntt_size" : 28,
+    "ntt_size" : 20,
     "modulus_q" : 21888242871839275222246405745257275088696311157297823662689037894645226208583,
     "bit_count_q" : 254,
     "limb_q" : 8,
-    "root_of_unity" : 19103219067921713944291392827692070036145651957329286315305642004821462161904,
+    "root_of_unity" : 19032961837237948602743626455740240236231119053033140765040043513661803148152,
     "weierstrass_b" : 3,
     "weierstrass_b_g2_re" : 19485874751759354771024239261021720505790618469301721065564631296452457478373,
     "weierstrass_b_g2_im" : 266929791119991161246907387137283842545076965332900288569378510910307636690,

curve_parameters/new_curve_script.py

@@ -204,14 +204,14 @@ newpath = f'./icicle/curves/{curve_name_lower}'
 if not os.path.exists(newpath):
     os.makedirs(newpath)
 
-with open("./icicle/curves/curve_template/params.cuh", "r") as params_file:
+with open("./icicle/curves/curve_template/params.cuh.tmpl", "r") as params_file:
     params_file_template = Template(params_file.read())
     params = get_params(config)
     params_content = params_file_template.safe_substitute(params)
     with open(f'./icicle/curves/{curve_name_lower}/params.cuh', 'w') as f:
         f.write(params_content)
 
-with open("./icicle/curves/curve_template/lde.cu", "r") as lde_file:
+with open("./icicle/curves/curve_template/lde.cu.tmpl", "r") as lde_file:
     template_content = Template(lde_file.read())
     lde_content = template_content.safe_substitute(
         CURVE_NAME_U=curve_name_upper, 
@@ -220,7 +220,7 @@ with open("./icicle/curves/curve_template/lde.cu", "r") as lde_file:
     with open(f'./icicle/curves/{curve_name_lower}/lde.cu', 'w') as f:
         f.write(lde_content)
     
-with open("./icicle/curves/curve_template/msm.cu", "r") as msm_file:
+with open("./icicle/curves/curve_template/msm.cu.tmpl", "r") as msm_file:
     template_content = Template(msm_file.read())
     msm_content = template_content.safe_substitute(
         CURVE_NAME_U=curve_name_upper, 
@@ -229,7 +229,7 @@ with open("./icicle/curves/curve_template/msm.cu", "r") as msm_file:
     with open(f'./icicle/curves/{curve_name_lower}/msm.cu', 'w') as f:
         f.write(msm_content)
 
-with open("./icicle/curves/curve_template/ve_mod_mult.cu", "r") as ve_mod_mult_file:
+with open("./icicle/curves/curve_template/ve_mod_mult.cu.tmpl", "r") as ve_mod_mult_file:
     template_content = Template(ve_mod_mult_file.read())
     ve_mod_mult_content = template_content.safe_substitute(
         CURVE_NAME_U=curve_name_upper, 
@@ -239,7 +239,7 @@ with open("./icicle/curves/curve_template/ve_mod_mult.cu", "r") as ve_mod_mult_f
         f.write(ve_mod_mult_content)
     
 
-with open(f'./icicle/curves/curve_template/curve_config.cuh', 'r') as cc:
+with open(f'./icicle/curves/curve_template/curve_config.cuh.tmpl', 'r') as cc:
     template_content = Template(cc.read())
     cc_content = template_content.safe_substitute(
         CURVE_NAME_U=curve_name_upper,
@@ -248,7 +248,7 @@ with open(f'./icicle/curves/curve_template/curve_config.cuh', 'r') as cc:
         f.write(cc_content)
     
 
-with open(f'./icicle/curves/curve_template/projective.cu', 'r') as proj:
+with open(f'./icicle/curves/curve_template/projective.cu.tmpl', 'r') as proj:
     template_content = Template(proj.read())
     proj_content = template_content.safe_substitute(
         CURVE_NAME_U=curve_name_upper, 
@@ -258,7 +258,7 @@ with open(f'./icicle/curves/curve_template/projective.cu', 'r') as proj:
         f.write(proj_content)
 
 
-with open(f'./icicle/curves/curve_template/supported_operations.cu', 'r') as supp_ops:
+with open(f'./icicle/curves/curve_template/supported_operations.cu.tmpl', 'r') as supp_ops:
     template_content = Template(supp_ops.read())
     supp_ops_content = template_content.safe_substitute()
     with open(f'./icicle/curves/{curve_name_lower}/supported_operations.cu', 'w') as f:
@@ -313,4 +313,4 @@ with open('./src/curves/mod.rs', 'r+') as f:
 with open('./src/lib.rs', 'r+') as f:
     lib_text = f.read()
     if lib_text.find(curve_name_lower) == -1:
-        f.write('\npub mod ' + curve_name_lower + ';')
+        f.write('\npub mod ' + curve_name_lower + ';')

examples/ntt/main.rs

@@ -0,0 +1,156 @@
+use std::time::Instant;
+
+use icicle_utils::{curves::bls12_381::ScalarField_BLS12_381, test_bls12_381::*};
+use rustacuda::prelude::DeviceBuffer;
+
+const LOG_NTT_SIZES: [usize; 3] = [20, 10, 9];
+const BATCH_SIZES: [usize; 3] = [1, 1 << 9, 1 << 10];
+
+const MAX_POINTS_LOG2: usize = 18;
+const MAX_SCALARS_LOG2: usize = 26;
+
+fn bench_lde() {
+    for log_ntt_size in LOG_NTT_SIZES {
+        for batch_size in BATCH_SIZES {
+            let ntt_size = 1 << log_ntt_size;
+
+            fn ntt_scalars_batch_bls12_381(
+                d_inout: &mut DeviceBuffer<ScalarField_BLS12_381>,
+                d_twiddles: &mut DeviceBuffer<ScalarField_BLS12_381>,
+                batch_size: usize,
+            ) -> i32 {
+                ntt_inplace_batch_bls12_381(d_inout, d_twiddles, batch_size, false, 0);
+                0
+            }
+
+            fn intt_scalars_batch_bls12_381(
+                d_inout: &mut DeviceBuffer<ScalarField_BLS12_381>,
+                d_twiddles: &mut DeviceBuffer<ScalarField_BLS12_381>,
+                batch_size: usize,
+            ) -> i32 {
+                ntt_inplace_batch_bls12_381(d_inout, d_twiddles, batch_size, true, 0);
+                0
+            }
+
+            // copy
+            bench_ntt_template(
+                MAX_SCALARS_LOG2,
+                ntt_size,
+                batch_size,
+                log_ntt_size,
+                set_up_scalars_bls12_381,
+                evaluate_scalars_batch_bls12_381,
+                "NTT",
+                false,
+                100,
+            );
+
+            bench_ntt_template(
+                MAX_SCALARS_LOG2,
+                ntt_size,
+                batch_size,
+                log_ntt_size,
+                set_up_scalars_bls12_381,
+                interpolate_scalars_batch_bls12_381,
+                "iNTT",
+                true,
+                100,
+            );
+
+            bench_ntt_template(
+                MAX_POINTS_LOG2,
+                ntt_size,
+                batch_size,
+                log_ntt_size,
+                set_up_points_bls12_381,
+                evaluate_points_batch_bls12_381,
+                "EC NTT",
+                false,
+                20,
+            );
+
+            bench_ntt_template(
+                MAX_POINTS_LOG2,
+                ntt_size,
+                batch_size,
+                log_ntt_size,
+                set_up_points_bls12_381,
+                interpolate_points_batch_bls12_381,
+                "EC iNTT",
+                true,
+                20,
+            );
+
+            // inplace
+            bench_ntt_template(
+                MAX_SCALARS_LOG2,
+                ntt_size,
+                batch_size,
+                log_ntt_size,
+                set_up_scalars_bls12_381,
+                ntt_scalars_batch_bls12_381,
+                "NTT inplace",
+                false,
+                100,
+            );
+
+            bench_ntt_template(
+                MAX_SCALARS_LOG2,
+                ntt_size,
+                batch_size,
+                log_ntt_size,
+                set_up_scalars_bls12_381,
+                intt_scalars_batch_bls12_381,
+                "iNTT inplace",
+                true,
+                100,
+            );
+        }
+    }
+}
+
+fn bench_ntt_template<E, S, R>(
+    log_max_size: usize,
+    ntt_size: usize,
+    batch_size: usize,
+    log_ntt_size: usize,
+    set_data: fn(test_size: usize, log_domain_size: usize, inverse: bool) -> (Vec<E>, DeviceBuffer<E>, DeviceBuffer<S>),
+    bench_fn: fn(d_evaluations: &mut DeviceBuffer<E>, d_domain: &mut DeviceBuffer<S>, batch_size: usize) -> R,
+    id: &str,
+    inverse: bool,
+    samples: usize,
+) -> Option<(Vec<E>, R)> {
+    let count = ntt_size * batch_size;
+
+    let bench_id = format!("{} of size 2^{} in batch {}", id, log_ntt_size, batch_size);
+
+    if count > 1 << log_max_size {
+        println!("Bench size exceeded: {}", bench_id);
+        return None;
+    }
+
+    println!("{}", bench_id);
+
+    let (input, mut d_evals, mut d_domain) = set_data(ntt_size * batch_size, log_ntt_size, inverse);
+
+    let first = bench_fn(&mut d_evals, &mut d_domain, batch_size);
+
+    let start = Instant::now();
+    for _ in 0..samples {
+        bench_fn(&mut d_evals, &mut d_domain, batch_size);
+    }
+    let elapsed = start.elapsed();
+    println!(
+        "{} {:0?} us x {} = {:?}",
+        bench_id,
+        elapsed.as_micros() as f32 / (samples as f32),
+        samples,
+        elapsed
+    );
+
+    Some((input, first))
+}
+
+fn main() {
+    bench_lde();
+}

go.mod

@@ -0,0 +1,23 @@
+module github.com/ingonyama-zk/icicle
+
+go 1.20
+
+require github.com/consensys/gnark-crypto v0.11.0
+
+require (
+	github.com/davecgh/go-spew v1.1.1 // indirect
+	github.com/jfeliu007/goplantuml v1.6.2 // indirect
+	github.com/pmezard/go-difflib v1.0.0 // indirect
+	github.com/spf13/afero v1.8.2 // indirect
+	golang.org/x/text v0.3.7 // indirect
+	gopkg.in/yaml.v3 v3.0.1 // indirect
+)
+
+require (
+	github.com/bits-and-blooms/bitset v1.5.0 // indirect
+	github.com/consensys/bavard v0.1.13
+	github.com/mmcloughlin/addchain v0.4.0 // indirect
+	github.com/stretchr/testify v1.8.3
+	golang.org/x/sys v0.2.0 // indirect
+	rsc.io/tmplfunc v0.0.3 // indirect
+)

go.sum

@@ -0,0 +1,459 @@
+cloud.google.com/go v0.26.0/go.mod h1:aQUYkXzVsufM+DwF1aE+0xfcU+56JwCaLick0ClmMTw=
+cloud.google.com/go v0.34.0/go.mod h1:aQUYkXzVsufM+DwF1aE+0xfcU+56JwCaLick0ClmMTw=
+cloud.google.com/go v0.38.0/go.mod h1:990N+gfupTy94rShfmMCWGDn0LpTmnzTp2qbd1dvSRU=
+cloud.google.com/go v0.44.1/go.mod h1:iSa0KzasP4Uvy3f1mN/7PiObzGgflwredwwASm/v6AU=
+cloud.google.com/go v0.44.2/go.mod h1:60680Gw3Yr4ikxnPRS/oxxkBccT6SA1yMk63TGekxKY=
+cloud.google.com/go v0.44.3/go.mod h1:60680Gw3Yr4ikxnPRS/oxxkBccT6SA1yMk63TGekxKY=
+cloud.google.com/go v0.45.1/go.mod h1:RpBamKRgapWJb87xiFSdk4g1CME7QZg3uwTez+TSTjc=
+cloud.google.com/go v0.46.3/go.mod h1:a6bKKbmY7er1mI7TEI4lsAkts/mkhTSZK8w33B4RAg0=
+cloud.google.com/go v0.50.0/go.mod h1:r9sluTvynVuxRIOHXQEHMFffphuXHOMZMycpNR5e6To=
+cloud.google.com/go v0.52.0/go.mod h1:pXajvRH/6o3+F9jDHZWQ5PbGhn+o8w9qiu/CffaVdO4=
+cloud.google.com/go v0.53.0/go.mod h1:fp/UouUEsRkN6ryDKNW/Upv/JBKnv6WDthjR6+vze6M=
+cloud.google.com/go v0.54.0/go.mod h1:1rq2OEkV3YMf6n/9ZvGWI3GWw0VoqH/1x2nd8Is/bPc=
+cloud.google.com/go v0.56.0/go.mod h1:jr7tqZxxKOVYizybht9+26Z/gUq7tiRzu+ACVAMbKVk=
+cloud.google.com/go v0.57.0/go.mod h1:oXiQ6Rzq3RAkkY7N6t3TcE6jE+CIBBbA36lwQ1JyzZs=
+cloud.google.com/go v0.62.0/go.mod h1:jmCYTdRCQuc1PHIIJ/maLInMho30T/Y0M4hTdTShOYc=
+cloud.google.com/go v0.65.0/go.mod h1:O5N8zS7uWy9vkA9vayVHs65eM1ubvY4h553ofrNHObY=
+cloud.google.com/go v0.72.0/go.mod h1:M+5Vjvlc2wnp6tjzE102Dw08nGShTscUx2nZMufOKPI=
+cloud.google.com/go v0.74.0/go.mod h1:VV1xSbzvo+9QJOxLDaJfTjx5e+MePCpCWwvftOeQmWk=
+cloud.google.com/go v0.75.0/go.mod h1:VGuuCn7PG0dwsd5XPVm2Mm3wlh3EL55/79EKB6hlPTY=
+cloud.google.com/go/bigquery v1.0.1/go.mod h1:i/xbL2UlR5RvWAURpBYZTtm/cXjCha9lbfbpx4poX+o=
+cloud.google.com/go/bigquery v1.3.0/go.mod h1:PjpwJnslEMmckchkHFfq+HTD2DmtT67aNFKH1/VBDHE=
+cloud.google.com/go/bigquery v1.4.0/go.mod h1:S8dzgnTigyfTmLBfrtrhyYhwRxG72rYxvftPBK2Dvzc=
+cloud.google.com/go/bigquery v1.5.0/go.mod h1:snEHRnqQbz117VIFhE8bmtwIDY80NLUZUMb4Nv6dBIg=
+cloud.google.com/go/bigquery v1.7.0/go.mod h1://okPTzCYNXSlb24MZs83e2Do+h+VXtc4gLoIoXIAPc=
+cloud.google.com/go/bigquery v1.8.0/go.mod h1:J5hqkt3O0uAFnINi6JXValWIb1v0goeZM77hZzJN/fQ=
+cloud.google.com/go/datastore v1.0.0/go.mod h1:LXYbyblFSglQ5pkeyhO+Qmw7ukd3C+pD7TKLgZqpHYE=
+cloud.google.com/go/datastore v1.1.0/go.mod h1:umbIZjpQpHh4hmRpGhH4tLFup+FVzqBi1b3c64qFpCk=
+cloud.google.com/go/pubsub v1.0.1/go.mod h1:R0Gpsv3s54REJCy4fxDixWD93lHJMoZTyQ2kNxGRt3I=
+cloud.google.com/go/pubsub v1.1.0/go.mod h1:EwwdRX2sKPjnvnqCa270oGRyludottCI76h+R3AArQw=
+cloud.google.com/go/pubsub v1.2.0/go.mod h1:jhfEVHT8odbXTkndysNHCcx0awwzvfOlguIAii9o8iA=
+cloud.google.com/go/pubsub v1.3.1/go.mod h1:i+ucay31+CNRpDW4Lu78I4xXG+O1r/MAHgjpRVR+TSU=
+cloud.google.com/go/storage v1.0.0/go.mod h1:IhtSnM/ZTZV8YYJWCY8RULGVqBDmpoyjwiyrjsg+URw=
+cloud.google.com/go/storage v1.5.0/go.mod h1:tpKbwo567HUNpVclU5sGELwQWBDZ8gh0ZeosJ0Rtdos=
+cloud.google.com/go/storage v1.6.0/go.mod h1:N7U0C8pVQ/+NIKOBQyamJIeKQKkZ+mxpohlUTyfDhBk=
+cloud.google.com/go/storage v1.8.0/go.mod h1:Wv1Oy7z6Yz3DshWRJFhqM/UCfaWIRTdp0RXyy7KQOVs=
+cloud.google.com/go/storage v1.10.0/go.mod h1:FLPqc6j+Ki4BU591ie1oL6qBQGu2Bl/tZ9ullr3+Kg0=
+cloud.google.com/go/storage v1.14.0/go.mod h1:GrKmX003DSIwi9o29oFT7YDnHYwZoctc3fOKtUw0Xmo=
+dmitri.shuralyov.com/gpu/mtl v0.0.0-20190408044501-666a987793e9/go.mod h1:H6x//7gZCb22OMCxBHrMx7a5I7Hp++hsVxbQ4BYO7hU=
+github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
+github.com/BurntSushi/xgb v0.0.0-20160522181843-27f122750802/go.mod h1:IVnqGOEym/WlBOVXweHU+Q+/VP0lqqI8lqeDx9IjBqo=
+github.com/bits-and-blooms/bitset v1.5.0 h1:NpE8frKRLGHIcEzkR+gZhiioW1+WbYV6fKwD6ZIpQT8=
+github.com/bits-and-blooms/bitset v1.5.0/go.mod h1:gIdJ4wp64HaoK2YrL1Q5/N7Y16edYb8uY+O0FJTyyDA=
+github.com/census-instrumentation/opencensus-proto v0.2.1/go.mod h1:f6KPmirojxKA12rnyqOA5BBL4O983OfeGPqjHWSTneU=
+github.com/chzyer/logex v1.1.10/go.mod h1:+Ywpsq7O8HXn0nuIou7OrIPyXbp3wmkHB+jjWRnGsAI=
+github.com/chzyer/readline v0.0.0-20180603132655-2972be24d48e/go.mod h1:nSuG5e5PlCu98SY8svDHJxuZscDgtXS6KTTbou5AhLI=
+github.com/chzyer/test v0.0.0-20180213035817-a1ea475d72b1/go.mod h1:Q3SI9o4m/ZMnBNeIyt5eFwwo7qiLfzFZmjNmxjkiQlU=
+github.com/client9/misspell v0.3.4/go.mod h1:qj6jICC3Q7zFZvVWo7KLAzC3yx5G7kyvSDkc90ppPyw=
+github.com/cncf/udpa/go v0.0.0-20191209042840-269d4d468f6f/go.mod h1:M8M6+tZqaGXZJjfX53e64911xZQV5JYwmTeXPW+k8Sc=
+github.com/cncf/udpa/go v0.0.0-20200629203442-efcf912fb354/go.mod h1:WmhPx2Nbnhtbo57+VJT5O0JRkEi1Wbu0z5j0R8u5Hbk=
+github.com/cncf/udpa/go v0.0.0-20201120205902-5459f2c99403/go.mod h1:WmhPx2Nbnhtbo57+VJT5O0JRkEi1Wbu0z5j0R8u5Hbk=
+github.com/consensys/bavard v0.1.13 h1:oLhMLOFGTLdlda/kma4VOJazblc7IM5y5QPd2A/YjhQ=
+github.com/consensys/bavard v0.1.13/go.mod h1:9ItSMtA/dXMAiL7BG6bqW2m3NdSEObYWoH223nGHukI=
+github.com/consensys/gnark-crypto v0.11.0 h1:QqzHQlwEqlQr5jfWblGDkwlKHpT+4QodYqqExkAtyks=
+github.com/consensys/gnark-crypto v0.11.0/go.mod h1:Iq/P3HHl0ElSjsg2E1gsMwhAyxnxoKK5nVyZKd+/KhU=
+github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
+github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
+github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
+github.com/envoyproxy/go-control-plane v0.9.0/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
+github.com/envoyproxy/go-control-plane v0.9.1-0.20191026205805-5f8ba28d4473/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
+github.com/envoyproxy/go-control-plane v0.9.4/go.mod h1:6rpuAdCZL397s3pYoYcLgu1mIlRU8Am5FuJP05cCM98=
+github.com/envoyproxy/go-control-plane v0.9.7/go.mod h1:cwu0lG7PUMfa9snN8LXBig5ynNVH9qI8YYLbd1fK2po=
+github.com/envoyproxy/go-control-plane v0.9.9-0.20201210154907-fd9021fe5dad/go.mod h1:cXg6YxExXjJnVBQHBLXeUAgxn2UodCpnH306RInaBQk=
+github.com/envoyproxy/protoc-gen-validate v0.1.0/go.mod h1:iSmxcyjqTsJpI2R4NaDN7+kN2VEUnK/pcBlmesArF7c=
+github.com/go-gl/glfw v0.0.0-20190409004039-e6da0acd62b1/go.mod h1:vR7hzQXu2zJy9AVAgeJqvqgH9Q5CA+iKCZ2gyEVpxRU=
+github.com/go-gl/glfw/v3.3/glfw v0.0.0-20191125211704-12ad95a8df72/go.mod h1:tQ2UAYgL5IevRw8kRxooKSPJfGvJ9fJQFa0TUsXzTg8=
+github.com/go-gl/glfw/v3.3/glfw v0.0.0-20200222043503-6f7a984d4dc4/go.mod h1:tQ2UAYgL5IevRw8kRxooKSPJfGvJ9fJQFa0TUsXzTg8=
+github.com/golang/glog v0.0.0-20160126235308-23def4e6c14b/go.mod h1:SBH7ygxi8pfUlaOkMMuAQtPIUF8ecWP5IEl/CR7VP2Q=
+github.com/golang/groupcache v0.0.0-20190702054246-869f871628b6/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
+github.com/golang/groupcache v0.0.0-20191227052852-215e87163ea7/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
+github.com/golang/groupcache v0.0.0-20200121045136-8c9f03a8e57e/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
+github.com/golang/mock v1.1.1/go.mod h1:oTYuIxOrZwtPieC+H1uAHpcLFnEyAGVDL/k47Jfbm0A=
+github.com/golang/mock v1.2.0/go.mod h1:oTYuIxOrZwtPieC+H1uAHpcLFnEyAGVDL/k47Jfbm0A=
+github.com/golang/mock v1.3.1/go.mod h1:sBzyDLLjw3U8JLTeZvSv8jJB+tU5PVekmnlKIyFUx0Y=
+github.com/golang/mock v1.4.0/go.mod h1:UOMv5ysSaYNkG+OFQykRIcU/QvvxJf3p21QfJ2Bt3cw=
+github.com/golang/mock v1.4.1/go.mod h1:UOMv5ysSaYNkG+OFQykRIcU/QvvxJf3p21QfJ2Bt3cw=
+github.com/golang/mock v1.4.3/go.mod h1:UOMv5ysSaYNkG+OFQykRIcU/QvvxJf3p21QfJ2Bt3cw=
+github.com/golang/mock v1.4.4/go.mod h1:l3mdAwkq5BuhzHwde/uurv3sEJeZMXNpwsxVWU71h+4=
+github.com/golang/protobuf v1.2.0/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
+github.com/golang/protobuf v1.3.1/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
+github.com/golang/protobuf v1.3.2/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
+github.com/golang/protobuf v1.3.3/go.mod h1:vzj43D7+SQXF/4pzW/hwtAqwc6iTitCiVSaWz5lYuqw=
+github.com/golang/protobuf v1.3.4/go.mod h1:vzj43D7+SQXF/4pzW/hwtAqwc6iTitCiVSaWz5lYuqw=
+github.com/golang/protobuf v1.3.5/go.mod h1:6O5/vntMXwX2lRkT1hjjk0nAC1IDOTvTlVgjlRvqsdk=
+github.com/golang/protobuf v1.4.0-rc.1/go.mod h1:ceaxUfeHdC40wWswd/P6IGgMaK3YpKi5j83Wpe3EHw8=
+github.com/golang/protobuf v1.4.0-rc.1.0.20200221234624-67d41d38c208/go.mod h1:xKAWHe0F5eneWXFV3EuXVDTCmh+JuBKY0li0aMyXATA=
+github.com/golang/protobuf v1.4.0-rc.2/go.mod h1:LlEzMj4AhA7rCAGe4KMBDvJI+AwstrUpVNzEA03Pprs=
+github.com/golang/protobuf v1.4.0-rc.4.0.20200313231945-b860323f09d0/go.mod h1:WU3c8KckQ9AFe+yFwt9sWVRKCVIyN9cPHBJSNnbL67w=
+github.com/golang/protobuf v1.4.0/go.mod h1:jodUvKwWbYaEsadDk5Fwe5c77LiNKVO9IDvqG2KuDX0=
+github.com/golang/protobuf v1.4.1/go.mod h1:U8fpvMrcmy5pZrNK1lt4xCsGvpyWQ/VVv6QDs8UjoX8=
+github.com/golang/protobuf v1.4.2/go.mod h1:oDoupMAO8OvCJWAcko0GGGIgR6R6ocIYbsSw735rRwI=
+github.com/golang/protobuf v1.4.3/go.mod h1:oDoupMAO8OvCJWAcko0GGGIgR6R6ocIYbsSw735rRwI=
+github.com/google/btree v0.0.0-20180813153112-4030bb1f1f0c/go.mod h1:lNA+9X1NB3Zf8V7Ke586lFgjr2dZNuvo3lPJSGZ5JPQ=
+github.com/google/btree v1.0.0/go.mod h1:lNA+9X1NB3Zf8V7Ke586lFgjr2dZNuvo3lPJSGZ5JPQ=
+github.com/google/go-cmp v0.2.0/go.mod h1:oXzfMopK8JAjlY9xF4vHSVASa0yLyX7SntLO5aqRK0M=
+github.com/google/go-cmp v0.3.0/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
+github.com/google/go-cmp v0.3.1/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
+github.com/google/go-cmp v0.4.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.4.1/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.5.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.5.1/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.5.2/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.5.4/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/martian v2.1.0+incompatible/go.mod h1:9I4somxYTbIHy5NJKHRl3wXiIaQGbYVAs8BPL6v8lEs=
+github.com/google/martian/v3 v3.0.0/go.mod h1:y5Zk1BBys9G+gd6Jrk0W3cC1+ELVxBWuIGO+w/tUAp0=
+github.com/google/martian/v3 v3.1.0/go.mod h1:y5Zk1BBys9G+gd6Jrk0W3cC1+ELVxBWuIGO+w/tUAp0=
+github.com/google/pprof v0.0.0-20181206194817-3ea8567a2e57/go.mod h1:zfwlbNMJ+OItoe0UupaVj+oy1omPYYDuagoSzA8v9mc=
+github.com/google/pprof v0.0.0-20190515194954-54271f7e092f/go.mod h1:zfwlbNMJ+OItoe0UupaVj+oy1omPYYDuagoSzA8v9mc=
+github.com/google/pprof v0.0.0-20191218002539-d4f498aebedc/go.mod h1:ZgVRPoUq/hfqzAqh7sHMqb3I9Rq5C59dIz2SbBwJ4eM=
+github.com/google/pprof v0.0.0-20200212024743-f11f1df84d12/go.mod h1:ZgVRPoUq/hfqzAqh7sHMqb3I9Rq5C59dIz2SbBwJ4eM=
+github.com/google/pprof v0.0.0-20200229191704-1ebb73c60ed3/go.mod h1:ZgVRPoUq/hfqzAqh7sHMqb3I9Rq5C59dIz2SbBwJ4eM=
+github.com/google/pprof v0.0.0-20200430221834-fc25d7d30c6d/go.mod h1:ZgVRPoUq/hfqzAqh7sHMqb3I9Rq5C59dIz2SbBwJ4eM=
+github.com/google/pprof v0.0.0-20200708004538-1a94d8640e99/go.mod h1:ZgVRPoUq/hfqzAqh7sHMqb3I9Rq5C59dIz2SbBwJ4eM=
+github.com/google/pprof v0.0.0-20201023163331-3e6fc7fc9c4c/go.mod h1:kpwsk12EmLew5upagYY7GY0pfYCcupk39gWOCRROcvE=
+github.com/google/pprof v0.0.0-20201203190320-1bf35d6f28c2/go.mod h1:kpwsk12EmLew5upagYY7GY0pfYCcupk39gWOCRROcvE=
+github.com/google/pprof v0.0.0-20201218002935-b9804c9f04c2/go.mod h1:kpwsk12EmLew5upagYY7GY0pfYCcupk39gWOCRROcvE=
+github.com/google/renameio v0.1.0/go.mod h1:KWCgfxg9yswjAJkECMjeO8J8rahYeXnNhOm40UhjYkI=
+github.com/google/subcommands v1.2.0/go.mod h1:ZjhPrFU+Olkh9WazFPsl27BQ4UPiG37m3yTrtFlrHVk=
+github.com/google/uuid v1.1.2/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
+github.com/googleapis/gax-go/v2 v2.0.4/go.mod h1:0Wqv26UfaUD9n4G6kQubkQ+KchISgw+vpHVxEJEs9eg=
+github.com/googleapis/gax-go/v2 v2.0.5/go.mod h1:DWXyrwAJ9X0FpwwEdw+IPEYBICEFu5mhpdKc/us6bOk=
+github.com/googleapis/google-cloud-go-testing v0.0.0-20200911160855-bcd43fbb19e8/go.mod h1:dvDLG8qkwmyD9a/MJJN3XJcT3xFxOKAvTZGvuZmac9g=
+github.com/hashicorp/golang-lru v0.5.0/go.mod h1:/m3WP610KZHVQ1SGc6re/UDhFvYD7pJ4Ao+sR/qLZy8=
+github.com/hashicorp/golang-lru v0.5.1/go.mod h1:/m3WP610KZHVQ1SGc6re/UDhFvYD7pJ4Ao+sR/qLZy8=
+github.com/ianlancetaylor/demangle v0.0.0-20181102032728-5e5cf60278f6/go.mod h1:aSSvb/t6k1mPoxDqO4vJh6VOCGPwU4O0C2/Eqndh1Sc=
+github.com/ianlancetaylor/demangle v0.0.0-20200824232613-28f6c0f3b639/go.mod h1:aSSvb/t6k1mPoxDqO4vJh6VOCGPwU4O0C2/Eqndh1Sc=
+github.com/jfeliu007/goplantuml v1.6.2 h1:WggxwLIBLF8P9OU/G/H2+IeHpEx5I7QK780jb78twEM=
+github.com/jfeliu007/goplantuml v1.6.2/go.mod h1:GnvyYGyIXD68akNFe2FBlNBypwfbpeNmVUQ4ZxJw8iI=
+github.com/jstemmer/go-junit-report v0.0.0-20190106144839-af01ea7f8024/go.mod h1:6v2b51hI/fHJwM22ozAgKL4VKDeJcHhJFhtBdhmNjmU=
+github.com/jstemmer/go-junit-report v0.9.1/go.mod h1:Brl9GWCQeLvo8nXZwPNNblvFj/XSXhF0NWZEnDohbsk=
+github.com/kisielk/gotool v1.0.0/go.mod h1:XhKaO+MFFWcvkIS/tQcRk01m1F5IRFswLeQ+oQHNcck=
+github.com/kr/fs v0.1.0/go.mod h1:FFnZGqtBN9Gxj7eW1uZ42v5BccTP0vu6NEaFoC2HwRg=
+github.com/kr/pretty v0.1.0/go.mod h1:dAy3ld7l9f0ibDNOQOHHMYYIIbhfbHSm3C4ZsoJORNo=
+github.com/kr/pty v1.1.1/go.mod h1:pFQYn66WHrOpPYNljwOMqo10TkYh1fy3cYio2l3bCsQ=
+github.com/kr/text v0.1.0/go.mod h1:4Jbv+DJW3UT/LiOwJeYQe1efqtUx/iVham/4vfdArNI=
+github.com/leanovate/gopter v0.2.9 h1:fQjYxZaynp97ozCzfOyOuAGOU4aU/z37zf/tOujFk7c=
+github.com/mmcloughlin/addchain v0.4.0 h1:SobOdjm2xLj1KkXN5/n0xTIWyZA2+s99UCY1iPfkHRY=
+github.com/mmcloughlin/addchain v0.4.0/go.mod h1:A86O+tHqZLMNO4w6ZZ4FlVQEadcoqkyU72HC5wJ4RlU=
+github.com/mmcloughlin/profile v0.1.1/go.mod h1:IhHD7q1ooxgwTgjxQYkACGA77oFTDdFVejUS1/tS/qU=
+github.com/pkg/errors v0.9.1/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
+github.com/pkg/sftp v1.13.1/go.mod h1:3HaPG6Dq1ILlpPZRO0HVMrsydcdLt6HRDccSgb87qRg=
+github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
+github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
+github.com/prometheus/client_model v0.0.0-20190812154241-14fe0d1b01d4/go.mod h1:xMI15A0UPsDsEKsMN9yxemIoYk6Tm2C1GtYGdfGttqA=
+github.com/rogpeppe/go-internal v1.3.0/go.mod h1:M8bDsm7K2OlrFYOpmOWEs/qY81heoFRclV5y23lUDJ4=
+github.com/spf13/afero v1.8.2 h1:xehSyVa0YnHWsJ49JFljMpg1HX19V6NDZ1fkm1Xznbo=
+github.com/spf13/afero v1.8.2/go.mod h1:CtAatgMJh6bJEIs48Ay/FOnkljP3WeGUG0MC1RfAqwo=
+github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
+github.com/stretchr/testify v1.4.0/go.mod h1:j7eGeouHqKxXV5pUuKE4zz7dFj8WfuZ+81PSLYec5m4=
+github.com/stretchr/testify v1.5.1/go.mod h1:5W2xD1RspED5o8YsWQXVCued0rvSQ+mT+I5cxcmMvtA=
+github.com/stretchr/testify v1.7.0/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
+github.com/stretchr/testify v1.8.3 h1:RP3t2pwF7cMEbC1dqtB6poj3niw/9gnV4Cjg5oW5gtY=
+github.com/stretchr/testify v1.8.3/go.mod h1:sz/lmYIOXD/1dqDmKjjqLyZ2RngseejIcXlSw2iwfAo=
+github.com/yuin/goldmark v1.1.25/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9decYSb74=
+github.com/yuin/goldmark v1.1.27/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9decYSb74=
+github.com/yuin/goldmark v1.1.32/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9decYSb74=
+github.com/yuin/goldmark v1.2.1/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9decYSb74=
+go.opencensus.io v0.21.0/go.mod h1:mSImk1erAIZhrmZN+AvHh14ztQfjbGwt4TtuofqLduU=
+go.opencensus.io v0.22.0/go.mod h1:+kGneAE2xo2IficOXnaByMWTGM9T73dGwxeWcUqIpI8=
+go.opencensus.io v0.22.2/go.mod h1:yxeiOL68Rb0Xd1ddK5vPZ/oVn4vY4Ynel7k9FzqtOIw=
+go.opencensus.io v0.22.3/go.mod h1:yxeiOL68Rb0Xd1ddK5vPZ/oVn4vY4Ynel7k9FzqtOIw=
+go.opencensus.io v0.22.4/go.mod h1:yxeiOL68Rb0Xd1ddK5vPZ/oVn4vY4Ynel7k9FzqtOIw=
+go.opencensus.io v0.22.5/go.mod h1:5pWMHQbX5EPX2/62yrJeAkowc+lfs/XD7Uxpq3pI6kk=
+golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
+golang.org/x/crypto v0.0.0-20190510104115-cbcb75029529/go.mod h1:yigFU9vqHzYiE8UmvKecakEJjdnWj3jj499lnFckfCI=
+golang.org/x/crypto v0.0.0-20190605123033-f99c8df09eb5/go.mod h1:yigFU9vqHzYiE8UmvKecakEJjdnWj3jj499lnFckfCI=
+golang.org/x/crypto v0.0.0-20191011191535-87dc89f01550/go.mod h1:yigFU9vqHzYiE8UmvKecakEJjdnWj3jj499lnFckfCI=
+golang.org/x/crypto v0.0.0-20200622213623-75b288015ac9/go.mod h1:LzIPMQfyMNhhGPhUkYOs5KpL4U8rLKemX1yGLhDgUto=
+golang.org/x/crypto v0.0.0-20210421170649-83a5a9bb288b/go.mod h1:T9bdIzuCu7OtxOm1hfPfRQxPLYneinmdGuTeoZ9dtd4=
+golang.org/x/crypto v0.0.0-20211108221036-ceb1ce70b4fa/go.mod h1:GvvjBRRGRdwPK5ydBHafDWAxML/pGHZbMvKqRZ5+Abc=
+golang.org/x/exp v0.0.0-20190121172915-509febef88a4/go.mod h1:CJ0aWSM057203Lf6IL+f9T1iT9GByDxfZKAQTCR3kQA=
+golang.org/x/exp v0.0.0-20190306152737-a1d7652674e8/go.mod h1:CJ0aWSM057203Lf6IL+f9T1iT9GByDxfZKAQTCR3kQA=
+golang.org/x/exp v0.0.0-20190510132918-efd6b22b2522/go.mod h1:ZjyILWgesfNpC6sMxTJOJm9Kp84zZh5NQWvqDGG3Qr8=
+golang.org/x/exp v0.0.0-20190829153037-c13cbed26979/go.mod h1:86+5VVa7VpoJ4kLfm080zCjGlMRFzhUhsZKEZO7MGek=
+golang.org/x/exp v0.0.0-20191030013958-a1ab85dbe136/go.mod h1:JXzH8nQsPlswgeRAPE3MuO9GYsAcnJvJ4vnMwN/5qkY=
+golang.org/x/exp v0.0.0-20191129062945-2f5052295587/go.mod h1:2RIsYlXP63K8oxa1u096TMicItID8zy7Y6sNkU49FU4=
+golang.org/x/exp v0.0.0-20191227195350-da58074b4299/go.mod h1:2RIsYlXP63K8oxa1u096TMicItID8zy7Y6sNkU49FU4=
+golang.org/x/exp v0.0.0-20200119233911-0405dc783f0a/go.mod h1:2RIsYlXP63K8oxa1u096TMicItID8zy7Y6sNkU49FU4=
+golang.org/x/exp v0.0.0-20200207192155-f17229e696bd/go.mod h1:J/WKrq2StrnmMY6+EHIKF9dgMWnmCNThgcyBT1FY9mM=
+golang.org/x/exp v0.0.0-20200224162631-6cc2880d07d6/go.mod h1:3jZMyOhIsHpP37uCMkUooju7aAi5cS1Q23tOzKc+0MU=
+golang.org/x/image v0.0.0-20190227222117-0694c2d4d067/go.mod h1:kZ7UVZpmo3dzQBMxlp+ypCbDeSB+sBbTgSJuh5dn5js=
+golang.org/x/image v0.0.0-20190802002840-cff245a6509b/go.mod h1:FeLwcggjj3mMvU+oOTbSwawSJRM1uh48EjtB4UJZlP0=
+golang.org/x/lint v0.0.0-20181026193005-c67002cb31c3/go.mod h1:UVdnD1Gm6xHRNCYTkRU2/jEulfH38KcIWyp/GAMgvoE=
+golang.org/x/lint v0.0.0-20190227174305-5b3e6a55c961/go.mod h1:wehouNa3lNwaWXcvxsM5YxQ5yQlVC4a0KAMCusXpPoU=
+golang.org/x/lint v0.0.0-20190301231843-5614ed5bae6f/go.mod h1:UVdnD1Gm6xHRNCYTkRU2/jEulfH38KcIWyp/GAMgvoE=
+golang.org/x/lint v0.0.0-20190313153728-d0100b6bd8b3/go.mod h1:6SW0HCj/g11FgYtHlgUYUwCkIfeOF89ocIRzGO/8vkc=
+golang.org/x/lint v0.0.0-20190409202823-959b441ac422/go.mod h1:6SW0HCj/g11FgYtHlgUYUwCkIfeOF89ocIRzGO/8vkc=
+golang.org/x/lint v0.0.0-20190909230951-414d861bb4ac/go.mod h1:6SW0HCj/g11FgYtHlgUYUwCkIfeOF89ocIRzGO/8vkc=
+golang.org/x/lint v0.0.0-20190930215403-16217165b5de/go.mod h1:6SW0HCj/g11FgYtHlgUYUwCkIfeOF89ocIRzGO/8vkc=
+golang.org/x/lint v0.0.0-20191125180803-fdd1cda4f05f/go.mod h1:5qLYkcX4OjUUV8bRuDixDT3tpyyb+LUpUlRWLxfhWrs=
+golang.org/x/lint v0.0.0-20200130185559-910be7a94367/go.mod h1:3xt1FjdF8hUf6vQPIChWIBhFzV8gjjsPE/fR3IyQdNY=
+golang.org/x/lint v0.0.0-20200302205851-738671d3881b/go.mod h1:3xt1FjdF8hUf6vQPIChWIBhFzV8gjjsPE/fR3IyQdNY=
+golang.org/x/lint v0.0.0-20201208152925-83fdc39ff7b5/go.mod h1:3xt1FjdF8hUf6vQPIChWIBhFzV8gjjsPE/fR3IyQdNY=
+golang.org/x/mobile v0.0.0-20190312151609-d3739f865fa6/go.mod h1:z+o9i4GpDbdi3rU15maQ/Ox0txvL9dWGYEHz965HBQE=
+golang.org/x/mobile v0.0.0-20190719004257-d2bd2a29d028/go.mod h1:E/iHnbuqvinMTCcRqshq8CkpyQDoeVncDDYHnLhea+o=
+golang.org/x/mod v0.0.0-20190513183733-4bf6d317e70e/go.mod h1:mXi4GBBbnImb6dmsKGUJ2LatrhH/nqhxcFungHvyanc=
+golang.org/x/mod v0.1.0/go.mod h1:0QHyrYULN0/3qlju5TqG8bIK38QM8yzMo5ekMj3DlcY=
+golang.org/x/mod v0.1.1-0.20191105210325-c90efee705ee/go.mod h1:QqPTAvyqsEbceGzBzNggFXnrqF1CaUcvgkdR5Ot7KZg=
+golang.org/x/mod v0.1.1-0.20191107180719-034126e5016b/go.mod h1:QqPTAvyqsEbceGzBzNggFXnrqF1CaUcvgkdR5Ot7KZg=
+golang.org/x/mod v0.2.0/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=
+golang.org/x/mod v0.3.0/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=
+golang.org/x/mod v0.4.0/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=
+golang.org/x/mod v0.4.1/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=
+golang.org/x/net v0.0.0-20180724234803-3673e40ba225/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20180826012351-8a410e7b638d/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20190108225652-1e06a53dbb7e/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20190213061140-3a22650c66bd/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20190311183353-d8887717615a/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20190404232315-eb5bcb51f2a3/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20190501004415-9ce7a6920f09/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20190503192946-f4e77d36d62c/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20190603091049-60506f45cf65/go.mod h1:HSz+uSET+XFnRR8LxR5pz3Of3rY3CfYBVs4xY44aLks=
+golang.org/x/net v0.0.0-20190620200207-3b0461eec859/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20190628185345-da137c7871d7/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20190724013045-ca1201d0de80/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20191209160850-c0dbc17a3553/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20200114155413-6afb5195e5aa/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20200202094626-16171245cfb2/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20200222125558-5a598a2470a0/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20200226121028-0de0cce0169b/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20200301022130-244492dfa37a/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
+golang.org/x/net v0.0.0-20200324143707-d3edc9973b7e/go.mod h1:qpuaurCH72eLCgpAm/N6yyVIVM9cpaDIP3A8BGJEC5A=
+golang.org/x/net v0.0.0-20200501053045-e0ff5e5a1de5/go.mod h1:qpuaurCH72eLCgpAm/N6yyVIVM9cpaDIP3A8BGJEC5A=
+golang.org/x/net v0.0.0-20200506145744-7e3656a0809f/go.mod h1:qpuaurCH72eLCgpAm/N6yyVIVM9cpaDIP3A8BGJEC5A=
+golang.org/x/net v0.0.0-20200513185701-a91f0712d120/go.mod h1:qpuaurCH72eLCgpAm/N6yyVIVM9cpaDIP3A8BGJEC5A=
+golang.org/x/net v0.0.0-20200520182314-0ba52f642ac2/go.mod h1:qpuaurCH72eLCgpAm/N6yyVIVM9cpaDIP3A8BGJEC5A=
+golang.org/x/net v0.0.0-20200625001655-4c5254603344/go.mod h1:/O7V0waA8r7cgGh81Ro3o1hOxt32SMVPicZroKQ2sZA=
+golang.org/x/net v0.0.0-20200707034311-ab3426394381/go.mod h1:/O7V0waA8r7cgGh81Ro3o1hOxt32SMVPicZroKQ2sZA=
+golang.org/x/net v0.0.0-20200822124328-c89045814202/go.mod h1:/O7V0waA8r7cgGh81Ro3o1hOxt32SMVPicZroKQ2sZA=
+golang.org/x/net v0.0.0-20201021035429-f5854403a974/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
+golang.org/x/net v0.0.0-20201031054903-ff519b6c9102/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
+golang.org/x/net v0.0.0-20201209123823-ac852fbbde11/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
+golang.org/x/net v0.0.0-20201224014010-6772e930b67b/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
+golang.org/x/net v0.0.0-20210226172049-e18ecbb05110/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
+golang.org/x/oauth2 v0.0.0-20180821212333-d2e6202438be/go.mod h1:N/0e6XlmueqKjAGxoOufVs8QHGRruUQn6yWY3a++T0U=
+golang.org/x/oauth2 v0.0.0-20190226205417-e64efc72b421/go.mod h1:gOpvHmFTYa4IltrdGE7lF6nIHvwfUNPOp7c8zoXwtLw=
+golang.org/x/oauth2 v0.0.0-20190604053449-0f29369cfe45/go.mod h1:gOpvHmFTYa4IltrdGE7lF6nIHvwfUNPOp7c8zoXwtLw=
+golang.org/x/oauth2 v0.0.0-20191202225959-858c2ad4c8b6/go.mod h1:gOpvHmFTYa4IltrdGE7lF6nIHvwfUNPOp7c8zoXwtLw=
+golang.org/x/oauth2 v0.0.0-20200107190931-bf48bf16ab8d/go.mod h1:gOpvHmFTYa4IltrdGE7lF6nIHvwfUNPOp7c8zoXwtLw=
+golang.org/x/oauth2 v0.0.0-20200902213428-5d25da1a8d43/go.mod h1:KelEdhl1UZF7XfJ4dDtk6s++YSgaE7mD/BuKKDLBl4A=
+golang.org/x/oauth2 v0.0.0-20201109201403-9fd604954f58/go.mod h1:KelEdhl1UZF7XfJ4dDtk6s++YSgaE7mD/BuKKDLBl4A=
+golang.org/x/oauth2 v0.0.0-20201208152858-08078c50e5b5/go.mod h1:KelEdhl1UZF7XfJ4dDtk6s++YSgaE7mD/BuKKDLBl4A=
+golang.org/x/oauth2 v0.0.0-20210218202405-ba52d332ba99/go.mod h1:KelEdhl1UZF7XfJ4dDtk6s++YSgaE7mD/BuKKDLBl4A=
+golang.org/x/sync v0.0.0-20180314180146-1d60e4601c6f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20181108010431-42b317875d0f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20181221193216-37e7f081c4d4/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20190227155943-e225da77a7e6/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20190423024810-112230192c58/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20190911185100-cd5d95a43a6e/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20200317015054-43a5402ce75a/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20200625203802-6e8e738ad208/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20201020160332-67f06af15bc9/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20201207232520-09787c993a3a/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sys v0.0.0-20180830151530-49385e6e1522/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20190312061237-fead79001313/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190412213103-97732733099d/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190502145724-3ef323f4f1fd/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190507160741-ecd444e8653b/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190606165138-5da285871e9c/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190624142023-c5567b49c5d0/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20190726091711-fc99dfbffb4e/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20191001151750-bb3f8db39f24/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20191204072324-ce4227a45e2e/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20191228213918-04cbcbbfeed8/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200113162924-86b910548bc1/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200122134326-e047566fdf82/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200202164722-d101bd2416d5/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200212091648-12a6c2dcc1e4/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200223170610-d5e6a3e2c0ae/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200302150141-5c8b2ff67527/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200323222414-85ca7c5b95cd/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200331124033-c3d80250170d/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200501052902-10377860bb8e/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200511232937-7e40ca221e25/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200515095857-1151b9dac4a9/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200523222454-059865788121/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200803210538-64077c9b5642/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200905004654-be1d3432aa8f/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200930185726-fdedc70b468f/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20201119102817-f84b799fce68/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20201201145000-ef89a241ccb3/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20210104204734-6f8348627aad/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20210119212857-b64e53b001e4/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20210225134936-a50acf3fe073/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20210423185535-09eb48e85fd7/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20210615035016-665e8c7367d1/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.2.0 h1:ljd4t30dBnAvMZaQCevtY0xLLD0A+bRZXbgLMLU1F/A=
+golang.org/x/sys v0.2.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
+golang.org/x/text v0.0.0-20170915032832-14c0d48ead0c/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
+golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
+golang.org/x/text v0.3.1-0.20180807135948-17ff2d5776d2/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
+golang.org/x/text v0.3.2/go.mod h1:bEr9sfX3Q8Zfm5fL9x+3itogRgK3+ptLWKqgva+5dAk=
+golang.org/x/text v0.3.3/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
+golang.org/x/text v0.3.4/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
+golang.org/x/text v0.3.7 h1:olpwvP2KacW1ZWvsR7uQhoyTYvKAupfQrRGBFM352Gk=
+golang.org/x/text v0.3.7/go.mod h1:u+2+/6zg+i71rQMx5EYifcz6MCKuco9NR6JIITiCfzQ=
+golang.org/x/time v0.0.0-20181108054448-85acf8d2951c/go.mod h1:tRJNPiyCQ0inRvYxbN9jk5I+vvW/OXSQhTDSoE431IQ=
+golang.org/x/time v0.0.0-20190308202827-9d24e82272b4/go.mod h1:tRJNPiyCQ0inRvYxbN9jk5I+vvW/OXSQhTDSoE431IQ=
+golang.org/x/time v0.0.0-20191024005414-555d28b269f0/go.mod h1:tRJNPiyCQ0inRvYxbN9jk5I+vvW/OXSQhTDSoE431IQ=
+golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20190114222345-bf090417da8b/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20190226205152-f727befe758c/go.mod h1:9Yl7xja0Znq3iFh3HoIrodX9oNMXvdceNzlUR8zjMvY=
+golang.org/x/tools v0.0.0-20190311212946-11955173bddd/go.mod h1:LCzVGOaR6xXOjkQ3onu1FJEFr0SW1gC7cKk1uF8kGRs=
+golang.org/x/tools v0.0.0-20190312151545-0bb0c0a6e846/go.mod h1:LCzVGOaR6xXOjkQ3onu1FJEFr0SW1gC7cKk1uF8kGRs=
+golang.org/x/tools v0.0.0-20190312170243-e65039ee4138/go.mod h1:LCzVGOaR6xXOjkQ3onu1FJEFr0SW1gC7cKk1uF8kGRs=
+golang.org/x/tools v0.0.0-20190425150028-36563e24a262/go.mod h1:RgjU9mgBXZiqYHBnxXauZ1Gv1EHHAz9KjViQ78xBX0Q=
+golang.org/x/tools v0.0.0-20190506145303-2d16b83fe98c/go.mod h1:RgjU9mgBXZiqYHBnxXauZ1Gv1EHHAz9KjViQ78xBX0Q=
+golang.org/x/tools v0.0.0-20190524140312-2c0ae7006135/go.mod h1:RgjU9mgBXZiqYHBnxXauZ1Gv1EHHAz9KjViQ78xBX0Q=
+golang.org/x/tools v0.0.0-20190606124116-d0a3d012864b/go.mod h1:/rFqwRUd4F7ZHNgwSSTFct+R/Kf4OFW1sUzUTQQTgfc=
+golang.org/x/tools v0.0.0-20190621195816-6e04913cbbac/go.mod h1:/rFqwRUd4F7ZHNgwSSTFct+R/Kf4OFW1sUzUTQQTgfc=
+golang.org/x/tools v0.0.0-20190628153133-6cdbf07be9d0/go.mod h1:/rFqwRUd4F7ZHNgwSSTFct+R/Kf4OFW1sUzUTQQTgfc=
+golang.org/x/tools v0.0.0-20190816200558-6889da9d5479/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20190911174233-4f2ddba30aff/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191012152004-8de300cfc20a/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191113191852-77e3bb0ad9e7/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191115202509-3a792d9c32b2/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191119224855-298f0cb1881e/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191125144606-a911d9008d1f/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191130070609-6e064ea0cf2d/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
+golang.org/x/tools v0.0.0-20191216173652-a0e659d51361/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20191227053925-7b8e75db28f4/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200117161641-43d50277825c/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200122220014-bf1340f18c4a/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200130002326-2f3ba24bd6e7/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200204074204-1cc6d1ef6c74/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200207183749-b753a1ba74fa/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200212150539-ea181f53ac56/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200224181240-023911ca70b2/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200227222343-706bc42d1f0d/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
+golang.org/x/tools v0.0.0-20200304193943-95d2e580d8eb/go.mod h1:o4KQGtdN14AW+yjsvvwRTJJuXz8XRtIHtEnmAXLyFUw=
+golang.org/x/tools v0.0.0-20200312045724-11d5b4c81c7d/go.mod h1:o4KQGtdN14AW+yjsvvwRTJJuXz8XRtIHtEnmAXLyFUw=
+golang.org/x/tools v0.0.0-20200331025713-a30bf2db82d4/go.mod h1:Sl4aGygMT6LrqrWclx+PTx3U+LnKx/seiNR+3G19Ar8=
+golang.org/x/tools v0.0.0-20200501065659-ab2804fb9c9d/go.mod h1:EkVYQZoAsY45+roYkvgYkIh4xh/qjgUK9TdY2XT94GE=
+golang.org/x/tools v0.0.0-20200512131952-2bc93b1c0c88/go.mod h1:EkVYQZoAsY45+roYkvgYkIh4xh/qjgUK9TdY2XT94GE=
+golang.org/x/tools v0.0.0-20200515010526-7d3b6ebf133d/go.mod h1:EkVYQZoAsY45+roYkvgYkIh4xh/qjgUK9TdY2XT94GE=
+golang.org/x/tools v0.0.0-20200618134242-20370b0cb4b2/go.mod h1:EkVYQZoAsY45+roYkvgYkIh4xh/qjgUK9TdY2XT94GE=
+golang.org/x/tools v0.0.0-20200729194436-6467de6f59a7/go.mod h1:njjCfa9FT2d7l9Bc6FUM5FLjQPp3cFF28FI3qnDFljA=
+golang.org/x/tools v0.0.0-20200804011535-6c149bb5ef0d/go.mod h1:njjCfa9FT2d7l9Bc6FUM5FLjQPp3cFF28FI3qnDFljA=
+golang.org/x/tools v0.0.0-20200825202427-b303f430e36d/go.mod h1:njjCfa9FT2d7l9Bc6FUM5FLjQPp3cFF28FI3qnDFljA=
+golang.org/x/tools v0.0.0-20200904185747-39188db58858/go.mod h1:Cj7w3i3Rnn0Xh82ur9kSqwfTHTeVxaDqrfMjpcNT6bE=
+golang.org/x/tools v0.0.0-20201110124207-079ba7bd75cd/go.mod h1:emZCQorbCU4vsT4fOWvOPXz4eW1wZW4PmDk9uLelYpA=
+golang.org/x/tools v0.0.0-20201201161351-ac6f37ff4c2a/go.mod h1:emZCQorbCU4vsT4fOWvOPXz4eW1wZW4PmDk9uLelYpA=
+golang.org/x/tools v0.0.0-20201208233053-a543418bbed2/go.mod h1:emZCQorbCU4vsT4fOWvOPXz4eW1wZW4PmDk9uLelYpA=
+golang.org/x/tools v0.0.0-20210105154028-b0ab187a4818/go.mod h1:emZCQorbCU4vsT4fOWvOPXz4eW1wZW4PmDk9uLelYpA=
+golang.org/x/tools v0.0.0-20210108195828-e2f9c7f1fc8e/go.mod h1:emZCQorbCU4vsT4fOWvOPXz4eW1wZW4PmDk9uLelYpA=
+golang.org/x/tools v0.1.0/go.mod h1:xkSsbof2nBLbhDlRMhhhyNLN/zl3eTqcnHD5viDpcZ0=
+golang.org/x/xerrors v0.0.0-20190717185122-a985d3407aa7/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
+golang.org/x/xerrors v0.0.0-20191011141410-1b5146add898/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
+golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
+golang.org/x/xerrors v0.0.0-20200804184101-5ec99f83aff1/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
+google.golang.org/api v0.4.0/go.mod h1:8k5glujaEP+g9n7WNsDg8QP6cUVNI86fCNMcbazEtwE=
+google.golang.org/api v0.7.0/go.mod h1:WtwebWUNSVBH/HAw79HIFXZNqEvBhG+Ra+ax0hx3E3M=
+google.golang.org/api v0.8.0/go.mod h1:o4eAsZoiT+ibD93RtjEohWalFOjRDx6CVaqeizhEnKg=
+google.golang.org/api v0.9.0/go.mod h1:o4eAsZoiT+ibD93RtjEohWalFOjRDx6CVaqeizhEnKg=
+google.golang.org/api v0.13.0/go.mod h1:iLdEw5Ide6rF15KTC1Kkl0iskquN2gFfn9o9XIsbkAI=
+google.golang.org/api v0.14.0/go.mod h1:iLdEw5Ide6rF15KTC1Kkl0iskquN2gFfn9o9XIsbkAI=
+google.golang.org/api v0.15.0/go.mod h1:iLdEw5Ide6rF15KTC1Kkl0iskquN2gFfn9o9XIsbkAI=
+google.golang.org/api v0.17.0/go.mod h1:BwFmGc8tA3vsd7r/7kR8DY7iEEGSU04BFxCo5jP/sfE=
+google.golang.org/api v0.18.0/go.mod h1:BwFmGc8tA3vsd7r/7kR8DY7iEEGSU04BFxCo5jP/sfE=
+google.golang.org/api v0.19.0/go.mod h1:BwFmGc8tA3vsd7r/7kR8DY7iEEGSU04BFxCo5jP/sfE=
+google.golang.org/api v0.20.0/go.mod h1:BwFmGc8tA3vsd7r/7kR8DY7iEEGSU04BFxCo5jP/sfE=
+google.golang.org/api v0.22.0/go.mod h1:BwFmGc8tA3vsd7r/7kR8DY7iEEGSU04BFxCo5jP/sfE=
+google.golang.org/api v0.24.0/go.mod h1:lIXQywCXRcnZPGlsd8NbLnOjtAoL6em04bJ9+z0MncE=
+google.golang.org/api v0.28.0/go.mod h1:lIXQywCXRcnZPGlsd8NbLnOjtAoL6em04bJ9+z0MncE=
+google.golang.org/api v0.29.0/go.mod h1:Lcubydp8VUV7KeIHD9z2Bys/sm/vGKnG1UHuDBSrHWM=
+google.golang.org/api v0.30.0/go.mod h1:QGmEvQ87FHZNiUVJkT14jQNYJ4ZJjdRF23ZXz5138Fc=
+google.golang.org/api v0.35.0/go.mod h1:/XrVsuzM0rZmrsbjJutiuftIzeuTQcEeaYcSk/mQ1dg=
+google.golang.org/api v0.36.0/go.mod h1:+z5ficQTmoYpPn8LCUNVpK5I7hwkpjbcgqA7I34qYtE=
+google.golang.org/api v0.40.0/go.mod h1:fYKFpnQN0DsDSKRVRcQSDQNtqWPfM9i+zNPxepjRCQ8=
+google.golang.org/appengine v1.1.0/go.mod h1:EbEs0AVv82hx2wNQdGPgUI5lhzA/G0D9YwlJXL52JkM=
+google.golang.org/appengine v1.4.0/go.mod h1:xpcJRLb0r/rnEns0DIKYYv+WjYCduHsrkT7/EB5XEv4=
+google.golang.org/appengine v1.5.0/go.mod h1:xpcJRLb0r/rnEns0DIKYYv+WjYCduHsrkT7/EB5XEv4=
+google.golang.org/appengine v1.6.1/go.mod h1:i06prIuMbXzDqacNJfV5OdTW448YApPu5ww/cMBSeb0=
+google.golang.org/appengine v1.6.5/go.mod h1:8WjMMxjGQR8xUklV/ARdw2HLXBOI7O7uCIDZVag1xfc=
+google.golang.org/appengine v1.6.6/go.mod h1:8WjMMxjGQR8xUklV/ARdw2HLXBOI7O7uCIDZVag1xfc=
+google.golang.org/appengine v1.6.7/go.mod h1:8WjMMxjGQR8xUklV/ARdw2HLXBOI7O7uCIDZVag1xfc=
+google.golang.org/genproto v0.0.0-20180817151627-c66870c02cf8/go.mod h1:JiN7NxoALGmiZfu7CAH4rXhgtRTLTxftemlI0sWmxmc=
+google.golang.org/genproto v0.0.0-20190307195333-5fe7a883aa19/go.mod h1:VzzqZJRnGkLBvHegQrXjBqPurQTc5/KpmUdxsrq26oE=
+google.golang.org/genproto v0.0.0-20190418145605-e7d98fc518a7/go.mod h1:VzzqZJRnGkLBvHegQrXjBqPurQTc5/KpmUdxsrq26oE=
+google.golang.org/genproto v0.0.0-20190425155659-357c62f0e4bb/go.mod h1:VzzqZJRnGkLBvHegQrXjBqPurQTc5/KpmUdxsrq26oE=
+google.golang.org/genproto v0.0.0-20190502173448-54afdca5d873/go.mod h1:VzzqZJRnGkLBvHegQrXjBqPurQTc5/KpmUdxsrq26oE=
+google.golang.org/genproto v0.0.0-20190801165951-fa694d86fc64/go.mod h1:DMBHOl98Agz4BDEuKkezgsaosCRResVns1a3J2ZsMNc=
+google.golang.org/genproto v0.0.0-20190819201941-24fa4b261c55/go.mod h1:DMBHOl98Agz4BDEuKkezgsaosCRResVns1a3J2ZsMNc=
+google.golang.org/genproto v0.0.0-20190911173649-1774047e7e51/go.mod h1:IbNlFCBrqXvoKpeg0TB2l7cyZUmoaFKYIwrEpbDKLA8=
+google.golang.org/genproto v0.0.0-20191108220845-16a3f7862a1a/go.mod h1:n3cpQtvxv34hfy77yVDNjmbRyujviMdxYliBSkLhpCc=
+google.golang.org/genproto v0.0.0-20191115194625-c23dd37a84c9/go.mod h1:n3cpQtvxv34hfy77yVDNjmbRyujviMdxYliBSkLhpCc=
+google.golang.org/genproto v0.0.0-20191216164720-4f79533eabd1/go.mod h1:n3cpQtvxv34hfy77yVDNjmbRyujviMdxYliBSkLhpCc=
+google.golang.org/genproto v0.0.0-20191230161307-f3c370f40bfb/go.mod h1:n3cpQtvxv34hfy77yVDNjmbRyujviMdxYliBSkLhpCc=
+google.golang.org/genproto v0.0.0-20200115191322-ca5a22157cba/go.mod h1:n3cpQtvxv34hfy77yVDNjmbRyujviMdxYliBSkLhpCc=
+google.golang.org/genproto v0.0.0-20200122232147-0452cf42e150/go.mod h1:n3cpQtvxv34hfy77yVDNjmbRyujviMdxYliBSkLhpCc=
+google.golang.org/genproto v0.0.0-20200204135345-fa8e72b47b90/go.mod h1:GmwEX6Z4W5gMy59cAlVYjN9JhxgbQH6Gn+gFDQe2lzA=
+google.golang.org/genproto v0.0.0-20200212174721-66ed5ce911ce/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200224152610-e50cd9704f63/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200228133532-8c2c7df3a383/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200305110556-506484158171/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200312145019-da6875a35672/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200331122359-1ee6d9798940/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200430143042-b979b6f78d84/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200511104702-f5ebc3bea380/go.mod h1:55QSHmfGQM9UVYDPBsyGGes0y52j32PQ3BqQfXhyH3c=
+google.golang.org/genproto v0.0.0-20200515170657-fc4c6c6a6587/go.mod h1:YsZOwe1myG/8QRHRsmBRE1LrgQY60beZKjly0O1fX9U=
+google.golang.org/genproto v0.0.0-20200526211855-cb27e3aa2013/go.mod h1:NbSheEEYHJ7i3ixzK3sjbqSGDJWnxyFXZblF3eUsNvo=
+google.golang.org/genproto v0.0.0-20200618031413-b414f8b61790/go.mod h1:jDfRM7FcilCzHH/e9qn6dsT145K34l5v+OpcnNgKAAA=
+google.golang.org/genproto v0.0.0-20200729003335-053ba62fc06f/go.mod h1:FWY/as6DDZQgahTzZj3fqbO1CbirC29ZNUFHwi0/+no=
+google.golang.org/genproto v0.0.0-20200804131852-c06518451d9c/go.mod h1:FWY/as6DDZQgahTzZj3fqbO1CbirC29ZNUFHwi0/+no=
+google.golang.org/genproto v0.0.0-20200825200019-8632dd797987/go.mod h1:FWY/as6DDZQgahTzZj3fqbO1CbirC29ZNUFHwi0/+no=
+google.golang.org/genproto v0.0.0-20200904004341-0bd0a958aa1d/go.mod h1:FWY/as6DDZQgahTzZj3fqbO1CbirC29ZNUFHwi0/+no=
+google.golang.org/genproto v0.0.0-20201109203340-2640f1f9cdfb/go.mod h1:FWY/as6DDZQgahTzZj3fqbO1CbirC29ZNUFHwi0/+no=
+google.golang.org/genproto v0.0.0-20201201144952-b05cb90ed32e/go.mod h1:FWY/as6DDZQgahTzZj3fqbO1CbirC29ZNUFHwi0/+no=
+google.golang.org/genproto v0.0.0-20201210142538-e3217bee35cc/go.mod h1:FWY/as6DDZQgahTzZj3fqbO1CbirC29ZNUFHwi0/+no=
+google.golang.org/genproto v0.0.0-20201214200347-8c77b98c765d/go.mod h1:FWY/as6DDZQgahTzZj3fqbO1CbirC29ZNUFHwi0/+no=
+google.golang.org/genproto v0.0.0-20210108203827-ffc7fda8c3d7/go.mod h1:FWY/as6DDZQgahTzZj3fqbO1CbirC29ZNUFHwi0/+no=
+google.golang.org/genproto v0.0.0-20210226172003-ab064af71705/go.mod h1:FWY/as6DDZQgahTzZj3fqbO1CbirC29ZNUFHwi0/+no=
+google.golang.org/grpc v1.19.0/go.mod h1:mqu4LbDTu4XGKhr4mRzUsmM4RtVoemTSY81AxZiDr8c=
+google.golang.org/grpc v1.20.1/go.mod h1:10oTOabMzJvdu6/UiuZezV6QK5dSlG84ov/aaiqXj38=
+google.golang.org/grpc v1.21.1/go.mod h1:oYelfM1adQP15Ek0mdvEgi9Df8B9CZIaU1084ijfRaM=
+google.golang.org/grpc v1.23.0/go.mod h1:Y5yQAOtifL1yxbo5wqy6BxZv8vAUGQwXBOALyacEbxg=
+google.golang.org/grpc v1.25.1/go.mod h1:c3i+UQWmh7LiEpx4sFZnkU36qjEYZ0imhYfXVyQciAY=
+google.golang.org/grpc v1.26.0/go.mod h1:qbnxyOmOxrQa7FizSgH+ReBfzJrCY1pSN7KXBS8abTk=
+google.golang.org/grpc v1.27.0/go.mod h1:qbnxyOmOxrQa7FizSgH+ReBfzJrCY1pSN7KXBS8abTk=
+google.golang.org/grpc v1.27.1/go.mod h1:qbnxyOmOxrQa7FizSgH+ReBfzJrCY1pSN7KXBS8abTk=
+google.golang.org/grpc v1.28.0/go.mod h1:rpkK4SK4GF4Ach/+MFLZUBavHOvF2JJB5uozKKal+60=
+google.golang.org/grpc v1.29.1/go.mod h1:itym6AZVZYACWQqET3MqgPpjcuV5QH3BxFS3IjizoKk=
+google.golang.org/grpc v1.30.0/go.mod h1:N36X2cJ7JwdamYAgDz+s+rVMFjt3numwzf/HckM8pak=
+google.golang.org/grpc v1.31.0/go.mod h1:N36X2cJ7JwdamYAgDz+s+rVMFjt3numwzf/HckM8pak=
+google.golang.org/grpc v1.31.1/go.mod h1:N36X2cJ7JwdamYAgDz+s+rVMFjt3numwzf/HckM8pak=
+google.golang.org/grpc v1.33.2/go.mod h1:JMHMWHQWaTccqQQlmk3MJZS+GWXOdAesneDmEnv2fbc=
+google.golang.org/grpc v1.34.0/go.mod h1:WotjhfgOW/POjDeRt8vscBtXq+2VjORFy659qA51WJ8=
+google.golang.org/grpc v1.35.0/go.mod h1:qjiiYl8FncCW8feJPdyg3v6XW24KsRHe+dy9BAGRRjU=
+google.golang.org/protobuf v0.0.0-20200109180630-ec00e32a8dfd/go.mod h1:DFci5gLYBciE7Vtevhsrf46CRTquxDuWsQurQQe4oz8=
+google.golang.org/protobuf v0.0.0-20200221191635-4d8936d0db64/go.mod h1:kwYJMbMJ01Woi6D6+Kah6886xMZcty6N08ah7+eCXa0=
+google.golang.org/protobuf v0.0.0-20200228230310-ab0ca4ff8a60/go.mod h1:cfTl7dwQJ+fmap5saPgwCLgHXTUD7jkjRqWcaiX5VyM=
+google.golang.org/protobuf v1.20.1-0.20200309200217-e05f789c0967/go.mod h1:A+miEFZTKqfCUM6K7xSMQL9OKL/b6hQv+e19PK+JZNE=
+google.golang.org/protobuf v1.21.0/go.mod h1:47Nbq4nVaFHyn7ilMalzfO3qCViNmqZ2kzikPIcrTAo=
+google.golang.org/protobuf v1.22.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
+google.golang.org/protobuf v1.23.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
+google.golang.org/protobuf v1.23.1-0.20200526195155-81db48ad09cc/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
+google.golang.org/protobuf v1.24.0/go.mod h1:r/3tXBNzIEhYS9I1OUVjXDlt8tc493IdKGjtUeSXeh4=
+google.golang.org/protobuf v1.25.0/go.mod h1:9JNX74DMeImyA3h4bdi1ymwjUzf21/xIlbajtzgsN7c=
+gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405 h1:yhCVgyC4o1eVCa2tZl7eS0r+SDo693bJlVdllGtEeKM=
+gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
+gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
+gopkg.in/errgo.v2 v2.1.0/go.mod h1:hNsd1EY+bozCKY1Ytp96fpM3vjJbqLJn88ws8XvfDNI=
+gopkg.in/yaml.v2 v2.2.2/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
+gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
+gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
+gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
+honnef.co/go/tools v0.0.0-20190102054323-c2f93a96b099/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
+honnef.co/go/tools v0.0.0-20190106161140-3f1c8253044a/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
+honnef.co/go/tools v0.0.0-20190418001031-e561f6794a2a/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
+honnef.co/go/tools v0.0.0-20190523083050-ea95bdfd59fc/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
+honnef.co/go/tools v0.0.1-2019.2.3/go.mod h1:a3bituU0lyd329TUQxRnasdCoJDkEUEAqEt0JzvZhAg=
+honnef.co/go/tools v0.0.1-2020.1.3/go.mod h1:X/FiERA/W4tHapMX5mGpAtMSVEeEUOyHaw9vFzvIQ3k=
+honnef.co/go/tools v0.0.1-2020.1.4/go.mod h1:X/FiERA/W4tHapMX5mGpAtMSVEeEUOyHaw9vFzvIQ3k=
+rsc.io/binaryregexp v0.2.0/go.mod h1:qTv7/COck+e2FymRvadv62gMdZztPaShugOCi3I+8D8=
+rsc.io/quote/v3 v3.1.0/go.mod h1:yEA65RcK8LyAZtP9Kv3t0HmxON59tX3rD+tICJqUlj0=
+rsc.io/sampler v1.3.0/go.mod h1:T1hPZKmBbMNahiBKFy5HrXp6adAjACjK9JXDnKaTXpA=
+rsc.io/tmplfunc v0.0.3 h1:53XFQh69AfOa8Tw0Jm7t+GV7KZhOi6jzsCzTtKbMvzU=
+rsc.io/tmplfunc v0.0.3/go.mod h1:AG3sTPzElb1Io3Yg4voV9AGZJuleGAwaVRxL9M49PhA=

goicicle/Makefile

@@ -0,0 +1,29 @@
+CUDA_ROOT_DIR = /usr/local/cuda
+NVCC = $(CUDA_ROOT_DIR)/bin/nvcc
+CFLAGS = -Xcompiler -fPIC -std=c++17
+LDFLAGS = -shared
+FEATURES = -DG2_DEFINED
+
+TARGET_BN254 = libbn254.so
+TARGET_BLS12_381 = libbls12_381.so
+TARGET_BLS12_377 = libbls12_377.so
+
+VPATH = ../icicle/curves/bn254:../icicle/curves/bls12_377:../icicle/curves/bls12_381
+
+SRCS_BN254 = lde.cu msm.cu projective.cu ve_mod_mult.cu
+SRCS_BLS12_381 = lde.cu msm.cu projective.cu ve_mod_mult.cu poseidon.cu
+SRCS_BLS12_377 = lde.cu msm.cu projective.cu ve_mod_mult.cu
+
+all: $(TARGET_BN254) $(TARGET_BLS12_381) $(TARGET_BLS12_377)
+
+$(TARGET_BN254): 
+	$(NVCC) $(FEATURES) $(CFLAGS) $(LDFLAGS) $(addprefix ../icicle/curves/bn254/, $(SRCS_BN254)) -o $@
+
+$(TARGET_BLS12_381):
+	$(NVCC) $(FEATURES) $(CFLAGS) $(LDFLAGS) $(addprefix ../icicle/curves/bls12_381/, $(SRCS_BLS12_381)) -o $@
+
+$(TARGET_BLS12_377):
+	$(NVCC) $(FEATURES) $(CFLAGS) $(LDFLAGS) $(addprefix ../icicle/curves/bls12_377/, $(SRCS_BLS12_377)) -o $@
+
+clean:
+	rm -f $(TARGET_BN254) $(TARGET_BLS12_381) $(TARGET_BLS12_377)

goicicle/README.md

@@ -0,0 +1,49 @@
+# ICICLE CUDA to Golang Binding Guide
+
+This guide provides instructions on how to compile CUDA code using the provided Makefile, and then how to use the resulting shared libraries to bind Golang to ICICLE's CUDA code.
+
+## Prerequisites
+
+To compile the CUDA files, you will need:
+
+- CUDA toolkit installed. The Makefile assumes CUDA is installed in `/usr/local/cuda`. If CUDA is installed in a different location, please adjust the `CUDA_ROOT_DIR` variable accordingly.
+- A compatible GPU and corresponding driver installed on your machine.
+
+## Structure of the Makefile
+
+The Makefile is designed to compile CUDA files for three curves: BN254, BLS12_381, and BLS12_377. The source files are located in the `icicle/curves/` directory.
+
+## Compiling CUDA Code
+
+1. Navigate to the directory containing the Makefile in your terminal.
+2. To compile all curve libraries, use the `make all` command. This will create three shared libraries: `libbn254.so`, `libbls12_381.so`, and `libbls12_377.so`.
+3. If you want to compile a specific curve, you can do so by specifying the target. For example, to compile only the BN254 curve, use `make libbn254.so`. Replace `libbn254.so` with `libbls12_381.so` or `libbls12_377.so` to compile those curves instead.
+
+The resulting `.so` files are the compiled shared libraries for each curve.
+
+## Golang Binding
+
+The shared libraries produced from the CUDA code compilation are used to bind Golang to ICICLE's CUDA code.
+
+1. These shared libraries (`libbn254.so`, `libbls12_381.so`, `libbls12_377.so`) can be imported in your Go project to leverage the GPU accelerated functionalities provided by ICICLE. 
+
+2. In your Go project, you can use `cgo` to link these shared libraries. Here's a basic example on how you can use `cgo` to link these libraries:
+
+```go
+/*
+#cgo LDFLAGS: -L/path/to/shared/libs -lbn254 -lbls12_381 -lbls12_377
+#include "icicle.h" // make sure you use the correct header file(s)
+*/
+import "C"
+
+func main() {
+    // Now you can call the C functions from the ICICLE libraries.
+    // Note that C function calls are prefixed with 'C.' in Go code.
+}
+```
+
+Replace `/path/to/shared/libs` with the actual path where the shared libraries are located on your system.
+
+## Cleaning up
+
+If you want to remove the compiled files, you can use the `make clean` command. This will remove the `libbn254.so`, `libbls12_381.so`, and `libbls12_377.so` files.

goicicle/curves/bls12377/g1.go

@@ -0,0 +1,324 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12377
+
+import (
+	"unsafe"
+
+	"encoding/binary"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbls12_377
+// #include "projective.h"
+// #include "ve_mod_mult.h"
+import "C"
+
+const SCALAR_SIZE = 8
+const BASE_SIZE = 12
+
+type G1ScalarField struct {
+	S [SCALAR_SIZE]uint32
+}
+
+type G1BaseField struct {
+	S [BASE_SIZE]uint32
+}
+
+/*
+ * BaseField Constrctors
+ */
+
+func (f *G1BaseField) SetZero() *G1BaseField {
+	var S [BASE_SIZE]uint32
+	f.S = S
+
+	return f
+}
+
+func (f *G1BaseField) SetOne() *G1BaseField {
+	var S [BASE_SIZE]uint32
+
+	S[0] = 1
+
+	f.S = S
+	return f
+}
+
+func (p *G1ProjectivePoint) FromAffine(affine *G1PointAffine) *G1ProjectivePoint {
+	out := (*C.BLS12_377_projective_t)(unsafe.Pointer(p))
+	in := (*C.BLS12_377_affine_t)(unsafe.Pointer(affine))
+
+	C.projective_from_affine_bls12_377(out, in)
+
+	return p
+}
+
+func (f *G1BaseField) FromLimbs(limbs [BASE_SIZE]uint32) *G1BaseField {
+	copy(f.S[:], limbs[:])
+
+	return f
+}
+
+/*
+ * BaseField methods
+ */
+
+func (f *G1BaseField) Limbs() [BASE_SIZE]uint32 {
+	return f.S
+}
+
+func (f *G1BaseField) ToBytesLe() []byte {
+	bytes := make([]byte, len(f.S)*4)
+	for i, v := range f.S {
+		binary.LittleEndian.PutUint32(bytes[i*4:], v)
+	}
+
+	return bytes
+}
+
+/*
+ * ScalarField methods
+ */
+
+func (p *G1ScalarField) Random() *G1ScalarField {
+	outC := (*C.BLS12_377_scalar_t)(unsafe.Pointer(p))
+	C.random_scalar_bls12_377(outC)
+
+	return p
+}
+
+func (f *G1ScalarField) SetZero() *G1ScalarField {
+	var S [SCALAR_SIZE]uint32
+	f.S = S
+
+	return f
+}
+
+func (f *G1ScalarField) SetOne() *G1ScalarField {
+	var S [SCALAR_SIZE]uint32
+	S[0] = 1
+	f.S = S
+
+	return f
+}
+
+func (a *G1ScalarField) Eq(b *G1ScalarField) bool {
+	for i, v := range a.S {
+		if b.S[i] != v {
+			return false
+		}
+	}
+	return true
+}
+
+/*
+ * ScalarField methods
+ */
+
+func (f *G1ScalarField) Limbs() [SCALAR_SIZE]uint32 {
+	return f.S
+}
+
+func (f *G1ScalarField) ToBytesLe() []byte {
+	bytes := make([]byte, len(f.S)*4)
+	for i, v := range f.S {
+		binary.LittleEndian.PutUint32(bytes[i*4:], v)
+	}
+
+	return bytes
+}
+
+/*
+ * PointBLS12_377
+ */
+
+type G1ProjectivePoint struct {
+	X, Y, Z G1BaseField
+}
+
+func (f *G1ProjectivePoint) SetZero() *G1ProjectivePoint {
+	var yOne G1BaseField
+	yOne.SetOne()
+
+	var xZero G1BaseField
+	xZero.SetZero()
+
+	var zZero G1BaseField
+	zZero.SetZero()
+
+	f.X = xZero
+	f.Y = yOne
+	f.Z = zZero
+
+	return f
+}
+
+func (p *G1ProjectivePoint) Eq(pCompare *G1ProjectivePoint) bool {
+	// Cast *PointBLS12_377 to *C.BLS12_377_projective_t
+	// The unsafe.Pointer cast is necessary because Go doesn't allow direct casts
+	// between different pointer types.
+	// It'S your responsibility to ensure that the types are compatible.
+	pC := (*C.BLS12_377_projective_t)(unsafe.Pointer(p))
+	pCompareC := (*C.BLS12_377_projective_t)(unsafe.Pointer(pCompare))
+
+	// Call the C function
+	// The C function doesn't keep any references to the data,
+	// so it'S fine if the Go garbage collector moves or deletes the data later.
+	return bool(C.eq_bls12_377(pC, pCompareC))
+}
+
+func (p *G1ProjectivePoint) IsOnCurve() bool {
+	point := (*C.BLS12_377_projective_t)(unsafe.Pointer(p))
+	res := C.projective_is_on_curve_bls12_377(point)
+
+	return bool(res)
+}
+
+func (p *G1ProjectivePoint) Random() *G1ProjectivePoint {
+	outC := (*C.BLS12_377_projective_t)(unsafe.Pointer(p))
+	C.random_projective_bls12_377(outC)
+
+	return p
+}
+
+func (p *G1ProjectivePoint) StripZ() *G1PointAffine {
+	return &G1PointAffine{
+		X: p.X,
+		Y: p.Y,
+	}
+}
+
+func (p *G1ProjectivePoint) FromLimbs(x, y, z *[]uint32) *G1ProjectivePoint {
+	var _x G1BaseField
+	var _y G1BaseField
+	var _z G1BaseField
+
+	_x.FromLimbs(GetFixedLimbs(x))
+	_y.FromLimbs(GetFixedLimbs(y))
+	_z.FromLimbs(GetFixedLimbs(z))
+
+	p.X = _x
+	p.Y = _y
+	p.Z = _z
+
+	return p
+}
+
+/*
+ * PointAffineNoInfinityBLS12_377
+ */
+
+type G1PointAffine struct {
+	X, Y G1BaseField
+}
+
+func (p *G1PointAffine) FromProjective(projective *G1ProjectivePoint) *G1PointAffine {
+	in := (*C.BLS12_377_projective_t)(unsafe.Pointer(projective))
+	out := (*C.BLS12_377_affine_t)(unsafe.Pointer(p))
+
+	C.projective_to_affine_bls12_377(out, in)
+
+	return p
+}
+
+func (p *G1PointAffine) ToProjective() *G1ProjectivePoint {
+	var Z G1BaseField
+	Z.SetOne()
+
+	return &G1ProjectivePoint{
+		X: p.X,
+		Y: p.Y,
+		Z: Z,
+	}
+}
+
+func (p *G1PointAffine) FromLimbs(X, Y *[]uint32) *G1PointAffine {
+	var _x G1BaseField
+	var _y G1BaseField
+
+	_x.FromLimbs(GetFixedLimbs(X))
+	_y.FromLimbs(GetFixedLimbs(Y))
+
+	return p
+}
+
+/*
+ * Multiplication
+ */
+
+func MultiplyVec(a []G1ProjectivePoint, b []G1ScalarField, deviceID int) {
+	if len(a) != len(b) {
+		panic("a and b have different lengths")
+	}
+
+	pointsC := (*C.BLS12_377_projective_t)(unsafe.Pointer(&a[0]))
+	scalarsC := (*C.BLS12_377_scalar_t)(unsafe.Pointer(&b[0]))
+	deviceIdC := C.size_t(deviceID)
+	nElementsC := C.size_t(len(a))
+
+	C.vec_mod_mult_point_bls12_377(pointsC, scalarsC, nElementsC, deviceIdC)
+}
+
+func MultiplyScalar(a []G1ScalarField, b []G1ScalarField, deviceID int) {
+	if len(a) != len(b) {
+		panic("a and b have different lengths")
+	}
+
+	aC := (*C.BLS12_377_scalar_t)(unsafe.Pointer(&a[0]))
+	bC := (*C.BLS12_377_scalar_t)(unsafe.Pointer(&b[0]))
+	deviceIdC := C.size_t(deviceID)
+	nElementsC := C.size_t(len(a))
+
+	C.vec_mod_mult_scalar_bls12_377(aC, bC, nElementsC, deviceIdC)
+}
+
+// Multiply a matrix by a scalar:
+//
+//	`a` - flattenned matrix;
+//	`b` - vector to multiply `a` by;
+func MultiplyMatrix(a []G1ScalarField, b []G1ScalarField, deviceID int) {
+	c := make([]G1ScalarField, len(b))
+	for i := range c {
+		var p G1ScalarField
+		p.SetZero()
+
+		c[i] = p
+	}
+
+	aC := (*C.BLS12_377_scalar_t)(unsafe.Pointer(&a[0]))
+	bC := (*C.BLS12_377_scalar_t)(unsafe.Pointer(&b[0]))
+	cC := (*C.BLS12_377_scalar_t)(unsafe.Pointer(&c[0]))
+	deviceIdC := C.size_t(deviceID)
+	nElementsC := C.size_t(len(a))
+
+	C.matrix_vec_mod_mult_bls12_377(aC, bC, cC, nElementsC, deviceIdC)
+}
+
+/*
+ * Utils
+ */
+
+func GetFixedLimbs(slice *[]uint32) [BASE_SIZE]uint32 {
+	if len(*slice) <= BASE_SIZE {
+		limbs := [BASE_SIZE]uint32{}
+		copy(limbs[:len(*slice)], *slice)
+		return limbs
+	}
+
+	panic("slice has too many elements")
+}

goicicle/curves/bls12377/g1_test.go

@@ -0,0 +1,198 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12377
+
+import (
+	"encoding/binary"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestNewFieldBLS12_377One(t *testing.T) {
+	var oneField G1BaseField
+	oneField.SetOne()
+
+	rawOneField := [8]uint32([8]uint32{0x1, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0})
+
+	assert.Equal(t, oneField.S, rawOneField)
+}
+
+func TestNewFieldBLS12_377Zero(t *testing.T) {
+	var zeroField G1BaseField
+	zeroField.SetZero()
+
+	rawZeroField := [8]uint32([8]uint32{0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0})
+
+	assert.Equal(t, zeroField.S, rawZeroField)
+}
+
+func TestFieldBLS12_377ToBytesLe(t *testing.T) {
+	var p G1ProjectivePoint
+	p.Random()
+
+	expected := make([]byte, len(p.X.S)*4) // each uint32 takes 4 bytes
+	for i, v := range p.X.S {
+		binary.LittleEndian.PutUint32(expected[i*4:], v)
+	}
+
+	assert.Equal(t, p.X.ToBytesLe(), expected)
+	assert.Equal(t, len(p.X.ToBytesLe()), 32)
+}
+
+func TestNewPointBLS12_377Zero(t *testing.T) {
+	var pointZero G1ProjectivePoint
+	pointZero.SetZero()
+
+	var baseOne G1BaseField
+	baseOne.SetOne()
+
+	var zeroSanity G1BaseField
+	zeroSanity.SetZero()
+
+	assert.Equal(t, pointZero.X, zeroSanity)
+	assert.Equal(t, pointZero.Y, baseOne)
+	assert.Equal(t, pointZero.Z, zeroSanity)
+}
+
+func TestFromProjectiveToAffine(t *testing.T) {
+	var projective G1ProjectivePoint
+	var affine G1PointAffine
+
+	projective.Random()
+
+	affine.FromProjective(&projective)
+	var projective2 G1ProjectivePoint
+	projective2.FromAffine(&affine)
+
+	assert.True(t, projective.IsOnCurve())
+	assert.True(t, projective2.IsOnCurve())
+	assert.True(t, projective.Eq(&projective2))
+}
+
+func TestBLS12_377Eq(t *testing.T) {
+	var p1 G1ProjectivePoint
+	p1.Random()
+	var p2 G1ProjectivePoint
+	p2.Random()
+
+	assert.Equal(t, p1.Eq(&p1), true)
+	assert.Equal(t, p1.Eq(&p2), false)
+}
+
+func TestBLS12_377StripZ(t *testing.T) {
+	var p1 G1ProjectivePoint
+	p1.Random()
+
+	p2ZLess := p1.StripZ()
+
+	assert.IsType(t, G1PointAffine{}, *p2ZLess)
+	assert.Equal(t, p1.X, p2ZLess.X)
+	assert.Equal(t, p1.Y, p2ZLess.Y)
+}
+
+func TestPointBLS12_377fromLimbs(t *testing.T) {
+	var p G1ProjectivePoint
+	p.Random()
+
+	x := p.X.Limbs()
+	y := p.Y.Limbs()
+	z := p.Z.Limbs()
+
+	xSlice := x[:]
+	ySlice := y[:]
+	zSlice := z[:]
+
+	var pFromLimbs G1ProjectivePoint
+	pFromLimbs.FromLimbs(&xSlice, &ySlice, &zSlice)
+
+	assert.Equal(t, pFromLimbs, p)
+}
+
+func TestNewPointAffineNoInfinityBLS12_377Zero(t *testing.T) {
+	var zeroP G1PointAffine
+
+	var zeroSanity G1BaseField
+	zeroSanity.SetZero()
+
+	assert.Equal(t, zeroP.X, zeroSanity)
+	assert.Equal(t, zeroP.Y, zeroSanity)
+}
+
+func TestPointAffineNoInfinityBLS12_377FromLimbs(t *testing.T) {
+	// Initialize your test values
+	x := [8]uint32{1, 2, 3, 4, 5, 6, 7, 8}
+	y := [8]uint32{9, 10, 11, 12, 13, 14, 15, 16}
+	xSlice := x[:]
+	ySlice := y[:]
+
+	// Execute your function
+	var result G1PointAffine
+	result.FromLimbs(&xSlice, &ySlice)
+
+	var xBase G1BaseField
+	var yBase G1BaseField
+	xBase.FromLimbs(x)
+	yBase.FromLimbs(y)
+
+	// Define your expected result
+	expected := &G1PointAffine{
+		X: xBase,
+		Y: yBase,
+	}
+
+	// Test if result is as expected
+	assert.Equal(t, result, expected)
+}
+
+func TestGetFixedLimbs(t *testing.T) {
+	t.Run("case of valid input of length less than 8", func(t *testing.T) {
+		slice := []uint32{1, 2, 3, 4, 5, 6, 7}
+		expected := [8]uint32{1, 2, 3, 4, 5, 6, 7, 0}
+
+		result := GetFixedLimbs(&slice)
+		assert.Equal(t, result, expected)
+	})
+
+	t.Run("case of valid input of length 8", func(t *testing.T) {
+		slice := []uint32{1, 2, 3, 4, 5, 6, 7, 8}
+		expected := [8]uint32{1, 2, 3, 4, 5, 6, 7, 8}
+
+		result := GetFixedLimbs(&slice)
+		assert.Equal(t, result, expected)
+	})
+
+	t.Run("case of empty input", func(t *testing.T) {
+		slice := []uint32{}
+		expected := [8]uint32{0, 0, 0, 0, 0, 0, 0, 0}
+
+		result := GetFixedLimbs(&slice)
+		assert.Equal(t, result, expected)
+	})
+
+	t.Run("case of input length greater than 8", func(t *testing.T) {
+		slice := []uint32{1, 2, 3, 4, 5, 6, 7, 8, 9}
+
+		defer func() {
+			if r := recover(); r == nil {
+				t.Errorf("the code did not panic")
+			}
+		}()
+
+		GetFixedLimbs(&slice)
+	})
+}

goicicle/curves/bls12377/g2.go

@@ -0,0 +1,112 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12377
+
+import (
+	"encoding/binary"
+	"unsafe"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbls12_377
+// #include "projective.h"
+// #include "ve_mod_mult.h"
+import "C"
+
+// G2 extension field
+
+type G2Element [6]uint64
+
+type ExtentionField struct {
+	A0, A1 G2Element
+}
+
+type G2PointAffine struct {
+	X, Y ExtentionField
+}
+
+type G2Point struct {
+	X, Y, Z ExtentionField
+}
+
+func (p *G2Point) Random() *G2Point {
+	outC := (*C.BLS12_377_g2_projective_t)(unsafe.Pointer(p))
+	C.random_g2_projective_bls12_377(outC)
+
+	return p
+}
+
+func (p *G2Point) FromAffine(affine *G2PointAffine) *G2Point {
+	out := (*C.BLS12_377_g2_projective_t)(unsafe.Pointer(p))
+	in := (*C.BLS12_377_g2_affine_t)(unsafe.Pointer(affine))
+
+	C.g2_projective_from_affine_bls12_377(out, in)
+
+	return p
+}
+
+func (p *G2Point) Eq(pCompare *G2Point) bool {
+	// Cast *PointBLS12_377 to *C.BLS12_377_projective_t
+	// The unsafe.Pointer cast is necessary because Go doesn't allow direct casts
+	// between different pointer types.
+	// It's your responsibility to ensure that the types are compatible.
+	pC := (*C.BLS12_377_g2_projective_t)(unsafe.Pointer(p))
+	pCompareC := (*C.BLS12_377_g2_projective_t)(unsafe.Pointer(pCompare))
+
+	// Call the C function
+	// The C function doesn't keep any references to the data,
+	// so it's fine if the Go garbage collector moves or deletes the data later.
+	return bool(C.eq_g2_bls12_377(pC, pCompareC))
+}
+
+func (f *G2Element) ToBytesLe() []byte {
+	var bytes []byte
+	for _, val := range f {
+		buf := make([]byte, 8) // 8 bytes because uint64 is 64-bit
+		binary.LittleEndian.PutUint64(buf, val)
+		bytes = append(bytes, buf...)
+	}
+	return bytes
+}
+
+func (p *G2PointAffine) ToProjective() G2Point {
+	return G2Point{
+		X: p.X,
+		Y: p.Y,
+		Z: ExtentionField{
+			A0: G2Element{1, 0, 0, 0},
+			A1: G2Element{0, 0, 0, 0},
+		},
+	}
+}
+
+func (p *G2PointAffine) FromProjective(projective *G2Point) *G2PointAffine {
+	out := (*C.BLS12_377_g2_affine_t)(unsafe.Pointer(p))
+	in := (*C.BLS12_377_g2_projective_t)(unsafe.Pointer(projective))
+
+	C.g2_projective_to_affine_bls12_377(out, in)
+
+	return p
+}
+
+func (p *G2Point) IsOnCurve() bool {
+	// Directly copy memory from the C struct to the Go struct
+	point := (*C.BLS12_377_g2_projective_t)(unsafe.Pointer(p))
+	res := C.g2_projective_is_on_curve_bls12_377(point)
+
+	return bool(res)
+}

goicicle/curves/bls12377/g2_test.go

@@ -0,0 +1,75 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12377
+
+import (
+	"github.com/stretchr/testify/assert"
+	"testing"
+)
+
+func TestG2Eqg2(t *testing.T) {
+	var point G2Point
+
+	point.Random()
+
+	assert.True(t, point.Eq(&point))
+}
+
+func TestG2FromProjectiveToAffine(t *testing.T) {
+	var projective G2Point
+	var affine G2PointAffine
+
+	projective.Random()
+
+	affine.FromProjective(&projective)
+	var projective2 G2Point
+	projective2.FromAffine(&affine)
+
+	assert.True(t, projective.IsOnCurve())
+	assert.True(t, projective2.IsOnCurve())
+	assert.True(t, projective.Eq(&projective2))
+}
+
+func TestG2Eqg2NotEqual(t *testing.T) {
+	var point G2Point
+	point.Random()
+
+	var point2 G2Point
+	point2.Random()
+
+	assert.False(t, point.Eq(&point2))
+}
+
+func TestG2ToBytes(t *testing.T) {
+	element := G2Element{0x6546098ea84b6298, 0x4a384533d1f68aca, 0xaa0666972d771336, 0x1569e4a34321993}
+	bytes := element.ToBytesLe()
+
+	assert.Equal(t, bytes, []byte{0x98, 0x62, 0x4b, 0xa8, 0x8e, 0x9, 0x46, 0x65, 0xca, 0x8a, 0xf6, 0xd1, 0x33, 0x45, 0x38, 0x4a, 0x36, 0x13, 0x77, 0x2d, 0x97, 0x66, 0x6, 0xaa, 0x93, 0x19, 0x32, 0x34, 0x4a, 0x9e, 0x56, 0x1})
+}
+
+func TestG2ShouldConvertToProjective(t *testing.T) {
+	var pointProjective G2Point
+	var pointAffine G2PointAffine
+
+	pointProjective.Random()
+	pointAffine.FromProjective(&pointProjective)
+
+	proj := pointAffine.ToProjective()
+
+	assert.True(t, proj.IsOnCurve())
+	assert.True(t, pointProjective.Eq(&proj))
+}

goicicle/curves/bls12377/include/msm.h

@@ -0,0 +1,98 @@
+
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+#include <cuda.h>
+#include <cuda_runtime.h>
+#include <stdbool.h>
+// msm.h
+
+#ifndef _BLS12_377_MSM_H
+#define _BLS12_377_MSM_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+// Incomplete declaration of BLS12_377 projective and affine structs
+typedef struct BLS12_377_projective_t BLS12_377_projective_t;
+typedef struct BLS12_377_g2_projective_t BLS12_377_g2_projective_t;
+typedef struct BLS12_377_affine_t BLS12_377_affine_t;
+typedef struct BLS12_377_g2_affine_t BLS12_377_g2_affine_t;
+typedef struct BLS12_377_scalar_t BLS12_377_scalar_t;
+typedef cudaStream_t CudaStream_t;
+
+int msm_cuda_bls12_377(
+  BLS12_377_projective_t* out, BLS12_377_affine_t* points, BLS12_377_scalar_t* scalars, size_t count, size_t device_id);
+
+int msm_batch_cuda_bls12_377(
+  BLS12_377_projective_t* out,
+  BLS12_377_affine_t* points,
+  BLS12_377_scalar_t* scalars,
+  size_t batch_size,
+  size_t msm_size,
+  size_t device_id);
+
+int commit_cuda_bls12_377(
+  BLS12_377_projective_t* d_out,
+  BLS12_377_scalar_t* d_scalars,
+  BLS12_377_affine_t* d_points,
+  size_t count,
+  unsigned large_bucket_factor,
+  size_t device_id);
+
+int commit_batch_cuda_bls12_377(
+  BLS12_377_projective_t* d_out,
+  BLS12_377_scalar_t* d_scalars,
+  BLS12_377_affine_t* d_points,
+  size_t count,
+  size_t batch_size,
+  size_t device_id);
+
+int msm_g2_cuda_bls12_377(
+  BLS12_377_g2_projective_t* out,
+  BLS12_377_g2_affine_t* points,
+  BLS12_377_scalar_t* scalars,
+  size_t count,
+  size_t device_id);
+int msm_batch_g2_cuda_bls12_377(
+  BLS12_377_g2_projective_t* out,
+  BLS12_377_g2_affine_t* points,
+  BLS12_377_scalar_t* scalars,
+  size_t batch_size,
+  size_t msm_size,
+  size_t device_id);
+int commit_g2_cuda_bls12_377(
+  BLS12_377_g2_projective_t* d_out,
+  BLS12_377_scalar_t* d_scalars,
+  BLS12_377_g2_affine_t* d_points,
+  size_t count,
+  unsigned large_bucket_factor,
+  size_t device_id);
+int commit_batch_g2_cuda_bls12_377(
+  BLS12_377_g2_projective_t* d_out,
+  BLS12_377_scalar_t* d_scalars,
+  BLS12_377_g2_affine_t* d_points,
+  size_t count,
+  size_t batch_size,
+  size_t device_id,
+  cudaStream_t stream);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _BLS12_377_MSM_H */

goicicle/curves/bls12377/include/ntt.h

@@ -0,0 +1,195 @@
+
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+#include <cuda.h>
+#include <stdbool.h>
+// ntt.h
+
+#ifndef _BLS12_377_NTT_H
+#define _BLS12_377_NTT_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+// Incomplete declaration of BLS12_377 projective and affine structs
+typedef struct BLS12_377_projective_t BLS12_377_projective_t;
+typedef struct BLS12_377_affine_t BLS12_377_affine_t;
+typedef struct BLS12_377_scalar_t BLS12_377_scalar_t;
+
+typedef struct BLS12_377_g2_projective_t BLS12_377_g2_projective_t;
+typedef struct BLS12_377_g2_affine_t BLS12_377_g2_affine_t;
+
+int ntt_cuda_bls12_377(BLS12_377_scalar_t* arr, uint32_t n, bool inverse, size_t device_id);
+int ntt_batch_cuda_bls12_377(
+  BLS12_377_scalar_t* arr, uint32_t arr_size, uint32_t batch_size, bool inverse, size_t device_id);
+
+int ecntt_cuda_bls12_377(BLS12_377_projective_t* arr, uint32_t n, bool inverse, size_t device_id);
+int ecntt_batch_cuda_bls12_377(
+  BLS12_377_projective_t* arr, uint32_t arr_size, uint32_t batch_size, bool inverse, size_t device_id);
+
+BLS12_377_scalar_t*
+build_domain_cuda_bls12_377(uint32_t domain_size, uint32_t logn, bool inverse, size_t device_id, size_t stream);
+int interpolate_scalars_cuda_bls12_377(
+  BLS12_377_scalar_t* d_out,
+  BLS12_377_scalar_t* d_evaluations,
+  BLS12_377_scalar_t* d_domain,
+  unsigned n,
+  unsigned device_id,
+  size_t stream);
+int interpolate_scalars_batch_cuda_bls12_377(
+  BLS12_377_scalar_t* d_out,
+  BLS12_377_scalar_t* d_evaluations,
+  BLS12_377_scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int interpolate_points_cuda_bls12_377(
+  BLS12_377_projective_t* d_out,
+  BLS12_377_projective_t* d_evaluations,
+  BLS12_377_scalar_t* d_domain,
+  unsigned n,
+  size_t device_id,
+  size_t stream);
+int interpolate_points_batch_cuda_bls12_377(
+  BLS12_377_projective_t* d_out,
+  BLS12_377_projective_t* d_evaluations,
+  BLS12_377_scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int interpolate_scalars_on_coset_cuda_bls12_377(
+  BLS12_377_scalar_t* d_out,
+  BLS12_377_scalar_t* d_evaluations,
+  BLS12_377_scalar_t* d_domain,
+  unsigned n,
+  BLS12_377_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int interpolate_scalars_batch_on_coset_cuda_bls12_377(
+  BLS12_377_scalar_t* d_out,
+  BLS12_377_scalar_t* d_evaluations,
+  BLS12_377_scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  BLS12_377_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int evaluate_scalars_cuda_bls12_377(
+  BLS12_377_scalar_t* d_out,
+  BLS12_377_scalar_t* d_coefficients,
+  BLS12_377_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned device_id,
+  size_t stream);
+int evaluate_scalars_batch_cuda_bls12_377(
+  BLS12_377_scalar_t* d_out,
+  BLS12_377_scalar_t* d_coefficients,
+  BLS12_377_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_cuda_bls12_377(
+  BLS12_377_projective_t* d_out,
+  BLS12_377_projective_t* d_coefficients,
+  BLS12_377_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_batch_cuda_bls12_377(
+  BLS12_377_projective_t* d_out,
+  BLS12_377_projective_t* d_coefficients,
+  BLS12_377_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int evaluate_scalars_on_coset_cuda_bls12_377(
+  BLS12_377_scalar_t* d_out,
+  BLS12_377_scalar_t* d_coefficients,
+  BLS12_377_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  BLS12_377_scalar_t* coset_powers,
+  unsigned device_id,
+  size_t stream);
+int evaluate_scalars_on_coset_batch_cuda_bls12_377(
+  BLS12_377_scalar_t* d_out,
+  BLS12_377_scalar_t* d_coefficients,
+  BLS12_377_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  BLS12_377_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_on_coset_cuda_bls12_377(
+  BLS12_377_projective_t* d_out,
+  BLS12_377_projective_t* d_coefficients,
+  BLS12_377_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  BLS12_377_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_on_coset_batch_cuda_bls12_377(
+  BLS12_377_projective_t* d_out,
+  BLS12_377_projective_t* d_coefficients,
+  BLS12_377_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  BLS12_377_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int reverse_order_scalars_cuda_bls12_377(BLS12_377_scalar_t* arr, int n, size_t device_id, size_t stream);
+int reverse_order_scalars_batch_cuda_bls12_377(
+  BLS12_377_scalar_t* arr, int n, int batch_size, size_t device_id, size_t stream);
+int reverse_order_points_cuda_bls12_377(BLS12_377_projective_t* arr, int n, size_t device_id, size_t stream);
+int reverse_order_points_batch_cuda_bls12_377(
+  BLS12_377_projective_t* arr, int n, int batch_size, size_t device_id, size_t stream);
+int add_scalars_cuda_bls12_377(
+  BLS12_377_scalar_t* d_out, BLS12_377_scalar_t* d_in1, BLS12_377_scalar_t* d_in2, unsigned n, size_t stream);
+int sub_scalars_cuda_bls12_377(
+  BLS12_377_scalar_t* d_out, BLS12_377_scalar_t* d_in1, BLS12_377_scalar_t* d_in2, unsigned n, size_t stream);
+int to_montgomery_scalars_cuda_bls12_377(BLS12_377_scalar_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_scalars_cuda_bls12_377(BLS12_377_scalar_t* d_inout, unsigned n, size_t stream);
+
+// points g1
+int to_montgomery_proj_points_cuda_bls12_377(BLS12_377_projective_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_proj_points_cuda_bls12_377(BLS12_377_projective_t* d_inout, unsigned n, size_t stream);
+int to_montgomery_aff_points_cuda_bls12_377(BLS12_377_affine_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_aff_points_cuda_bls12_377(BLS12_377_affine_t* d_inout, unsigned n, size_t stream);
+
+// points g2
+int to_montgomery_proj_points_g2_cuda_bls12_377(BLS12_377_g2_projective_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_proj_points_g2_cuda_bls12_377(BLS12_377_g2_projective_t* d_inout, unsigned n, size_t stream);
+int to_montgomery_aff_points_g2_cuda_bls12_377(BLS12_377_g2_affine_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_aff_points_g2_cuda_bls12_377(BLS12_377_g2_affine_t* d_inout, unsigned n, size_t stream);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _BLS12_377_NTT_H */

goicicle/curves/bls12377/include/projective.h

@@ -0,0 +1,50 @@
+
+	// Copyright 2023 Ingonyama
+	//
+	// Licensed under the Apache License, Version 2.0 (the "License");
+	// you may not use this file except in compliance with the License.
+	// You may obtain a copy of the License at
+	//
+	//     http://www.apache.org/licenses/LICENSE-2.0
+	//
+	// Unless required by applicable law or agreed to in writing, software
+	// distributed under the License is distributed on an "AS IS" BASIS,
+	// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+	// See the License for the specific language governing permissions and
+	// limitations under the License.
+	
+// Code generated by Ingonyama DO NOT EDIT
+
+#include <cuda.h>
+#include <stdbool.h>
+// projective.h
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+typedef struct BLS12_377_projective_t BLS12_377_projective_t;
+typedef struct BLS12_377_g2_projective_t BLS12_377_g2_projective_t;
+typedef struct BLS12_377_affine_t BLS12_377_affine_t;
+typedef struct BLS12_377_g2_affine_t BLS12_377_g2_affine_t;
+typedef struct BLS12_377_scalar_t BLS12_377_scalar_t;
+
+bool projective_is_on_curve_bls12_377(BLS12_377_projective_t* point1);
+
+int random_scalar_bls12_377(BLS12_377_scalar_t* out); 
+int random_projective_bls12_377(BLS12_377_projective_t* out);
+BLS12_377_projective_t* projective_zero_bls12_377();
+int projective_to_affine_bls12_377(BLS12_377_affine_t* out, BLS12_377_projective_t* point1);
+int projective_from_affine_bls12_377(BLS12_377_projective_t* out, BLS12_377_affine_t* point1);
+
+int random_g2_projective_bls12_377(BLS12_377_g2_projective_t* out);
+int g2_projective_to_affine_bls12_377(BLS12_377_g2_affine_t* out, BLS12_377_g2_projective_t* point1);
+int g2_projective_from_affine_bls12_377(BLS12_377_g2_projective_t* out, BLS12_377_g2_affine_t* point1);
+bool g2_projective_is_on_curve_bls12_377(BLS12_377_g2_projective_t* point1);
+
+bool eq_bls12_377(BLS12_377_projective_t* point1, BLS12_377_projective_t* point2);
+bool eq_g2_bls12_377(BLS12_377_g2_projective_t* point1, BLS12_377_g2_projective_t* point2);
+
+#ifdef __cplusplus
+}
+#endif

goicicle/curves/bls12377/include/ve_mod_mult.h

@@ -0,0 +1,49 @@
+
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+#include <cuda.h>
+#include <stdbool.h>
+// ve_mod_mult.h
+
+#ifndef _BLS12_377_VEC_MULT_H
+#define _BLS12_377_VEC_MULT_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+typedef struct BLS12_377_projective_t BLS12_377_projective_t;
+typedef struct BLS12_377_scalar_t BLS12_377_scalar_t;
+
+int32_t vec_mod_mult_point_bls12_377(
+  BLS12_377_projective_t* inout, BLS12_377_scalar_t* scalar_vec, size_t n_elments, size_t device_id);
+int32_t vec_mod_mult_scalar_bls12_377(
+  BLS12_377_scalar_t* inout, BLS12_377_scalar_t* scalar_vec, size_t n_elments, size_t device_id);
+int32_t vec_mod_mult_device_scalar_bls12_377(
+  BLS12_377_scalar_t* inout, BLS12_377_scalar_t* scalar_vec, size_t n_elements, size_t device_id);
+int32_t matrix_vec_mod_mult_bls12_377(
+  BLS12_377_scalar_t* matrix_flattened,
+  BLS12_377_scalar_t* input,
+  BLS12_377_scalar_t* output,
+  size_t n_elments,
+  size_t device_id);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _BLS12_377_VEC_MULT_H */

goicicle/curves/bls12377/msm.go

@@ -0,0 +1,208 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12377
+
+import (
+	"errors"
+	"fmt"
+	"unsafe"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbls12_377
+// #include "msm.h"
+import "C"
+
+func Msm(out *G1ProjectivePoint, points []G1PointAffine, scalars []G1ScalarField, device_id int) (*G1ProjectivePoint, error) {
+	if len(points) != len(scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	pointsC := (*C.BLS12_377_affine_t)(unsafe.Pointer(&points[0]))
+	scalarsC := (*C.BLS12_377_scalar_t)(unsafe.Pointer(&scalars[0]))
+	outC := (*C.BLS12_377_projective_t)(unsafe.Pointer(out))
+	ret := C.msm_cuda_bls12_377(outC, pointsC, scalarsC, C.size_t(len(points)), C.size_t(device_id))
+
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_cuda_bls12_377 returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func MsmG2(out *G2Point, points []G2PointAffine, scalars []G1ScalarField, device_id int) (*G2Point, error) {
+	if len(points) != len(scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	pointsC := (*C.BLS12_377_g2_affine_t)(unsafe.Pointer(&points[0]))
+	scalarsC := (*C.BLS12_377_scalar_t)(unsafe.Pointer(&scalars[0]))
+	outC := (*C.BLS12_377_g2_projective_t)(unsafe.Pointer(out))
+
+	ret := C.msm_g2_cuda_bls12_377(outC, pointsC, scalarsC, C.size_t(len(points)), C.size_t(device_id))
+
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_g2_cuda_bls12_377 returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func MsmBatch(points *[]G1PointAffine, scalars *[]G1ScalarField, batchSize, deviceId int) ([]*G1ProjectivePoint, error) {
+	// Check for nil pointers
+	if points == nil || scalars == nil {
+		return nil, errors.New("points or scalars is nil")
+	}
+
+	if len(*points) != len(*scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	// Check for empty slices
+	if len(*points) == 0 || len(*scalars) == 0 {
+		return nil, errors.New("points or scalars is empty")
+	}
+
+	// Check for zero batchSize
+	if batchSize <= 0 {
+		return nil, errors.New("error on: batchSize must be greater than zero")
+	}
+
+	out := make([]*G1ProjectivePoint, batchSize)
+
+	for i := 0; i < len(out); i++ {
+		var p G1ProjectivePoint
+		p.SetZero()
+
+		out[i] = &p
+	}
+
+	outC := (*C.BLS12_377_projective_t)(unsafe.Pointer(&out[0]))
+	pointsC := (*C.BLS12_377_affine_t)(unsafe.Pointer(&(*points)[0]))
+	scalarsC := (*C.BLS12_377_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+	msmSizeC := C.size_t(len(*points) / batchSize)
+	deviceIdC := C.size_t(deviceId)
+	batchSizeC := C.size_t(batchSize)
+
+	ret := C.msm_batch_cuda_bls12_377(outC, pointsC, scalarsC, batchSizeC, msmSizeC, deviceIdC)
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_batch_cuda_bls12_377 returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func MsmG2Batch(points *[]G2PointAffine, scalars *[]G1ScalarField, batchSize, deviceId int) ([]*G2Point, error) {
+	// Check for nil pointers
+	if points == nil || scalars == nil {
+		return nil, errors.New("points or scalars is nil")
+	}
+
+	if len(*points) != len(*scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	// Check for empty slices
+	if len(*points) == 0 || len(*scalars) == 0 {
+		return nil, errors.New("points or scalars is empty")
+	}
+
+	// Check for zero batchSize
+	if batchSize <= 0 {
+		return nil, errors.New("error on: batchSize must be greater than zero")
+	}
+
+	out := make([]*G2Point, batchSize)
+
+	outC := (*C.BLS12_377_g2_projective_t)(unsafe.Pointer(&out[0]))
+	pointsC := (*C.BLS12_377_g2_affine_t)(unsafe.Pointer(&(*points)[0]))
+	scalarsC := (*C.BLS12_377_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+	msmSizeC := C.size_t(len(*points) / batchSize)
+	deviceIdC := C.size_t(deviceId)
+	batchSizeC := C.size_t(batchSize)
+
+	ret := C.msm_batch_g2_cuda_bls12_377(outC, pointsC, scalarsC, batchSizeC, msmSizeC, deviceIdC)
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_batch_cuda_bls12_377 returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func Commit(d_out, d_scalars, d_points unsafe.Pointer, count, bucketFactor int) int {
+	d_outC := (*C.BLS12_377_projective_t)(d_out)
+	scalarsC := (*C.BLS12_377_scalar_t)(d_scalars)
+	pointsC := (*C.BLS12_377_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	largeBucketFactorC := C.uint(bucketFactor)
+
+	ret := C.commit_cuda_bls12_377(d_outC, scalarsC, pointsC, countC, largeBucketFactorC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}
+
+func CommitG2(d_out, d_scalars, d_points unsafe.Pointer, count, bucketFactor int) int {
+	d_outC := (*C.BLS12_377_g2_projective_t)(d_out)
+	scalarsC := (*C.BLS12_377_scalar_t)(d_scalars)
+	pointsC := (*C.BLS12_377_g2_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	largeBucketFactorC := C.uint(bucketFactor)
+
+	ret := C.commit_g2_cuda_bls12_377(d_outC, scalarsC, pointsC, countC, largeBucketFactorC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}
+
+func CommitBatch(d_out, d_scalars, d_points unsafe.Pointer, count, batch_size int) int {
+	d_outC := (*C.BLS12_377_projective_t)(d_out)
+	scalarsC := (*C.BLS12_377_scalar_t)(d_scalars)
+	pointsC := (*C.BLS12_377_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	batch_sizeC := (C.size_t)(batch_size)
+
+	ret := C.commit_batch_cuda_bls12_377(d_outC, scalarsC, pointsC, countC, batch_sizeC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}
+
+func CommitG2Batch(d_out, d_scalars, d_points unsafe.Pointer, count, batch_size int) int {
+	d_outC := (*C.BLS12_377_g2_projective_t)(d_out)
+	scalarsC := (*C.BLS12_377_scalar_t)(d_scalars)
+	pointsC := (*C.BLS12_377_g2_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	batch_sizeC := (C.size_t)(batch_size)
+
+	ret := C.msm_batch_g2_cuda_bls12_377(d_outC, pointsC, scalarsC, countC, batch_sizeC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}

goicicle/curves/bls12377/msm_test.go

@@ -0,0 +1,355 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12377
+
+import (
+	"fmt"
+	"math"
+	"testing"
+	"time"
+	"unsafe"
+
+	"github.com/ingonyama-zk/icicle/goicicle"
+	"github.com/stretchr/testify/assert"
+)
+
+func GeneratePoints(count int) []G1PointAffine {
+	// Declare a slice of integers
+	var points []G1PointAffine
+
+	// populate the slice
+	for i := 0; i < 10; i++ {
+		var pointProjective G1ProjectivePoint
+		pointProjective.Random()
+
+		var pointAffine G1PointAffine
+		pointAffine.FromProjective(&pointProjective)
+
+		points = append(points, pointAffine)
+	}
+
+	log2_10 := math.Log2(10)
+	log2Count := math.Log2(float64(count))
+	log2Size := int(math.Ceil(log2Count - log2_10))
+
+	for i := 0; i < log2Size; i++ {
+		points = append(points, points...)
+	}
+
+	return points[:count]
+}
+
+func GeneratePointsProj(count int) []G1ProjectivePoint {
+	// Declare a slice of integers
+	var points []G1ProjectivePoint
+	// Use a loop to populate the slice
+	for i := 0; i < count; i++ {
+		var p G1ProjectivePoint
+		p.Random()
+
+		points = append(points, p)
+	}
+
+	return points
+}
+
+func GenerateScalars(count int, skewed bool) []G1ScalarField {
+	// Declare a slice of integers
+	var scalars []G1ScalarField
+
+	var rand G1ScalarField
+	var zero G1ScalarField
+	var one G1ScalarField
+	var randLarge G1ScalarField
+
+	zero.SetZero()
+	one.SetOne()
+	randLarge.Random()
+
+	if skewed && count > 1_200_000 {
+		for i := 0; i < count-1_200_000; i++ {
+			rand.Random()
+			scalars = append(scalars, rand)
+		}
+
+		for i := 0; i < 600_000; i++ {
+			scalars = append(scalars, randLarge)
+		}
+		for i := 0; i < 400_000; i++ {
+			scalars = append(scalars, zero)
+		}
+		for i := 0; i < 200_000; i++ {
+			scalars = append(scalars, one)
+		}
+	} else {
+		for i := 0; i < count; i++ {
+			rand.Random()
+			scalars = append(scalars, rand)
+		}
+	}
+
+	return scalars[:count]
+}
+
+func TestMSM(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1 << v
+
+		points := GeneratePoints(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, true)
+		fmt.Print("Finished generating scalars\n")
+
+		out := new(G1ProjectivePoint)
+		startTime := time.Now()
+		_, e := Msm(out, points, scalars, 0) // non mont
+		fmt.Printf("icicle MSM took: %d ms\n", time.Since(startTime).Milliseconds())
+
+		assert.Equal(t, e, nil, "error should be nil")
+
+		assert.True(t, out.IsOnCurve())
+	}
+}
+
+func TestCommitMSM(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1<<v - 1
+
+		points := GeneratePoints(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, true)
+		fmt.Print("Finished generating scalars\n")
+
+		out_d, _ := goicicle.CudaMalloc(96)
+
+		pointsBytes := count * 64
+		points_d, _ := goicicle.CudaMalloc(pointsBytes)
+		goicicle.CudaMemCpyHtoD[G1PointAffine](points_d, points, pointsBytes)
+
+		scalarBytes := count * 32
+		scalars_d, _ := goicicle.CudaMalloc(scalarBytes)
+		goicicle.CudaMemCpyHtoD[G1ScalarField](scalars_d, scalars, scalarBytes)
+
+		startTime := time.Now()
+		e := Commit(out_d, scalars_d, points_d, count, 10)
+		fmt.Printf("icicle MSM took: %d ms\n", time.Since(startTime).Milliseconds())
+
+		outHost := make([]G1ProjectivePoint, 1)
+		goicicle.CudaMemCpyDtoH[G1ProjectivePoint](outHost, out_d, 96)
+
+		assert.Equal(t, e, 0, "error should be 0")
+		assert.True(t, outHost[0].IsOnCurve())
+	}
+}
+
+func BenchmarkCommit(b *testing.B) {
+	LOG_MSM_SIZES := []int{20, 21, 22, 23, 24, 25, 26}
+
+	for _, logMsmSize := range LOG_MSM_SIZES {
+		msmSize := 1 << logMsmSize
+		points := GeneratePoints(msmSize)
+		scalars := GenerateScalars(msmSize, false)
+
+		out_d, _ := goicicle.CudaMalloc(96)
+
+		pointsBytes := msmSize * 64
+		points_d, _ := goicicle.CudaMalloc(pointsBytes)
+		goicicle.CudaMemCpyHtoD[G1PointAffine](points_d, points, pointsBytes)
+
+		scalarBytes := msmSize * 32
+		scalars_d, _ := goicicle.CudaMalloc(scalarBytes)
+		goicicle.CudaMemCpyHtoD[G1ScalarField](scalars_d, scalars, scalarBytes)
+
+		b.Run(fmt.Sprintf("MSM %d", logMsmSize), func(b *testing.B) {
+			for n := 0; n < b.N; n++ {
+				e := Commit(out_d, scalars_d, points_d, msmSize, 10)
+
+				if e != 0 {
+					panic("Error occured")
+				}
+			}
+		})
+	}
+}
+
+func TestBenchMSM(t *testing.T) {
+	for _, batchPow2 := range []int{2, 4} {
+		for _, pow2 := range []int{4, 6} {
+			msmSize := 1 << pow2
+			batchSize := 1 << batchPow2
+			count := msmSize * batchSize
+
+			points := GeneratePoints(count)
+			scalars := GenerateScalars(count, false)
+
+			a, e := MsmBatch(&points, &scalars, batchSize, 0)
+
+			if e != nil {
+				t.Errorf("MsmBatchBLS12_377 returned an error: %v", e)
+			}
+
+			if len(a) != batchSize {
+				t.Errorf("Expected length %d, but got %d", batchSize, len(a))
+			}
+		}
+	}
+}
+
+func BenchmarkMSM(b *testing.B) {
+	LOG_MSM_SIZES := []int{20, 21, 22, 23, 24, 25, 26}
+
+	for _, logMsmSize := range LOG_MSM_SIZES {
+		msmSize := 1 << logMsmSize
+		points := GeneratePoints(msmSize)
+		scalars := GenerateScalars(msmSize, false)
+		b.Run(fmt.Sprintf("MSM %d", logMsmSize), func(b *testing.B) {
+			for n := 0; n < b.N; n++ {
+				out := new(G1ProjectivePoint)
+				_, e := Msm(out, points, scalars, 0)
+
+				if e != nil {
+					panic("Error occured")
+				}
+			}
+		})
+	}
+}
+
+// G2
+func GenerateG2Points(count int) []G2PointAffine {
+	// Declare a slice of integers
+	var points []G2PointAffine
+
+	// populate the slice
+	for i := 0; i < 10; i++ {
+		var p G2Point
+		p.Random()
+		var affine G2PointAffine
+		affine.FromProjective(&p)
+
+		points = append(points, affine)
+	}
+
+	log2_10 := math.Log2(10)
+	log2Count := math.Log2(float64(count))
+	log2Size := int(math.Ceil(log2Count - log2_10))
+
+	for i := 0; i < log2Size; i++ {
+		points = append(points, points...)
+	}
+
+	return points[:count]
+}
+
+func TestMsmG2BLS12_377(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1 << v
+		points := GenerateG2Points(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, false)
+		fmt.Print("Finished generating scalars\n")
+
+		out := new(G2Point)
+		_, e := MsmG2(out, points, scalars, 0)
+		assert.Equal(t, e, nil, "error should be nil")
+		assert.True(t, out.IsOnCurve())
+	}
+}
+
+func BenchmarkMsmG2BLS12_377(b *testing.B) {
+	LOG_MSM_SIZES := []int{20, 21, 22, 23, 24, 25, 26}
+
+	for _, logMsmSize := range LOG_MSM_SIZES {
+		msmSize := 1 << logMsmSize
+		points := GenerateG2Points(msmSize)
+		scalars := GenerateScalars(msmSize, false)
+		b.Run(fmt.Sprintf("MSM G2 %d", logMsmSize), func(b *testing.B) {
+			for n := 0; n < b.N; n++ {
+				out := new(G2Point)
+				_, e := MsmG2(out, points, scalars, 0)
+
+				if e != nil {
+					panic("Error occured")
+				}
+			}
+		})
+	}
+}
+
+func TestCommitG2MSM(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1 << v
+
+		points := GenerateG2Points(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, true)
+		fmt.Print("Finished generating scalars\n")
+
+		var sizeCheckG2PointAffine G2PointAffine
+		inputPointsBytes := count * int(unsafe.Sizeof(sizeCheckG2PointAffine))
+
+		var sizeCheckG2Point G2Point
+		out_d, _ := goicicle.CudaMalloc(int(unsafe.Sizeof(sizeCheckG2Point)))
+
+		points_d, _ := goicicle.CudaMalloc(inputPointsBytes)
+		goicicle.CudaMemCpyHtoD[G2PointAffine](points_d, points, inputPointsBytes)
+
+		scalarBytes := count * 32
+		scalars_d, _ := goicicle.CudaMalloc(scalarBytes)
+		goicicle.CudaMemCpyHtoD[G1ScalarField](scalars_d, scalars, scalarBytes)
+
+		startTime := time.Now()
+		e := CommitG2(out_d, scalars_d, points_d, count, 10)
+		fmt.Printf("icicle MSM took: %d ms\n", time.Since(startTime).Milliseconds())
+
+		outHost := make([]G2Point, 1)
+		goicicle.CudaMemCpyDtoH[G2Point](outHost, out_d, int(unsafe.Sizeof(sizeCheckG2Point)))
+
+		assert.Equal(t, e, 0, "error should be 0")
+		assert.Equal(t, len(outHost), 1)
+		result := outHost[0]
+
+		assert.True(t, result.IsOnCurve())
+	}
+}
+
+func TestBatchG2MSM(t *testing.T) {
+	for _, batchPow2 := range []int{2, 4} {
+		for _, pow2 := range []int{4, 6} {
+			msmSize := 1 << pow2
+			batchSize := 1 << batchPow2
+			count := msmSize * batchSize
+
+			points := GenerateG2Points(count)
+			scalars := GenerateScalars(count, false)
+
+			pointsResults, e := MsmG2Batch(&points, &scalars, batchSize, 0)
+
+			if e != nil {
+				t.Errorf("MsmBatchBLS12_377 returned an error: %v", e)
+			}
+
+			if len(pointsResults) != batchSize {
+				t.Errorf("Expected length %d, but got %d", batchSize, len(pointsResults))
+			}
+
+			for _, s := range pointsResults {
+				assert.True(t, s.IsOnCurve())
+			}
+		}
+	}
+}

goicicle/curves/bls12377/ntt.go

@@ -0,0 +1,221 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12377
+
+import (
+	"errors"
+	"fmt"
+	"unsafe"
+
+	"github.com/ingonyama-zk/icicle/goicicle"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbls12_377
+// #include "ntt.h"
+import "C"
+
+const (
+	NONE = 0
+	DIF  = 1
+	DIT  = 2
+)
+
+func Ntt(scalars *[]G1ScalarField, isInverse bool, deviceId int) uint64 {
+	scalarsC := (*C.BLS12_377_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+
+	ret := C.ntt_cuda_bls12_377(scalarsC, C.uint32_t(len(*scalars)), C.bool(isInverse), C.size_t(deviceId))
+
+	return uint64(ret)
+}
+
+func NttBatch(scalars *[]G1ScalarField, isInverse bool, batchSize, deviceId int) uint64 {
+	scalarsC := (*C.BLS12_377_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+	isInverseC := C.bool(isInverse)
+	batchSizeC := C.uint32_t(batchSize)
+	deviceIdC := C.size_t(deviceId)
+
+	ret := C.ntt_batch_cuda_bls12_377(scalarsC, C.uint32_t(len(*scalars)), batchSizeC, isInverseC, deviceIdC)
+
+	return uint64(ret)
+}
+
+func EcNtt(values *[]G1ProjectivePoint, isInverse bool, deviceId int) uint64 {
+	valuesC := (*C.BLS12_377_projective_t)(unsafe.Pointer(&(*values)[0]))
+	deviceIdC := C.size_t(deviceId)
+	isInverseC := C.bool(isInverse)
+	n := C.uint32_t(len(*values))
+
+	ret := C.ecntt_cuda_bls12_377(valuesC, n, isInverseC, deviceIdC)
+
+	return uint64(ret)
+}
+
+func EcNttBatch(values *[]G1ProjectivePoint, isInverse bool, batchSize, deviceId int) uint64 {
+	valuesC := (*C.BLS12_377_projective_t)(unsafe.Pointer(&(*values)[0]))
+	deviceIdC := C.size_t(deviceId)
+	isInverseC := C.bool(isInverse)
+	n := C.uint32_t(len(*values))
+	batchSizeC := C.uint32_t(batchSize)
+
+	ret := C.ecntt_batch_cuda_bls12_377(valuesC, n, batchSizeC, isInverseC, deviceIdC)
+
+	return uint64(ret)
+}
+
+func GenerateTwiddles(d_size int, log_d_size int, inverse bool) (up unsafe.Pointer, err error) {
+	domain_size := C.uint32_t(d_size)
+	logn := C.uint32_t(log_d_size)
+	is_inverse := C.bool(inverse)
+
+	dp := C.build_domain_cuda_bls12_377(domain_size, logn, is_inverse, 0, 0)
+
+	if dp == nil {
+		err = errors.New("nullptr returned from generating twiddles")
+		return unsafe.Pointer(nil), err
+	}
+
+	return unsafe.Pointer(dp), nil
+}
+
+// Reverses d_scalars in-place
+func ReverseScalars(d_scalars unsafe.Pointer, len int) (int, error) {
+	scalarsC := (*C.BLS12_377_scalar_t)(d_scalars)
+	lenC := C.int(len)
+	if success := C.reverse_order_scalars_cuda_bls12_377(scalarsC, lenC, 0, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func Interpolate(scalars, twiddles, cosetPowers unsafe.Pointer, size int, isCoset bool) unsafe.Pointer {
+	size_d := size * 32
+	dp, err := goicicle.CudaMalloc(size_d)
+
+	if err != nil {
+		return nil
+	}
+
+	d_out := (*C.BLS12_377_scalar_t)(dp)
+	scalarsC := (*C.BLS12_377_scalar_t)(scalars)
+	twiddlesC := (*C.BLS12_377_scalar_t)(twiddles)
+	cosetPowersC := (*C.BLS12_377_scalar_t)(cosetPowers)
+	sizeC := C.uint(size)
+
+	var ret C.int
+	if isCoset {
+		ret = C.interpolate_scalars_on_coset_cuda_bls12_377(d_out, scalarsC, twiddlesC, sizeC, cosetPowersC, 0, 0)
+	} else {
+		ret = C.interpolate_scalars_cuda_bls12_377(d_out, scalarsC, twiddlesC, sizeC, 0, 0)
+	}
+	if ret != 0 {
+		fmt.Print("error interpolating")
+	}
+
+	return unsafe.Pointer(d_out)
+}
+
+func Evaluate(scalars_out, scalars, twiddles, coset_powers unsafe.Pointer, scalars_size, twiddles_size int, isCoset bool) int {
+	scalars_outC := (*C.BLS12_377_scalar_t)(scalars_out)
+	scalarsC := (*C.BLS12_377_scalar_t)(scalars)
+	twiddlesC := (*C.BLS12_377_scalar_t)(twiddles)
+	coset_powersC := (*C.BLS12_377_scalar_t)(coset_powers)
+	sizeC := C.uint(scalars_size)
+	twiddlesC_size := C.uint(twiddles_size)
+
+	var ret C.int
+	if isCoset {
+		ret = C.evaluate_scalars_on_coset_cuda_bls12_377(scalars_outC, scalarsC, twiddlesC, twiddlesC_size, sizeC, coset_powersC, 0, 0)
+	} else {
+		ret = C.evaluate_scalars_cuda_bls12_377(scalars_outC, scalarsC, twiddlesC, twiddlesC_size, sizeC, 0, 0)
+	}
+
+	if ret != 0 {
+		fmt.Print("error interpolating")
+		return -1
+	}
+
+	return 0
+}
+
+func VecScalarAdd(in1_d, in2_d unsafe.Pointer, size int) int {
+	in1_dC := (*C.BLS12_377_scalar_t)(in1_d)
+	in2_dC := (*C.BLS12_377_scalar_t)(in2_d)
+	sizeC := C.uint(size)
+
+	ret := C.add_scalars_cuda_bls12_377(in1_dC, in1_dC, in2_dC, sizeC, 0)
+
+	if ret != 0 {
+		fmt.Print("error adding scalar vectors")
+		return -1
+	}
+
+	return 0
+}
+
+func VecScalarSub(in1_d, in2_d unsafe.Pointer, size int) int {
+	in1_dC := (*C.BLS12_377_scalar_t)(in1_d)
+	in2_dC := (*C.BLS12_377_scalar_t)(in2_d)
+	sizeC := C.uint(size)
+
+	ret := C.sub_scalars_cuda_bls12_377(in1_dC, in1_dC, in2_dC, sizeC, 0)
+
+	if ret != 0 {
+		fmt.Print("error subtracting scalar vectors")
+		return -1
+	}
+
+	return 0
+}
+
+func ToMontgomery(d_scalars unsafe.Pointer, len int) (int, error) {
+	scalarsC := (*C.BLS12_377_scalar_t)(d_scalars)
+	lenC := C.uint(len)
+	if success := C.to_montgomery_scalars_cuda_bls12_377(scalarsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func FromMontgomery(d_scalars unsafe.Pointer, len int) (int, error) {
+	scalarsC := (*C.BLS12_377_scalar_t)(d_scalars)
+	lenC := C.uint(len)
+	if success := C.from_montgomery_scalars_cuda_bls12_377(scalarsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func AffinePointFromMontgomery(d_points unsafe.Pointer, len int) (int, error) {
+	pointsC := (*C.BLS12_377_affine_t)(d_points)
+	lenC := C.uint(len)
+
+	if success := C.from_montgomery_aff_points_cuda_bls12_377(pointsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func G2AffinePointFromMontgomery(d_points unsafe.Pointer, len int) (int, error) {
+	pointsC := (*C.BLS12_377_g2_affine_t)(d_points)
+	lenC := C.uint(len)
+
+	if success := C.from_montgomery_aff_points_g2_cuda_bls12_377(pointsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}

goicicle/curves/bls12377/ntt_test.go

@@ -0,0 +1,148 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12377
+
+import (
+	"fmt"
+	"github.com/stretchr/testify/assert"
+	"reflect"
+	"testing"
+)
+
+func TestNttBLS12_377BBB(t *testing.T) {
+	count := 1 << 20
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	NttBatch(&nttResult, false, count, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	assert.Equal(t, nttResult, nttResult)
+}
+
+func TestNttBLS12_377CompareToGnarkDIF(t *testing.T) {
+	count := 1 << 2
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	Ntt(&nttResult, false, DIF, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	assert.Equal(t, nttResult, nttResult)
+}
+
+func TestINttBLS12_377CompareToGnarkDIT(t *testing.T) {
+	count := 1 << 3
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	Ntt(&nttResult, true, DIT, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	assert.Equal(t, nttResult, nttResult)
+}
+
+func TestNttBLS12_377(t *testing.T) {
+	count := 1 << 3
+
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	Ntt(&nttResult, false, NONE, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	inttResult := make([]G1ScalarField, len(nttResult))
+	copy(inttResult, nttResult)
+
+	assert.Equal(t, inttResult, nttResult)
+	Ntt(&inttResult, true, NONE, 0)
+	assert.Equal(t, inttResult, scalars)
+}
+
+func TestNttBatchBLS12_377(t *testing.T) {
+	count := 1 << 5
+	batches := 4
+
+	scalars := GenerateScalars(count*batches, false)
+
+	var scalarVecOfVec [][]G1ScalarField = make([][]G1ScalarField, 0)
+
+	for i := 0; i < batches; i++ {
+		start := i * count
+		end := (i + 1) * count
+		batch := make([]G1ScalarField, len(scalars[start:end]))
+		copy(batch, scalars[start:end])
+		scalarVecOfVec = append(scalarVecOfVec, batch)
+	}
+
+	nttBatchResult := make([]G1ScalarField, len(scalars))
+	copy(nttBatchResult, scalars)
+
+	NttBatch(&nttBatchResult, false, count, 0)
+
+	var nttResultVecOfVec [][]G1ScalarField
+
+	for i := 0; i < batches; i++ {
+		// Clone the slice
+		clone := make([]G1ScalarField, len(scalarVecOfVec[i]))
+		copy(clone, scalarVecOfVec[i])
+
+		// Add it to the result vector of vectors
+		nttResultVecOfVec = append(nttResultVecOfVec, clone)
+
+		// Call the ntt_bls12_377 function
+		Ntt(&nttResultVecOfVec[i], false, NONE, 0)
+	}
+
+	assert.NotEqual(t, nttBatchResult, scalars)
+
+	// Check that the ntt of each vec of scalars is equal to the intt of the specific batch
+	for i := 0; i < batches; i++ {
+		if !reflect.DeepEqual(nttResultVecOfVec[i], nttBatchResult[i*count:((i+1)*count)]) {
+			t.Errorf("ntt of vec of scalars not equal to intt of specific batch")
+		}
+	}
+}
+
+func BenchmarkNTT(b *testing.B) {
+	LOG_NTT_SIZES := []int{12, 15, 20, 21, 22, 23, 24, 25, 26}
+
+	for _, logNTTSize := range LOG_NTT_SIZES {
+		nttSize := 1 << logNTTSize
+		b.Run(fmt.Sprintf("NTT %d", logNTTSize), func(b *testing.B) {
+			scalars := GenerateScalars(nttSize, false)
+
+			nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+			copy(nttResult, scalars)
+			for n := 0; n < b.N; n++ {
+				Ntt(&nttResult, false, NONE, 0)
+			}
+		})
+	}
+}

goicicle/curves/bls12377/utils.go

@@ -0,0 +1,38 @@
+package bls12377
+
+import "encoding/binary"
+
+// Function to convert [8]uint32 to [4]uint64
+func ConvertUint32ArrToUint64Arr(arr32 [8]uint32) [4]uint64 {
+	var arr64 [4]uint64
+	for i := 0; i < len(arr32); i += 2 {
+		arr64[i/2] = (uint64(arr32[i]) << 32) | uint64(arr32[i+1])
+	}
+	return arr64
+}
+
+func ConvertUint64ArrToUint32Arr4(arr64 [4]uint64) [8]uint32 {
+	var arr32 [8]uint32
+	for i, v := range arr64 {
+		b := make([]byte, 8)
+		binary.LittleEndian.PutUint64(b, v)
+
+		arr32[i*2] = binary.LittleEndian.Uint32(b[0:4])
+		arr32[i*2+1] = binary.LittleEndian.Uint32(b[4:8])
+	}
+
+	return arr32
+}
+
+func ConvertUint64ArrToUint32Arr6(arr64 [6]uint64) [12]uint32 {
+	var arr32 [12]uint32
+	for i, v := range arr64 {
+		b := make([]byte, 8)
+		binary.LittleEndian.PutUint64(b, v)
+
+		arr32[i*2] = binary.LittleEndian.Uint32(b[0:4])
+		arr32[i*2+1] = binary.LittleEndian.Uint32(b[4:8])
+	}
+
+	return arr32
+}

goicicle/curves/bls12377/vec_mod.go

@@ -0,0 +1,41 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12377
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbls12_377
+// #include "ve_mod_mult.h"
+import "C"
+import (
+	"fmt"
+	"unsafe"
+)
+
+func VecScalarMulMod(scalarVec1, scalarVec2 unsafe.Pointer, size int) int {
+	scalarVec1C := (*C.BLS12_377_scalar_t)(scalarVec1)
+	scalarVec2C := (*C.BLS12_377_scalar_t)(scalarVec2)
+	sizeC := C.size_t(size)
+
+	ret := C.vec_mod_mult_device_scalar_bls12_377(scalarVec1C, scalarVec2C, sizeC, 0)
+
+	if ret != 0 {
+		fmt.Print("error multiplying scalar vectors")
+		return -1
+	}
+
+	return 0
+}

goicicle/curves/bls12381/g1.go

@@ -0,0 +1,324 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12381
+
+import (
+	"unsafe"
+
+	"encoding/binary"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbls12_381
+// #include "projective.h"
+// #include "ve_mod_mult.h"
+import "C"
+
+const SCALAR_SIZE = 8
+const BASE_SIZE = 12
+
+type G1ScalarField struct {
+	S [SCALAR_SIZE]uint32
+}
+
+type G1BaseField struct {
+	S [BASE_SIZE]uint32
+}
+
+/*
+ * BaseField Constrctors
+ */
+
+func (f *G1BaseField) SetZero() *G1BaseField {
+	var S [BASE_SIZE]uint32
+	f.S = S
+
+	return f
+}
+
+func (f *G1BaseField) SetOne() *G1BaseField {
+	var S [BASE_SIZE]uint32
+
+	S[0] = 1
+
+	f.S = S
+	return f
+}
+
+func (p *G1ProjectivePoint) FromAffine(affine *G1PointAffine) *G1ProjectivePoint {
+	out := (*C.BLS12_381_projective_t)(unsafe.Pointer(p))
+	in := (*C.BLS12_381_affine_t)(unsafe.Pointer(affine))
+
+	C.projective_from_affine_bls12_381(out, in)
+
+	return p
+}
+
+func (f *G1BaseField) FromLimbs(limbs [BASE_SIZE]uint32) *G1BaseField {
+	copy(f.S[:], limbs[:])
+
+	return f
+}
+
+/*
+ * BaseField methods
+ */
+
+func (f *G1BaseField) Limbs() [BASE_SIZE]uint32 {
+	return f.S
+}
+
+func (f *G1BaseField) ToBytesLe() []byte {
+	bytes := make([]byte, len(f.S)*4)
+	for i, v := range f.S {
+		binary.LittleEndian.PutUint32(bytes[i*4:], v)
+	}
+
+	return bytes
+}
+
+/*
+ * ScalarField methods
+ */
+
+func (p *G1ScalarField) Random() *G1ScalarField {
+	outC := (*C.BLS12_381_scalar_t)(unsafe.Pointer(p))
+	C.random_scalar_bls12_381(outC)
+
+	return p
+}
+
+func (f *G1ScalarField) SetZero() *G1ScalarField {
+	var S [SCALAR_SIZE]uint32
+	f.S = S
+
+	return f
+}
+
+func (f *G1ScalarField) SetOne() *G1ScalarField {
+	var S [SCALAR_SIZE]uint32
+	S[0] = 1
+	f.S = S
+
+	return f
+}
+
+func (a *G1ScalarField) Eq(b *G1ScalarField) bool {
+	for i, v := range a.S {
+		if b.S[i] != v {
+			return false
+		}
+	}
+	return true
+}
+
+/*
+ * ScalarField methods
+ */
+
+func (f *G1ScalarField) Limbs() [SCALAR_SIZE]uint32 {
+	return f.S
+}
+
+func (f *G1ScalarField) ToBytesLe() []byte {
+	bytes := make([]byte, len(f.S)*4)
+	for i, v := range f.S {
+		binary.LittleEndian.PutUint32(bytes[i*4:], v)
+	}
+
+	return bytes
+}
+
+/*
+ * PointBLS12_381
+ */
+
+type G1ProjectivePoint struct {
+	X, Y, Z G1BaseField
+}
+
+func (f *G1ProjectivePoint) SetZero() *G1ProjectivePoint {
+	var yOne G1BaseField
+	yOne.SetOne()
+
+	var xZero G1BaseField
+	xZero.SetZero()
+
+	var zZero G1BaseField
+	zZero.SetZero()
+
+	f.X = xZero
+	f.Y = yOne
+	f.Z = zZero
+
+	return f
+}
+
+func (p *G1ProjectivePoint) Eq(pCompare *G1ProjectivePoint) bool {
+	// Cast *PointBLS12_381 to *C.BLS12_381_projective_t
+	// The unsafe.Pointer cast is necessary because Go doesn't allow direct casts
+	// between different pointer types.
+	// It'S your responsibility to ensure that the types are compatible.
+	pC := (*C.BLS12_381_projective_t)(unsafe.Pointer(p))
+	pCompareC := (*C.BLS12_381_projective_t)(unsafe.Pointer(pCompare))
+
+	// Call the C function
+	// The C function doesn't keep any references to the data,
+	// so it'S fine if the Go garbage collector moves or deletes the data later.
+	return bool(C.eq_bls12_381(pC, pCompareC))
+}
+
+func (p *G1ProjectivePoint) IsOnCurve() bool {
+	point := (*C.BLS12_381_projective_t)(unsafe.Pointer(p))
+	res := C.projective_is_on_curve_bls12_381(point)
+
+	return bool(res)
+}
+
+func (p *G1ProjectivePoint) Random() *G1ProjectivePoint {
+	outC := (*C.BLS12_381_projective_t)(unsafe.Pointer(p))
+	C.random_projective_bls12_381(outC)
+
+	return p
+}
+
+func (p *G1ProjectivePoint) StripZ() *G1PointAffine {
+	return &G1PointAffine{
+		X: p.X,
+		Y: p.Y,
+	}
+}
+
+func (p *G1ProjectivePoint) FromLimbs(x, y, z *[]uint32) *G1ProjectivePoint {
+	var _x G1BaseField
+	var _y G1BaseField
+	var _z G1BaseField
+
+	_x.FromLimbs(GetFixedLimbs(x))
+	_y.FromLimbs(GetFixedLimbs(y))
+	_z.FromLimbs(GetFixedLimbs(z))
+
+	p.X = _x
+	p.Y = _y
+	p.Z = _z
+
+	return p
+}
+
+/*
+ * PointAffineNoInfinityBLS12_381
+ */
+
+type G1PointAffine struct {
+	X, Y G1BaseField
+}
+
+func (p *G1PointAffine) FromProjective(projective *G1ProjectivePoint) *G1PointAffine {
+	in := (*C.BLS12_381_projective_t)(unsafe.Pointer(projective))
+	out := (*C.BLS12_381_affine_t)(unsafe.Pointer(p))
+
+	C.projective_to_affine_bls12_381(out, in)
+
+	return p
+}
+
+func (p *G1PointAffine) ToProjective() *G1ProjectivePoint {
+	var Z G1BaseField
+	Z.SetOne()
+
+	return &G1ProjectivePoint{
+		X: p.X,
+		Y: p.Y,
+		Z: Z,
+	}
+}
+
+func (p *G1PointAffine) FromLimbs(X, Y *[]uint32) *G1PointAffine {
+	var _x G1BaseField
+	var _y G1BaseField
+
+	_x.FromLimbs(GetFixedLimbs(X))
+	_y.FromLimbs(GetFixedLimbs(Y))
+
+	return p
+}
+
+/*
+ * Multiplication
+ */
+
+func MultiplyVec(a []G1ProjectivePoint, b []G1ScalarField, deviceID int) {
+	if len(a) != len(b) {
+		panic("a and b have different lengths")
+	}
+
+	pointsC := (*C.BLS12_381_projective_t)(unsafe.Pointer(&a[0]))
+	scalarsC := (*C.BLS12_381_scalar_t)(unsafe.Pointer(&b[0]))
+	deviceIdC := C.size_t(deviceID)
+	nElementsC := C.size_t(len(a))
+
+	C.vec_mod_mult_point_bls12_381(pointsC, scalarsC, nElementsC, deviceIdC)
+}
+
+func MultiplyScalar(a []G1ScalarField, b []G1ScalarField, deviceID int) {
+	if len(a) != len(b) {
+		panic("a and b have different lengths")
+	}
+
+	aC := (*C.BLS12_381_scalar_t)(unsafe.Pointer(&a[0]))
+	bC := (*C.BLS12_381_scalar_t)(unsafe.Pointer(&b[0]))
+	deviceIdC := C.size_t(deviceID)
+	nElementsC := C.size_t(len(a))
+
+	C.vec_mod_mult_scalar_bls12_381(aC, bC, nElementsC, deviceIdC)
+}
+
+// Multiply a matrix by a scalar:
+//
+//	`a` - flattenned matrix;
+//	`b` - vector to multiply `a` by;
+func MultiplyMatrix(a []G1ScalarField, b []G1ScalarField, deviceID int) {
+	c := make([]G1ScalarField, len(b))
+	for i := range c {
+		var p G1ScalarField
+		p.SetZero()
+
+		c[i] = p
+	}
+
+	aC := (*C.BLS12_381_scalar_t)(unsafe.Pointer(&a[0]))
+	bC := (*C.BLS12_381_scalar_t)(unsafe.Pointer(&b[0]))
+	cC := (*C.BLS12_381_scalar_t)(unsafe.Pointer(&c[0]))
+	deviceIdC := C.size_t(deviceID)
+	nElementsC := C.size_t(len(a))
+
+	C.matrix_vec_mod_mult_bls12_381(aC, bC, cC, nElementsC, deviceIdC)
+}
+
+/*
+ * Utils
+ */
+
+func GetFixedLimbs(slice *[]uint32) [BASE_SIZE]uint32 {
+	if len(*slice) <= BASE_SIZE {
+		limbs := [BASE_SIZE]uint32{}
+		copy(limbs[:len(*slice)], *slice)
+		return limbs
+	}
+
+	panic("slice has too many elements")
+}

goicicle/curves/bls12381/g1_test.go

@@ -0,0 +1,198 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12381
+
+import (
+	"encoding/binary"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestNewFieldBLS12_381One(t *testing.T) {
+	var oneField G1BaseField
+	oneField.SetOne()
+
+	rawOneField := [8]uint32([8]uint32{0x1, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0})
+
+	assert.Equal(t, oneField.S, rawOneField)
+}
+
+func TestNewFieldBLS12_381Zero(t *testing.T) {
+	var zeroField G1BaseField
+	zeroField.SetZero()
+
+	rawZeroField := [8]uint32([8]uint32{0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0})
+
+	assert.Equal(t, zeroField.S, rawZeroField)
+}
+
+func TestFieldBLS12_381ToBytesLe(t *testing.T) {
+	var p G1ProjectivePoint
+	p.Random()
+
+	expected := make([]byte, len(p.X.S)*4) // each uint32 takes 4 bytes
+	for i, v := range p.X.S {
+		binary.LittleEndian.PutUint32(expected[i*4:], v)
+	}
+
+	assert.Equal(t, p.X.ToBytesLe(), expected)
+	assert.Equal(t, len(p.X.ToBytesLe()), 32)
+}
+
+func TestNewPointBLS12_381Zero(t *testing.T) {
+	var pointZero G1ProjectivePoint
+	pointZero.SetZero()
+
+	var baseOne G1BaseField
+	baseOne.SetOne()
+
+	var zeroSanity G1BaseField
+	zeroSanity.SetZero()
+
+	assert.Equal(t, pointZero.X, zeroSanity)
+	assert.Equal(t, pointZero.Y, baseOne)
+	assert.Equal(t, pointZero.Z, zeroSanity)
+}
+
+func TestFromProjectiveToAffine(t *testing.T) {
+	var projective G1ProjectivePoint
+	var affine G1PointAffine
+
+	projective.Random()
+
+	affine.FromProjective(&projective)
+	var projective2 G1ProjectivePoint
+	projective2.FromAffine(&affine)
+
+	assert.True(t, projective.IsOnCurve())
+	assert.True(t, projective2.IsOnCurve())
+	assert.True(t, projective.Eq(&projective2))
+}
+
+func TestBLS12_381Eq(t *testing.T) {
+	var p1 G1ProjectivePoint
+	p1.Random()
+	var p2 G1ProjectivePoint
+	p2.Random()
+
+	assert.Equal(t, p1.Eq(&p1), true)
+	assert.Equal(t, p1.Eq(&p2), false)
+}
+
+func TestBLS12_381StripZ(t *testing.T) {
+	var p1 G1ProjectivePoint
+	p1.Random()
+
+	p2ZLess := p1.StripZ()
+
+	assert.IsType(t, G1PointAffine{}, *p2ZLess)
+	assert.Equal(t, p1.X, p2ZLess.X)
+	assert.Equal(t, p1.Y, p2ZLess.Y)
+}
+
+func TestPointBLS12_381fromLimbs(t *testing.T) {
+	var p G1ProjectivePoint
+	p.Random()
+
+	x := p.X.Limbs()
+	y := p.Y.Limbs()
+	z := p.Z.Limbs()
+
+	xSlice := x[:]
+	ySlice := y[:]
+	zSlice := z[:]
+
+	var pFromLimbs G1ProjectivePoint
+	pFromLimbs.FromLimbs(&xSlice, &ySlice, &zSlice)
+
+	assert.Equal(t, pFromLimbs, p)
+}
+
+func TestNewPointAffineNoInfinityBLS12_381Zero(t *testing.T) {
+	var zeroP G1PointAffine
+
+	var zeroSanity G1BaseField
+	zeroSanity.SetZero()
+
+	assert.Equal(t, zeroP.X, zeroSanity)
+	assert.Equal(t, zeroP.Y, zeroSanity)
+}
+
+func TestPointAffineNoInfinityBLS12_381FromLimbs(t *testing.T) {
+	// Initialize your test values
+	x := [12]uint32{1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12}
+	y := [12]uint32{9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20}
+	xSlice := x[:]
+	ySlice := y[:]
+
+	// Execute your function
+	var result G1PointAffine
+	result.FromLimbs(&xSlice, &ySlice)
+
+	var xBase G1BaseField
+	var yBase G1BaseField
+	xBase.FromLimbs(x)
+	yBase.FromLimbs(y)
+
+	// Define your expected result
+	expected := &G1PointAffine{
+		X: xBase,
+		Y: yBase,
+	}
+
+	// Test if result is as expected
+	assert.Equal(t, result, expected)
+}
+
+func TestGetFixedLimbs(t *testing.T) {
+	t.Run("case of valid input of length less than 8", func(t *testing.T) {
+		slice := []uint32{1, 2, 3, 4, 5, 6, 7}
+		expected := [8]uint32{1, 2, 3, 4, 5, 6, 7, 0}
+
+		result := GetFixedLimbs(&slice)
+		assert.Equal(t, result, expected)
+	})
+
+	t.Run("case of valid input of length 8", func(t *testing.T) {
+		slice := []uint32{1, 2, 3, 4, 5, 6, 7, 8}
+		expected := [8]uint32{1, 2, 3, 4, 5, 6, 7, 8}
+
+		result := GetFixedLimbs(&slice)
+		assert.Equal(t, result, expected)
+	})
+
+	t.Run("case of empty input", func(t *testing.T) {
+		slice := []uint32{}
+		expected := [8]uint32{0, 0, 0, 0, 0, 0, 0, 0}
+
+		result := GetFixedLimbs(&slice)
+		assert.Equal(t, result, expected)
+	})
+
+	t.Run("case of input length greater than 8", func(t *testing.T) {
+		slice := []uint32{1, 2, 3, 4, 5, 6, 7, 8, 9}
+
+		defer func() {
+			if r := recover(); r == nil {
+				t.Errorf("the code did not panic")
+			}
+		}()
+
+		GetFixedLimbs(&slice)
+	})
+}

goicicle/curves/bls12381/g2.go

@@ -0,0 +1,112 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12381
+
+import (
+	"encoding/binary"
+	"unsafe"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbls12_381
+// #include "projective.h"
+// #include "ve_mod_mult.h"
+import "C"
+
+// G2 extension field
+
+type G2Element [6]uint64
+
+type ExtentionField struct {
+	A0, A1 G2Element
+}
+
+type G2PointAffine struct {
+	X, Y ExtentionField
+}
+
+type G2Point struct {
+	X, Y, Z ExtentionField
+}
+
+func (p *G2Point) Random() *G2Point {
+	outC := (*C.BLS12_381_g2_projective_t)(unsafe.Pointer(p))
+	C.random_g2_projective_bls12_381(outC)
+
+	return p
+}
+
+func (p *G2Point) FromAffine(affine *G2PointAffine) *G2Point {
+	out := (*C.BLS12_381_g2_projective_t)(unsafe.Pointer(p))
+	in := (*C.BLS12_381_g2_affine_t)(unsafe.Pointer(affine))
+
+	C.g2_projective_from_affine_bls12_381(out, in)
+
+	return p
+}
+
+func (p *G2Point) Eq(pCompare *G2Point) bool {
+	// Cast *PointBLS12_381 to *C.BLS12_381_projective_t
+	// The unsafe.Pointer cast is necessary because Go doesn't allow direct casts
+	// between different pointer types.
+	// It's your responsibility to ensure that the types are compatible.
+	pC := (*C.BLS12_381_g2_projective_t)(unsafe.Pointer(p))
+	pCompareC := (*C.BLS12_381_g2_projective_t)(unsafe.Pointer(pCompare))
+
+	// Call the C function
+	// The C function doesn't keep any references to the data,
+	// so it's fine if the Go garbage collector moves or deletes the data later.
+	return bool(C.eq_g2_bls12_381(pC, pCompareC))
+}
+
+func (f *G2Element) ToBytesLe() []byte {
+	var bytes []byte
+	for _, val := range f {
+		buf := make([]byte, 8) // 8 bytes because uint64 is 64-bit
+		binary.LittleEndian.PutUint64(buf, val)
+		bytes = append(bytes, buf...)
+	}
+	return bytes
+}
+
+func (p *G2PointAffine) ToProjective() G2Point {
+	return G2Point{
+		X: p.X,
+		Y: p.Y,
+		Z: ExtentionField{
+			A0: G2Element{1, 0, 0, 0},
+			A1: G2Element{0, 0, 0, 0},
+		},
+	}
+}
+
+func (p *G2PointAffine) FromProjective(projective *G2Point) *G2PointAffine {
+	out := (*C.BLS12_381_g2_affine_t)(unsafe.Pointer(p))
+	in := (*C.BLS12_381_g2_projective_t)(unsafe.Pointer(projective))
+
+	C.g2_projective_to_affine_bls12_381(out, in)
+
+	return p
+}
+
+func (p *G2Point) IsOnCurve() bool {
+	// Directly copy memory from the C struct to the Go struct
+	point := (*C.BLS12_381_g2_projective_t)(unsafe.Pointer(p))
+	res := C.g2_projective_is_on_curve_bls12_381(point)
+
+	return bool(res)
+}

goicicle/curves/bls12381/g2_test.go

@@ -0,0 +1,75 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12381
+
+import (
+	"github.com/stretchr/testify/assert"
+	"testing"
+)
+
+func TestG2Eqg2(t *testing.T) {
+	var point G2Point
+
+	point.Random()
+
+	assert.True(t, point.Eq(&point))
+}
+
+func TestG2FromProjectiveToAffine(t *testing.T) {
+	var projective G2Point
+	var affine G2PointAffine
+
+	projective.Random()
+
+	affine.FromProjective(&projective)
+	var projective2 G2Point
+	projective2.FromAffine(&affine)
+
+	assert.True(t, projective.IsOnCurve())
+	assert.True(t, projective2.IsOnCurve())
+	assert.True(t, projective.Eq(&projective2))
+}
+
+func TestG2Eqg2NotEqual(t *testing.T) {
+	var point G2Point
+	point.Random()
+
+	var point2 G2Point
+	point2.Random()
+
+	assert.False(t, point.Eq(&point2))
+}
+
+func TestG2ToBytes(t *testing.T) {
+	element := G2Element{0x6546098ea84b6298, 0x4a384533d1f68aca, 0xaa0666972d771336, 0x1569e4a34321993}
+	bytes := element.ToBytesLe()
+
+	assert.Equal(t, bytes, []byte{0x98, 0x62, 0x4b, 0xa8, 0x8e, 0x9, 0x46, 0x65, 0xca, 0x8a, 0xf6, 0xd1, 0x33, 0x45, 0x38, 0x4a, 0x36, 0x13, 0x77, 0x2d, 0x97, 0x66, 0x6, 0xaa, 0x93, 0x19, 0x32, 0x34, 0x4a, 0x9e, 0x56, 0x1})
+}
+
+func TestG2ShouldConvertToProjective(t *testing.T) {
+	var pointProjective G2Point
+	var pointAffine G2PointAffine
+
+	pointProjective.Random()
+	pointAffine.FromProjective(&pointProjective)
+
+	proj := pointAffine.ToProjective()
+
+	assert.True(t, proj.IsOnCurve())
+	assert.True(t, pointProjective.Eq(&proj))
+}

goicicle/curves/bls12381/include/msm.h

@@ -0,0 +1,98 @@
+
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+#include <cuda.h>
+#include <cuda_runtime.h>
+#include <stdbool.h>
+// msm.h
+
+#ifndef _BLS12_381_MSM_H
+#define _BLS12_381_MSM_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+// Incomplete declaration of BLS12_381 projective and affine structs
+typedef struct BLS12_381_projective_t BLS12_381_projective_t;
+typedef struct BLS12_381_g2_projective_t BLS12_381_g2_projective_t;
+typedef struct BLS12_381_affine_t BLS12_381_affine_t;
+typedef struct BLS12_381_g2_affine_t BLS12_381_g2_affine_t;
+typedef struct BLS12_381_scalar_t BLS12_381_scalar_t;
+typedef cudaStream_t CudaStream_t;
+
+int msm_cuda_bls12_381(
+  BLS12_381_projective_t* out, BLS12_381_affine_t* points, BLS12_381_scalar_t* scalars, size_t count, size_t device_id);
+
+int msm_batch_cuda_bls12_381(
+  BLS12_381_projective_t* out,
+  BLS12_381_affine_t* points,
+  BLS12_381_scalar_t* scalars,
+  size_t batch_size,
+  size_t msm_size,
+  size_t device_id);
+
+int commit_cuda_bls12_381(
+  BLS12_381_projective_t* d_out,
+  BLS12_381_scalar_t* d_scalars,
+  BLS12_381_affine_t* d_points,
+  size_t count,
+  unsigned large_bucket_factor,
+  size_t device_id);
+
+int commit_batch_cuda_bls12_381(
+  BLS12_381_projective_t* d_out,
+  BLS12_381_scalar_t* d_scalars,
+  BLS12_381_affine_t* d_points,
+  size_t count,
+  size_t batch_size,
+  size_t device_id);
+
+int msm_g2_cuda_bls12_381(
+  BLS12_381_g2_projective_t* out,
+  BLS12_381_g2_affine_t* points,
+  BLS12_381_scalar_t* scalars,
+  size_t count,
+  size_t device_id);
+int msm_batch_g2_cuda_bls12_381(
+  BLS12_381_g2_projective_t* out,
+  BLS12_381_g2_affine_t* points,
+  BLS12_381_scalar_t* scalars,
+  size_t batch_size,
+  size_t msm_size,
+  size_t device_id);
+int commit_g2_cuda_bls12_381(
+  BLS12_381_g2_projective_t* d_out,
+  BLS12_381_scalar_t* d_scalars,
+  BLS12_381_g2_affine_t* d_points,
+  size_t count,
+  unsigned large_bucket_factor,
+  size_t device_id);
+int commit_batch_g2_cuda_bls12_381(
+  BLS12_381_g2_projective_t* d_out,
+  BLS12_381_scalar_t* d_scalars,
+  BLS12_381_g2_affine_t* d_points,
+  size_t count,
+  size_t batch_size,
+  size_t device_id,
+  cudaStream_t stream);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _BLS12_381_MSM_H */

goicicle/curves/bls12381/include/ntt.h

@@ -0,0 +1,195 @@
+
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+#include <cuda.h>
+#include <stdbool.h>
+// ntt.h
+
+#ifndef _BLS12_381_NTT_H
+#define _BLS12_381_NTT_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+// Incomplete declaration of BLS12_381 projective and affine structs
+typedef struct BLS12_381_projective_t BLS12_381_projective_t;
+typedef struct BLS12_381_affine_t BLS12_381_affine_t;
+typedef struct BLS12_381_scalar_t BLS12_381_scalar_t;
+
+typedef struct BLS12_381_g2_projective_t BLS12_381_g2_projective_t;
+typedef struct BLS12_381_g2_affine_t BLS12_381_g2_affine_t;
+
+int ntt_cuda_bls12_381(BLS12_381_scalar_t* arr, uint32_t n, bool inverse, size_t device_id);
+int ntt_batch_cuda_bls12_381(
+  BLS12_381_scalar_t* arr, uint32_t arr_size, uint32_t batch_size, bool inverse, size_t device_id);
+
+int ecntt_cuda_bls12_381(BLS12_381_projective_t* arr, uint32_t n, bool inverse, size_t device_id);
+int ecntt_batch_cuda_bls12_381(
+  BLS12_381_projective_t* arr, uint32_t arr_size, uint32_t batch_size, bool inverse, size_t device_id);
+
+BLS12_381_scalar_t*
+build_domain_cuda_bls12_381(uint32_t domain_size, uint32_t logn, bool inverse, size_t device_id, size_t stream);
+int interpolate_scalars_cuda_bls12_381(
+  BLS12_381_scalar_t* d_out,
+  BLS12_381_scalar_t* d_evaluations,
+  BLS12_381_scalar_t* d_domain,
+  unsigned n,
+  unsigned device_id,
+  size_t stream);
+int interpolate_scalars_batch_cuda_bls12_381(
+  BLS12_381_scalar_t* d_out,
+  BLS12_381_scalar_t* d_evaluations,
+  BLS12_381_scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int interpolate_points_cuda_bls12_381(
+  BLS12_381_projective_t* d_out,
+  BLS12_381_projective_t* d_evaluations,
+  BLS12_381_scalar_t* d_domain,
+  unsigned n,
+  size_t device_id,
+  size_t stream);
+int interpolate_points_batch_cuda_bls12_381(
+  BLS12_381_projective_t* d_out,
+  BLS12_381_projective_t* d_evaluations,
+  BLS12_381_scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int interpolate_scalars_on_coset_cuda_bls12_381(
+  BLS12_381_scalar_t* d_out,
+  BLS12_381_scalar_t* d_evaluations,
+  BLS12_381_scalar_t* d_domain,
+  unsigned n,
+  BLS12_381_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int interpolate_scalars_batch_on_coset_cuda_bls12_381(
+  BLS12_381_scalar_t* d_out,
+  BLS12_381_scalar_t* d_evaluations,
+  BLS12_381_scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  BLS12_381_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int evaluate_scalars_cuda_bls12_381(
+  BLS12_381_scalar_t* d_out,
+  BLS12_381_scalar_t* d_coefficients,
+  BLS12_381_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned device_id,
+  size_t stream);
+int evaluate_scalars_batch_cuda_bls12_381(
+  BLS12_381_scalar_t* d_out,
+  BLS12_381_scalar_t* d_coefficients,
+  BLS12_381_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_cuda_bls12_381(
+  BLS12_381_projective_t* d_out,
+  BLS12_381_projective_t* d_coefficients,
+  BLS12_381_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_batch_cuda_bls12_381(
+  BLS12_381_projective_t* d_out,
+  BLS12_381_projective_t* d_coefficients,
+  BLS12_381_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int evaluate_scalars_on_coset_cuda_bls12_381(
+  BLS12_381_scalar_t* d_out,
+  BLS12_381_scalar_t* d_coefficients,
+  BLS12_381_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  BLS12_381_scalar_t* coset_powers,
+  unsigned device_id,
+  size_t stream);
+int evaluate_scalars_on_coset_batch_cuda_bls12_381(
+  BLS12_381_scalar_t* d_out,
+  BLS12_381_scalar_t* d_coefficients,
+  BLS12_381_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  BLS12_381_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_on_coset_cuda_bls12_381(
+  BLS12_381_projective_t* d_out,
+  BLS12_381_projective_t* d_coefficients,
+  BLS12_381_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  BLS12_381_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_on_coset_batch_cuda_bls12_381(
+  BLS12_381_projective_t* d_out,
+  BLS12_381_projective_t* d_coefficients,
+  BLS12_381_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  BLS12_381_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int reverse_order_scalars_cuda_bls12_381(BLS12_381_scalar_t* arr, int n, size_t device_id, size_t stream);
+int reverse_order_scalars_batch_cuda_bls12_381(
+  BLS12_381_scalar_t* arr, int n, int batch_size, size_t device_id, size_t stream);
+int reverse_order_points_cuda_bls12_381(BLS12_381_projective_t* arr, int n, size_t device_id, size_t stream);
+int reverse_order_points_batch_cuda_bls12_381(
+  BLS12_381_projective_t* arr, int n, int batch_size, size_t device_id, size_t stream);
+int add_scalars_cuda_bls12_381(
+  BLS12_381_scalar_t* d_out, BLS12_381_scalar_t* d_in1, BLS12_381_scalar_t* d_in2, unsigned n, size_t stream);
+int sub_scalars_cuda_bls12_381(
+  BLS12_381_scalar_t* d_out, BLS12_381_scalar_t* d_in1, BLS12_381_scalar_t* d_in2, unsigned n, size_t stream);
+int to_montgomery_scalars_cuda_bls12_381(BLS12_381_scalar_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_scalars_cuda_bls12_381(BLS12_381_scalar_t* d_inout, unsigned n, size_t stream);
+
+// points g1
+int to_montgomery_proj_points_cuda_bls12_381(BLS12_381_projective_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_proj_points_cuda_bls12_381(BLS12_381_projective_t* d_inout, unsigned n, size_t stream);
+int to_montgomery_aff_points_cuda_bls12_381(BLS12_381_affine_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_aff_points_cuda_bls12_381(BLS12_381_affine_t* d_inout, unsigned n, size_t stream);
+
+// points g2
+int to_montgomery_proj_points_g2_cuda_bls12_381(BLS12_381_g2_projective_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_proj_points_g2_cuda_bls12_381(BLS12_381_g2_projective_t* d_inout, unsigned n, size_t stream);
+int to_montgomery_aff_points_g2_cuda_bls12_381(BLS12_381_g2_affine_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_aff_points_g2_cuda_bls12_381(BLS12_381_g2_affine_t* d_inout, unsigned n, size_t stream);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _BLS12_381_NTT_H */

goicicle/curves/bls12381/include/projective.h

@@ -0,0 +1,50 @@
+
+	// Copyright 2023 Ingonyama
+	//
+	// Licensed under the Apache License, Version 2.0 (the "License");
+	// you may not use this file except in compliance with the License.
+	// You may obtain a copy of the License at
+	//
+	//     http://www.apache.org/licenses/LICENSE-2.0
+	//
+	// Unless required by applicable law or agreed to in writing, software
+	// distributed under the License is distributed on an "AS IS" BASIS,
+	// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+	// See the License for the specific language governing permissions and
+	// limitations under the License.
+	
+// Code generated by Ingonyama DO NOT EDIT
+
+#include <cuda.h>
+#include <stdbool.h>
+// projective.h
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+typedef struct BLS12_381_projective_t BLS12_381_projective_t;
+typedef struct BLS12_381_g2_projective_t BLS12_381_g2_projective_t;
+typedef struct BLS12_381_affine_t BLS12_381_affine_t;
+typedef struct BLS12_381_g2_affine_t BLS12_381_g2_affine_t;
+typedef struct BLS12_381_scalar_t BLS12_381_scalar_t;
+
+bool projective_is_on_curve_bls12_381(BLS12_381_projective_t* point1);
+
+int random_scalar_bls12_381(BLS12_381_scalar_t* out); 
+int random_projective_bls12_381(BLS12_381_projective_t* out);
+BLS12_381_projective_t* projective_zero_bls12_381();
+int projective_to_affine_bls12_381(BLS12_381_affine_t* out, BLS12_381_projective_t* point1);
+int projective_from_affine_bls12_381(BLS12_381_projective_t* out, BLS12_381_affine_t* point1);
+
+int random_g2_projective_bls12_381(BLS12_381_g2_projective_t* out);
+int g2_projective_to_affine_bls12_381(BLS12_381_g2_affine_t* out, BLS12_381_g2_projective_t* point1);
+int g2_projective_from_affine_bls12_381(BLS12_381_g2_projective_t* out, BLS12_381_g2_affine_t* point1);
+bool g2_projective_is_on_curve_bls12_381(BLS12_381_g2_projective_t* point1);
+
+bool eq_bls12_381(BLS12_381_projective_t* point1, BLS12_381_projective_t* point2);
+bool eq_g2_bls12_381(BLS12_381_g2_projective_t* point1, BLS12_381_g2_projective_t* point2);
+
+#ifdef __cplusplus
+}
+#endif

goicicle/curves/bls12381/include/ve_mod_mult.h

@@ -0,0 +1,49 @@
+
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+#include <cuda.h>
+#include <stdbool.h>
+// ve_mod_mult.h
+
+#ifndef _BLS12_381_VEC_MULT_H
+#define _BLS12_381_VEC_MULT_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+typedef struct BLS12_381_projective_t BLS12_381_projective_t;
+typedef struct BLS12_381_scalar_t BLS12_381_scalar_t;
+
+int32_t vec_mod_mult_point_bls12_381(
+  BLS12_381_projective_t* inout, BLS12_381_scalar_t* scalar_vec, size_t n_elments, size_t device_id);
+int32_t vec_mod_mult_scalar_bls12_381(
+  BLS12_381_scalar_t* inout, BLS12_381_scalar_t* scalar_vec, size_t n_elments, size_t device_id);
+int32_t vec_mod_mult_device_scalar_bls12_381(
+  BLS12_381_scalar_t* inout, BLS12_381_scalar_t* scalar_vec, size_t n_elements, size_t device_id);
+int32_t matrix_vec_mod_mult_bls12_381(
+  BLS12_381_scalar_t* matrix_flattened,
+  BLS12_381_scalar_t* input,
+  BLS12_381_scalar_t* output,
+  size_t n_elments,
+  size_t device_id);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _BLS12_381_VEC_MULT_H */

goicicle/curves/bls12381/msm.go

@@ -0,0 +1,208 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12381
+
+import (
+	"errors"
+	"fmt"
+	"unsafe"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbls12_381
+// #include "msm.h"
+import "C"
+
+func Msm(out *G1ProjectivePoint, points []G1PointAffine, scalars []G1ScalarField, device_id int) (*G1ProjectivePoint, error) {
+	if len(points) != len(scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	pointsC := (*C.BLS12_381_affine_t)(unsafe.Pointer(&points[0]))
+	scalarsC := (*C.BLS12_381_scalar_t)(unsafe.Pointer(&scalars[0]))
+	outC := (*C.BLS12_381_projective_t)(unsafe.Pointer(out))
+	ret := C.msm_cuda_bls12_381(outC, pointsC, scalarsC, C.size_t(len(points)), C.size_t(device_id))
+
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_cuda_bls12_381 returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func MsmG2(out *G2Point, points []G2PointAffine, scalars []G1ScalarField, device_id int) (*G2Point, error) {
+	if len(points) != len(scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	pointsC := (*C.BLS12_381_g2_affine_t)(unsafe.Pointer(&points[0]))
+	scalarsC := (*C.BLS12_381_scalar_t)(unsafe.Pointer(&scalars[0]))
+	outC := (*C.BLS12_381_g2_projective_t)(unsafe.Pointer(out))
+
+	ret := C.msm_g2_cuda_bls12_381(outC, pointsC, scalarsC, C.size_t(len(points)), C.size_t(device_id))
+
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_g2_cuda_bls12_381 returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func MsmBatch(points *[]G1PointAffine, scalars *[]G1ScalarField, batchSize, deviceId int) ([]*G1ProjectivePoint, error) {
+	// Check for nil pointers
+	if points == nil || scalars == nil {
+		return nil, errors.New("points or scalars is nil")
+	}
+
+	if len(*points) != len(*scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	// Check for empty slices
+	if len(*points) == 0 || len(*scalars) == 0 {
+		return nil, errors.New("points or scalars is empty")
+	}
+
+	// Check for zero batchSize
+	if batchSize <= 0 {
+		return nil, errors.New("error on: batchSize must be greater than zero")
+	}
+
+	out := make([]*G1ProjectivePoint, batchSize)
+
+	for i := 0; i < len(out); i++ {
+		var p G1ProjectivePoint
+		p.SetZero()
+
+		out[i] = &p
+	}
+
+	outC := (*C.BLS12_381_projective_t)(unsafe.Pointer(&out[0]))
+	pointsC := (*C.BLS12_381_affine_t)(unsafe.Pointer(&(*points)[0]))
+	scalarsC := (*C.BLS12_381_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+	msmSizeC := C.size_t(len(*points) / batchSize)
+	deviceIdC := C.size_t(deviceId)
+	batchSizeC := C.size_t(batchSize)
+
+	ret := C.msm_batch_cuda_bls12_381(outC, pointsC, scalarsC, batchSizeC, msmSizeC, deviceIdC)
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_batch_cuda_bls12_381 returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func MsmG2Batch(points *[]G2PointAffine, scalars *[]G1ScalarField, batchSize, deviceId int) ([]*G2Point, error) {
+	// Check for nil pointers
+	if points == nil || scalars == nil {
+		return nil, errors.New("points or scalars is nil")
+	}
+
+	if len(*points) != len(*scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	// Check for empty slices
+	if len(*points) == 0 || len(*scalars) == 0 {
+		return nil, errors.New("points or scalars is empty")
+	}
+
+	// Check for zero batchSize
+	if batchSize <= 0 {
+		return nil, errors.New("error on: batchSize must be greater than zero")
+	}
+
+	out := make([]*G2Point, batchSize)
+
+	outC := (*C.BLS12_381_g2_projective_t)(unsafe.Pointer(&out[0]))
+	pointsC := (*C.BLS12_381_g2_affine_t)(unsafe.Pointer(&(*points)[0]))
+	scalarsC := (*C.BLS12_381_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+	msmSizeC := C.size_t(len(*points) / batchSize)
+	deviceIdC := C.size_t(deviceId)
+	batchSizeC := C.size_t(batchSize)
+
+	ret := C.msm_batch_g2_cuda_bls12_381(outC, pointsC, scalarsC, batchSizeC, msmSizeC, deviceIdC)
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_batch_cuda_bls12_381 returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func Commit(d_out, d_scalars, d_points unsafe.Pointer, count, bucketFactor int) int {
+	d_outC := (*C.BLS12_381_projective_t)(d_out)
+	scalarsC := (*C.BLS12_381_scalar_t)(d_scalars)
+	pointsC := (*C.BLS12_381_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	largeBucketFactorC := C.uint(bucketFactor)
+
+	ret := C.commit_cuda_bls12_381(d_outC, scalarsC, pointsC, countC, largeBucketFactorC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}
+
+func CommitG2(d_out, d_scalars, d_points unsafe.Pointer, count, bucketFactor int) int {
+	d_outC := (*C.BLS12_381_g2_projective_t)(d_out)
+	scalarsC := (*C.BLS12_381_scalar_t)(d_scalars)
+	pointsC := (*C.BLS12_381_g2_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	largeBucketFactorC := C.uint(bucketFactor)
+
+	ret := C.commit_g2_cuda_bls12_381(d_outC, scalarsC, pointsC, countC, largeBucketFactorC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}
+
+func CommitBatch(d_out, d_scalars, d_points unsafe.Pointer, count, batch_size int) int {
+	d_outC := (*C.BLS12_381_projective_t)(d_out)
+	scalarsC := (*C.BLS12_381_scalar_t)(d_scalars)
+	pointsC := (*C.BLS12_381_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	batch_sizeC := (C.size_t)(batch_size)
+
+	ret := C.commit_batch_cuda_bls12_381(d_outC, scalarsC, pointsC, countC, batch_sizeC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}
+
+func CommitG2Batch(d_out, d_scalars, d_points unsafe.Pointer, count, batch_size int) int {
+	d_outC := (*C.BLS12_381_g2_projective_t)(d_out)
+	scalarsC := (*C.BLS12_381_scalar_t)(d_scalars)
+	pointsC := (*C.BLS12_381_g2_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	batch_sizeC := (C.size_t)(batch_size)
+
+	ret := C.msm_batch_g2_cuda_bls12_381(d_outC, pointsC, scalarsC, countC, batch_sizeC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}

goicicle/curves/bls12381/msm_test.go

@@ -0,0 +1,355 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12381
+
+import (
+	"fmt"
+	"math"
+	"testing"
+	"time"
+	"unsafe"
+
+	"github.com/ingonyama-zk/icicle/goicicle"
+	"github.com/stretchr/testify/assert"
+)
+
+func GeneratePoints(count int) []G1PointAffine {
+	// Declare a slice of integers
+	var points []G1PointAffine
+
+	// populate the slice
+	for i := 0; i < 10; i++ {
+		var pointProjective G1ProjectivePoint
+		pointProjective.Random()
+
+		var pointAffine G1PointAffine
+		pointAffine.FromProjective(&pointProjective)
+
+		points = append(points, pointAffine)
+	}
+
+	log2_10 := math.Log2(10)
+	log2Count := math.Log2(float64(count))
+	log2Size := int(math.Ceil(log2Count - log2_10))
+
+	for i := 0; i < log2Size; i++ {
+		points = append(points, points...)
+	}
+
+	return points[:count]
+}
+
+func GeneratePointsProj(count int) []G1ProjectivePoint {
+	// Declare a slice of integers
+	var points []G1ProjectivePoint
+	// Use a loop to populate the slice
+	for i := 0; i < count; i++ {
+		var p G1ProjectivePoint
+		p.Random()
+
+		points = append(points, p)
+	}
+
+	return points
+}
+
+func GenerateScalars(count int, skewed bool) []G1ScalarField {
+	// Declare a slice of integers
+	var scalars []G1ScalarField
+
+	var rand G1ScalarField
+	var zero G1ScalarField
+	var one G1ScalarField
+	var randLarge G1ScalarField
+
+	zero.SetZero()
+	one.SetOne()
+	randLarge.Random()
+
+	if skewed && count > 1_200_000 {
+		for i := 0; i < count-1_200_000; i++ {
+			rand.Random()
+			scalars = append(scalars, rand)
+		}
+
+		for i := 0; i < 600_000; i++ {
+			scalars = append(scalars, randLarge)
+		}
+		for i := 0; i < 400_000; i++ {
+			scalars = append(scalars, zero)
+		}
+		for i := 0; i < 200_000; i++ {
+			scalars = append(scalars, one)
+		}
+	} else {
+		for i := 0; i < count; i++ {
+			rand.Random()
+			scalars = append(scalars, rand)
+		}
+	}
+
+	return scalars[:count]
+}
+
+func TestMSM(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1 << v
+
+		points := GeneratePoints(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, true)
+		fmt.Print("Finished generating scalars\n")
+
+		out := new(G1ProjectivePoint)
+		startTime := time.Now()
+		_, e := Msm(out, points, scalars, 0) // non mont
+		fmt.Printf("icicle MSM took: %d ms\n", time.Since(startTime).Milliseconds())
+
+		assert.Equal(t, e, nil, "error should be nil")
+
+		assert.True(t, out.IsOnCurve())
+	}
+}
+
+func TestCommitMSM(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1<<v - 1
+
+		points := GeneratePoints(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, true)
+		fmt.Print("Finished generating scalars\n")
+
+		out_d, _ := goicicle.CudaMalloc(96)
+
+		pointsBytes := count * 64
+		points_d, _ := goicicle.CudaMalloc(pointsBytes)
+		goicicle.CudaMemCpyHtoD[G1PointAffine](points_d, points, pointsBytes)
+
+		scalarBytes := count * 32
+		scalars_d, _ := goicicle.CudaMalloc(scalarBytes)
+		goicicle.CudaMemCpyHtoD[G1ScalarField](scalars_d, scalars, scalarBytes)
+
+		startTime := time.Now()
+		e := Commit(out_d, scalars_d, points_d, count, 10)
+		fmt.Printf("icicle MSM took: %d ms\n", time.Since(startTime).Milliseconds())
+
+		outHost := make([]G1ProjectivePoint, 1)
+		goicicle.CudaMemCpyDtoH[G1ProjectivePoint](outHost, out_d, 96)
+
+		assert.Equal(t, e, 0, "error should be 0")
+		assert.True(t, outHost[0].IsOnCurve())
+	}
+}
+
+func BenchmarkCommit(b *testing.B) {
+	LOG_MSM_SIZES := []int{20, 21, 22, 23, 24, 25, 26}
+
+	for _, logMsmSize := range LOG_MSM_SIZES {
+		msmSize := 1 << logMsmSize
+		points := GeneratePoints(msmSize)
+		scalars := GenerateScalars(msmSize, false)
+
+		out_d, _ := goicicle.CudaMalloc(96)
+
+		pointsBytes := msmSize * 64
+		points_d, _ := goicicle.CudaMalloc(pointsBytes)
+		goicicle.CudaMemCpyHtoD[G1PointAffine](points_d, points, pointsBytes)
+
+		scalarBytes := msmSize * 32
+		scalars_d, _ := goicicle.CudaMalloc(scalarBytes)
+		goicicle.CudaMemCpyHtoD[G1ScalarField](scalars_d, scalars, scalarBytes)
+
+		b.Run(fmt.Sprintf("MSM %d", logMsmSize), func(b *testing.B) {
+			for n := 0; n < b.N; n++ {
+				e := Commit(out_d, scalars_d, points_d, msmSize, 10)
+
+				if e != 0 {
+					panic("Error occured")
+				}
+			}
+		})
+	}
+}
+
+func TestBenchMSM(t *testing.T) {
+	for _, batchPow2 := range []int{2, 4} {
+		for _, pow2 := range []int{4, 6} {
+			msmSize := 1 << pow2
+			batchSize := 1 << batchPow2
+			count := msmSize * batchSize
+
+			points := GeneratePoints(count)
+			scalars := GenerateScalars(count, false)
+
+			a, e := MsmBatch(&points, &scalars, batchSize, 0)
+
+			if e != nil {
+				t.Errorf("MsmBatchBLS12_381 returned an error: %v", e)
+			}
+
+			if len(a) != batchSize {
+				t.Errorf("Expected length %d, but got %d", batchSize, len(a))
+			}
+		}
+	}
+}
+
+func BenchmarkMSM(b *testing.B) {
+	LOG_MSM_SIZES := []int{20, 21, 22, 23, 24, 25, 26}
+
+	for _, logMsmSize := range LOG_MSM_SIZES {
+		msmSize := 1 << logMsmSize
+		points := GeneratePoints(msmSize)
+		scalars := GenerateScalars(msmSize, false)
+		b.Run(fmt.Sprintf("MSM %d", logMsmSize), func(b *testing.B) {
+			for n := 0; n < b.N; n++ {
+				out := new(G1ProjectivePoint)
+				_, e := Msm(out, points, scalars, 0)
+
+				if e != nil {
+					panic("Error occured")
+				}
+			}
+		})
+	}
+}
+
+// G2
+func GenerateG2Points(count int) []G2PointAffine {
+	// Declare a slice of integers
+	var points []G2PointAffine
+
+	// populate the slice
+	for i := 0; i < 10; i++ {
+		var p G2Point
+		p.Random()
+		var affine G2PointAffine
+		affine.FromProjective(&p)
+
+		points = append(points, affine)
+	}
+
+	log2_10 := math.Log2(10)
+	log2Count := math.Log2(float64(count))
+	log2Size := int(math.Ceil(log2Count - log2_10))
+
+	for i := 0; i < log2Size; i++ {
+		points = append(points, points...)
+	}
+
+	return points[:count]
+}
+
+func TestMsmG2BLS12381(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1 << v
+		points := GenerateG2Points(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, false)
+		fmt.Print("Finished generating scalars\n")
+
+		out := new(G2Point)
+		_, e := MsmG2(out, points, scalars, 0)
+		assert.Equal(t, e, nil, "error should be nil")
+		assert.True(t, out.IsOnCurve())
+	}
+}
+
+func BenchmarkMsmG2BLS12381(b *testing.B) {
+	LOG_MSM_SIZES := []int{20, 21, 22, 23, 24, 25, 26}
+
+	for _, logMsmSize := range LOG_MSM_SIZES {
+		msmSize := 1 << logMsmSize
+		points := GenerateG2Points(msmSize)
+		scalars := GenerateScalars(msmSize, false)
+		b.Run(fmt.Sprintf("MSM G2 %d", logMsmSize), func(b *testing.B) {
+			for n := 0; n < b.N; n++ {
+				out := new(G2Point)
+				_, e := MsmG2(out, points, scalars, 0)
+
+				if e != nil {
+					panic("Error occured")
+				}
+			}
+		})
+	}
+}
+
+func TestCommitG2MSM(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1 << v
+
+		points := GenerateG2Points(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, true)
+		fmt.Print("Finished generating scalars\n")
+
+		var sizeCheckG2PointAffine G2PointAffine
+		inputPointsBytes := count * int(unsafe.Sizeof(sizeCheckG2PointAffine))
+
+		var sizeCheckG2Point G2Point
+		out_d, _ := goicicle.CudaMalloc(int(unsafe.Sizeof(sizeCheckG2Point)))
+
+		points_d, _ := goicicle.CudaMalloc(inputPointsBytes)
+		goicicle.CudaMemCpyHtoD[G2PointAffine](points_d, points, inputPointsBytes)
+
+		scalarBytes := count * 32
+		scalars_d, _ := goicicle.CudaMalloc(scalarBytes)
+		goicicle.CudaMemCpyHtoD[G1ScalarField](scalars_d, scalars, scalarBytes)
+
+		startTime := time.Now()
+		e := CommitG2(out_d, scalars_d, points_d, count, 10)
+		fmt.Printf("icicle MSM took: %d ms\n", time.Since(startTime).Milliseconds())
+
+		outHost := make([]G2Point, 1)
+		goicicle.CudaMemCpyDtoH[G2Point](outHost, out_d, int(unsafe.Sizeof(sizeCheckG2Point)))
+
+		assert.Equal(t, e, 0, "error should be 0")
+		assert.Equal(t, len(outHost), 1)
+		result := outHost[0]
+
+		assert.True(t, result.IsOnCurve())
+	}
+}
+
+func TestBatchG2MSM(t *testing.T) {
+	for _, batchPow2 := range []int{2, 4} {
+		for _, pow2 := range []int{4, 6} {
+			msmSize := 1 << pow2
+			batchSize := 1 << batchPow2
+			count := msmSize * batchSize
+
+			points := GenerateG2Points(count)
+			scalars := GenerateScalars(count, false)
+
+			pointsResults, e := MsmG2Batch(&points, &scalars, batchSize, 0)
+
+			if e != nil {
+				t.Errorf("MsmBatchBLS12_381 returned an error: %v", e)
+			}
+
+			if len(pointsResults) != batchSize {
+				t.Errorf("Expected length %d, but got %d", batchSize, len(pointsResults))
+			}
+
+			for _, s := range pointsResults {
+				assert.True(t, s.IsOnCurve())
+			}
+		}
+	}
+}

goicicle/curves/bls12381/ntt.go

@@ -0,0 +1,221 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12381
+
+import (
+	"errors"
+	"fmt"
+	"unsafe"
+
+	"github.com/ingonyama-zk/icicle/goicicle"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbls12_381
+// #include "ntt.h"
+import "C"
+
+const (
+	NONE = 0
+	DIF  = 1
+	DIT  = 2
+)
+
+func Ntt(scalars *[]G1ScalarField, isInverse bool, deviceId int) uint64 {
+	scalarsC := (*C.BLS12_381_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+
+	ret := C.ntt_cuda_bls12_381(scalarsC, C.uint32_t(len(*scalars)), C.bool(isInverse), C.size_t(deviceId))
+
+	return uint64(ret)
+}
+
+func NttBatch(scalars *[]G1ScalarField, isInverse bool, batchSize, deviceId int) uint64 {
+	scalarsC := (*C.BLS12_381_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+	isInverseC := C.bool(isInverse)
+	batchSizeC := C.uint32_t(batchSize)
+	deviceIdC := C.size_t(deviceId)
+
+	ret := C.ntt_batch_cuda_bls12_381(scalarsC, C.uint32_t(len(*scalars)), batchSizeC, isInverseC, deviceIdC)
+
+	return uint64(ret)
+}
+
+func EcNtt(values *[]G1ProjectivePoint, isInverse bool, deviceId int) uint64 {
+	valuesC := (*C.BLS12_381_projective_t)(unsafe.Pointer(&(*values)[0]))
+	deviceIdC := C.size_t(deviceId)
+	isInverseC := C.bool(isInverse)
+	n := C.uint32_t(len(*values))
+
+	ret := C.ecntt_cuda_bls12_381(valuesC, n, isInverseC, deviceIdC)
+
+	return uint64(ret)
+}
+
+func EcNttBatch(values *[]G1ProjectivePoint, isInverse bool, batchSize, deviceId int) uint64 {
+	valuesC := (*C.BLS12_381_projective_t)(unsafe.Pointer(&(*values)[0]))
+	deviceIdC := C.size_t(deviceId)
+	isInverseC := C.bool(isInverse)
+	n := C.uint32_t(len(*values))
+	batchSizeC := C.uint32_t(batchSize)
+
+	ret := C.ecntt_batch_cuda_bls12_381(valuesC, n, batchSizeC, isInverseC, deviceIdC)
+
+	return uint64(ret)
+}
+
+func GenerateTwiddles(d_size int, log_d_size int, inverse bool) (up unsafe.Pointer, err error) {
+	domain_size := C.uint32_t(d_size)
+	logn := C.uint32_t(log_d_size)
+	is_inverse := C.bool(inverse)
+
+	dp := C.build_domain_cuda_bls12_381(domain_size, logn, is_inverse, 0, 0)
+
+	if dp == nil {
+		err = errors.New("nullptr returned from generating twiddles")
+		return unsafe.Pointer(nil), err
+	}
+
+	return unsafe.Pointer(dp), nil
+}
+
+// Reverses d_scalars in-place
+func ReverseScalars(d_scalars unsafe.Pointer, len int) (int, error) {
+	scalarsC := (*C.BLS12_381_scalar_t)(d_scalars)
+	lenC := C.int(len)
+	if success := C.reverse_order_scalars_cuda_bls12_381(scalarsC, lenC, 0, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func Interpolate(scalars, twiddles, cosetPowers unsafe.Pointer, size int, isCoset bool) unsafe.Pointer {
+	size_d := size * 32
+	dp, err := goicicle.CudaMalloc(size_d)
+
+	if err != nil {
+		return nil
+	}
+
+	d_out := (*C.BLS12_381_scalar_t)(dp)
+	scalarsC := (*C.BLS12_381_scalar_t)(scalars)
+	twiddlesC := (*C.BLS12_381_scalar_t)(twiddles)
+	cosetPowersC := (*C.BLS12_381_scalar_t)(cosetPowers)
+	sizeC := C.uint(size)
+
+	var ret C.int
+	if isCoset {
+		ret = C.interpolate_scalars_on_coset_cuda_bls12_381(d_out, scalarsC, twiddlesC, sizeC, cosetPowersC, 0, 0)
+	} else {
+		ret = C.interpolate_scalars_cuda_bls12_381(d_out, scalarsC, twiddlesC, sizeC, 0, 0)
+	}
+	if ret != 0 {
+		fmt.Print("error interpolating")
+	}
+
+	return unsafe.Pointer(d_out)
+}
+
+func Evaluate(scalars_out, scalars, twiddles, coset_powers unsafe.Pointer, scalars_size, twiddles_size int, isCoset bool) int {
+	scalars_outC := (*C.BLS12_381_scalar_t)(scalars_out)
+	scalarsC := (*C.BLS12_381_scalar_t)(scalars)
+	twiddlesC := (*C.BLS12_381_scalar_t)(twiddles)
+	coset_powersC := (*C.BLS12_381_scalar_t)(coset_powers)
+	sizeC := C.uint(scalars_size)
+	twiddlesC_size := C.uint(twiddles_size)
+
+	var ret C.int
+	if isCoset {
+		ret = C.evaluate_scalars_on_coset_cuda_bls12_381(scalars_outC, scalarsC, twiddlesC, twiddlesC_size, sizeC, coset_powersC, 0, 0)
+	} else {
+		ret = C.evaluate_scalars_cuda_bls12_381(scalars_outC, scalarsC, twiddlesC, twiddlesC_size, sizeC, 0, 0)
+	}
+
+	if ret != 0 {
+		fmt.Print("error interpolating")
+		return -1
+	}
+
+	return 0
+}
+
+func VecScalarAdd(in1_d, in2_d unsafe.Pointer, size int) int {
+	in1_dC := (*C.BLS12_381_scalar_t)(in1_d)
+	in2_dC := (*C.BLS12_381_scalar_t)(in2_d)
+	sizeC := C.uint(size)
+
+	ret := C.add_scalars_cuda_bls12_381(in1_dC, in1_dC, in2_dC, sizeC, 0)
+
+	if ret != 0 {
+		fmt.Print("error adding scalar vectors")
+		return -1
+	}
+
+	return 0
+}
+
+func VecScalarSub(in1_d, in2_d unsafe.Pointer, size int) int {
+	in1_dC := (*C.BLS12_381_scalar_t)(in1_d)
+	in2_dC := (*C.BLS12_381_scalar_t)(in2_d)
+	sizeC := C.uint(size)
+
+	ret := C.sub_scalars_cuda_bls12_381(in1_dC, in1_dC, in2_dC, sizeC, 0)
+
+	if ret != 0 {
+		fmt.Print("error subtracting scalar vectors")
+		return -1
+	}
+
+	return 0
+}
+
+func ToMontgomery(d_scalars unsafe.Pointer, len int) (int, error) {
+	scalarsC := (*C.BLS12_381_scalar_t)(d_scalars)
+	lenC := C.uint(len)
+	if success := C.to_montgomery_scalars_cuda_bls12_381(scalarsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func FromMontgomery(d_scalars unsafe.Pointer, len int) (int, error) {
+	scalarsC := (*C.BLS12_381_scalar_t)(d_scalars)
+	lenC := C.uint(len)
+	if success := C.from_montgomery_scalars_cuda_bls12_381(scalarsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func AffinePointFromMontgomery(d_points unsafe.Pointer, len int) (int, error) {
+	pointsC := (*C.BLS12_381_affine_t)(d_points)
+	lenC := C.uint(len)
+
+	if success := C.from_montgomery_aff_points_cuda_bls12_381(pointsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func G2AffinePointFromMontgomery(d_points unsafe.Pointer, len int) (int, error) {
+	pointsC := (*C.BLS12_381_g2_affine_t)(d_points)
+	lenC := C.uint(len)
+
+	if success := C.from_montgomery_aff_points_g2_cuda_bls12_381(pointsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}

goicicle/curves/bls12381/ntt_test.go

@@ -0,0 +1,148 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12381
+
+import (
+	"fmt"
+	"github.com/stretchr/testify/assert"
+	"reflect"
+	"testing"
+)
+
+func TestNttBLS12381BBB(t *testing.T) {
+	count := 1 << 20
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	NttBatch(&nttResult, false, count, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	assert.Equal(t, nttResult, nttResult)
+}
+
+func TestNttBLS12381CompareToGnarkDIF(t *testing.T) {
+	count := 1 << 2
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	Ntt(&nttResult, false, DIF, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	assert.Equal(t, nttResult, nttResult)
+}
+
+func TestINttBLS12381CompareToGnarkDIT(t *testing.T) {
+	count := 1 << 3
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	Ntt(&nttResult, true, DIT, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	assert.Equal(t, nttResult, nttResult)
+}
+
+func TestNttBLS12381(t *testing.T) {
+	count := 1 << 3
+
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	Ntt(&nttResult, false, NONE, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	inttResult := make([]G1ScalarField, len(nttResult))
+	copy(inttResult, nttResult)
+
+	assert.Equal(t, inttResult, nttResult)
+	Ntt(&inttResult, true, NONE, 0)
+	assert.Equal(t, inttResult, scalars)
+}
+
+func TestNttBatchBLS12381(t *testing.T) {
+	count := 1 << 5
+	batches := 4
+
+	scalars := GenerateScalars(count*batches, false)
+
+	var scalarVecOfVec [][]G1ScalarField = make([][]G1ScalarField, 0)
+
+	for i := 0; i < batches; i++ {
+		start := i * count
+		end := (i + 1) * count
+		batch := make([]G1ScalarField, len(scalars[start:end]))
+		copy(batch, scalars[start:end])
+		scalarVecOfVec = append(scalarVecOfVec, batch)
+	}
+
+	nttBatchResult := make([]G1ScalarField, len(scalars))
+	copy(nttBatchResult, scalars)
+
+	NttBatch(&nttBatchResult, false, count, 0)
+
+	var nttResultVecOfVec [][]G1ScalarField
+
+	for i := 0; i < batches; i++ {
+		// Clone the slice
+		clone := make([]G1ScalarField, len(scalarVecOfVec[i]))
+		copy(clone, scalarVecOfVec[i])
+
+		// Add it to the result vector of vectors
+		nttResultVecOfVec = append(nttResultVecOfVec, clone)
+
+		// Call the ntt_bls12_381 function
+		Ntt(&nttResultVecOfVec[i], false, NONE, 0)
+	}
+
+	assert.NotEqual(t, nttBatchResult, scalars)
+
+	// Check that the ntt of each vec of scalars is equal to the intt of the specific batch
+	for i := 0; i < batches; i++ {
+		if !reflect.DeepEqual(nttResultVecOfVec[i], nttBatchResult[i*count:((i+1)*count)]) {
+			t.Errorf("ntt of vec of scalars not equal to intt of specific batch")
+		}
+	}
+}
+
+func BenchmarkNTT(b *testing.B) {
+	LOG_NTT_SIZES := []int{12, 15, 20, 21, 22, 23, 24, 25, 26}
+
+	for _, logNTTSize := range LOG_NTT_SIZES {
+		nttSize := 1 << logNTTSize
+		b.Run(fmt.Sprintf("NTT %d", logNTTSize), func(b *testing.B) {
+			scalars := GenerateScalars(nttSize, false)
+
+			nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+			copy(nttResult, scalars)
+			for n := 0; n < b.N; n++ {
+				Ntt(&nttResult, false, NONE, 0)
+			}
+		})
+	}
+}

goicicle/curves/bls12381/utils.go

@@ -0,0 +1,38 @@
+package bls12381
+
+import "encoding/binary"
+
+// Function to convert [8]uint32 to [4]uint64
+func ConvertUint32ArrToUint64Arr(arr32 [8]uint32) [4]uint64 {
+	var arr64 [4]uint64
+	for i := 0; i < len(arr32); i += 2 {
+		arr64[i/2] = (uint64(arr32[i]) << 32) | uint64(arr32[i+1])
+	}
+	return arr64
+}
+
+func ConvertUint64ArrToUint32Arr4(arr64 [4]uint64) [8]uint32 {
+	var arr32 [8]uint32
+	for i, v := range arr64 {
+		b := make([]byte, 8)
+		binary.LittleEndian.PutUint64(b, v)
+
+		arr32[i*2] = binary.LittleEndian.Uint32(b[0:4])
+		arr32[i*2+1] = binary.LittleEndian.Uint32(b[4:8])
+	}
+
+	return arr32
+}
+
+func ConvertUint64ArrToUint32Arr6(arr64 [6]uint64) [12]uint32 {
+	var arr32 [12]uint32
+	for i, v := range arr64 {
+		b := make([]byte, 8)
+		binary.LittleEndian.PutUint64(b, v)
+
+		arr32[i*2] = binary.LittleEndian.Uint32(b[0:4])
+		arr32[i*2+1] = binary.LittleEndian.Uint32(b[4:8])
+	}
+
+	return arr32
+}

goicicle/curves/bls12381/utils_test.go

@@ -0,0 +1,81 @@
+package bls12381
+
+import (
+	"testing"
+)
+
+func TestConvertUint32ArrToUint64Arr(t *testing.T) {
+	testCases := []struct {
+		name  string
+		input [8]uint32
+		want  [4]uint64
+	}{
+		{
+			name:  "Test with incremental array",
+			input: [8]uint32{1, 2, 3, 4, 5, 6, 7, 8},
+			want:  [4]uint64{4294967298, 12884901892, 21474836486, 30064771080},
+		},
+		{
+			name:  "Test with all zeros",
+			input: [8]uint32{0, 0, 0, 0, 0, 0, 0, 0},
+			want:  [4]uint64{0, 0, 0, 0},
+		},
+		{
+			name:  "Test with maximum uint32 values",
+			input: [8]uint32{4294967295, 4294967295, 4294967295, 4294967295, 4294967295, 4294967295, 4294967295, 4294967295},
+			want:  [4]uint64{18446744073709551615, 18446744073709551615, 18446744073709551615, 18446744073709551615},
+		},
+		{
+			name:  "Test with alternating min and max uint32 values",
+			input: [8]uint32{0, 4294967295, 0, 4294967295, 0, 4294967295, 0, 4294967295},
+			want:  [4]uint64{4294967295, 4294967295, 4294967295, 4294967295},
+		},
+		{
+			name:  "Test with alternating max and min uint32 values",
+			input: [8]uint32{4294967295, 0, 4294967295, 0, 4294967295, 0, 4294967295, 0},
+			want:  [4]uint64{18446744069414584320, 18446744069414584320, 18446744069414584320, 18446744069414584320},
+		},
+	}
+
+	for _, tc := range testCases {
+		t.Run(tc.name, func(t *testing.T) {
+			got := ConvertUint32ArrToUint64Arr(tc.input)
+			if got != tc.want {
+				t.Errorf("got %v, want %v", got, tc.want)
+			}
+		})
+	}
+}
+
+func TestConvertUint64ArrToUint32Arr(t *testing.T) {
+	testCases := []struct {
+		name     string
+		input    [6]uint64
+		expected [12]uint32
+	}{
+		{
+			name:     "test one",
+			input:    [6]uint64{1, 2, 3, 4, 5, 6},
+			expected: [12]uint32{1, 0, 2, 0, 3, 0, 4, 0, 5, 0, 6, 0},
+		},
+		{
+			name:     "test two",
+			input:    [6]uint64{100, 200, 300, 400, 500, 600},
+			expected: [12]uint32{100, 0, 200, 0, 300, 0, 400, 0, 500, 0, 600, 0},
+		},
+		{
+			name:     "test three",
+			input:    [6]uint64{1000, 2000, 3000, 4000, 5000, 6000},
+			expected: [12]uint32{1000, 0, 2000, 0, 3000, 0, 4000, 0, 5000, 0, 6000, 0},
+		},
+	}
+
+	for _, tc := range testCases {
+		t.Run(tc.name, func(t *testing.T) {
+			got := ConvertUint64ArrToUint32Arr6(tc.input)
+			if got != tc.expected {
+				t.Errorf("got %v, want %v", got, tc.expected)
+			}
+		})
+	}
+}

goicicle/curves/bls12381/vec_mod.go

@@ -0,0 +1,41 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bls12381
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbls12_381
+// #include "ve_mod_mult.h"
+import "C"
+import (
+	"fmt"
+	"unsafe"
+)
+
+func VecScalarMulMod(scalarVec1, scalarVec2 unsafe.Pointer, size int) int {
+	scalarVec1C := (*C.BLS12_381_scalar_t)(scalarVec1)
+	scalarVec2C := (*C.BLS12_381_scalar_t)(scalarVec2)
+	sizeC := C.size_t(size)
+
+	ret := C.vec_mod_mult_device_scalar_bls12_381(scalarVec1C, scalarVec2C, sizeC, 0)
+
+	if ret != 0 {
+		fmt.Print("error multiplying scalar vectors")
+		return -1
+	}
+
+	return 0
+}

goicicle/curves/bn254/g1.go

@@ -0,0 +1,324 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bn254
+
+import (
+	"unsafe"
+
+	"encoding/binary"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbn254
+// #include "projective.h"
+// #include "ve_mod_mult.h"
+import "C"
+
+const SCALAR_SIZE = 8
+const BASE_SIZE = 8
+
+type G1ScalarField struct {
+	S [SCALAR_SIZE]uint32
+}
+
+type G1BaseField struct {
+	S [BASE_SIZE]uint32
+}
+
+/*
+ * BaseField Constrctors
+ */
+
+func (f *G1BaseField) SetZero() *G1BaseField {
+	var S [BASE_SIZE]uint32
+	f.S = S
+
+	return f
+}
+
+func (f *G1BaseField) SetOne() *G1BaseField {
+	var S [BASE_SIZE]uint32
+
+	S[0] = 1
+
+	f.S = S
+	return f
+}
+
+func (p *G1ProjectivePoint) FromAffine(affine *G1PointAffine) *G1ProjectivePoint {
+	out := (*C.BN254_projective_t)(unsafe.Pointer(p))
+	in := (*C.BN254_affine_t)(unsafe.Pointer(affine))
+
+	C.projective_from_affine_bn254(out, in)
+
+	return p
+}
+
+func (f *G1BaseField) FromLimbs(limbs [BASE_SIZE]uint32) *G1BaseField {
+	copy(f.S[:], limbs[:])
+
+	return f
+}
+
+/*
+ * BaseField methods
+ */
+
+func (f *G1BaseField) Limbs() [BASE_SIZE]uint32 {
+	return f.S
+}
+
+func (f *G1BaseField) ToBytesLe() []byte {
+	bytes := make([]byte, len(f.S)*4)
+	for i, v := range f.S {
+		binary.LittleEndian.PutUint32(bytes[i*4:], v)
+	}
+
+	return bytes
+}
+
+/*
+ * ScalarField methods
+ */
+
+func (p *G1ScalarField) Random() *G1ScalarField {
+	outC := (*C.BN254_scalar_t)(unsafe.Pointer(p))
+	C.random_scalar_bn254(outC)
+
+	return p
+}
+
+func (f *G1ScalarField) SetZero() *G1ScalarField {
+	var S [SCALAR_SIZE]uint32
+	f.S = S
+
+	return f
+}
+
+func (f *G1ScalarField) SetOne() *G1ScalarField {
+	var S [SCALAR_SIZE]uint32
+	S[0] = 1
+	f.S = S
+
+	return f
+}
+
+func (a *G1ScalarField) Eq(b *G1ScalarField) bool {
+	for i, v := range a.S {
+		if b.S[i] != v {
+			return false
+		}
+	}
+	return true
+}
+
+/*
+ * ScalarField methods
+ */
+
+func (f *G1ScalarField) Limbs() [SCALAR_SIZE]uint32 {
+	return f.S
+}
+
+func (f *G1ScalarField) ToBytesLe() []byte {
+	bytes := make([]byte, len(f.S)*4)
+	for i, v := range f.S {
+		binary.LittleEndian.PutUint32(bytes[i*4:], v)
+	}
+
+	return bytes
+}
+
+/*
+ * PointBN254
+ */
+
+type G1ProjectivePoint struct {
+	X, Y, Z G1BaseField
+}
+
+func (f *G1ProjectivePoint) SetZero() *G1ProjectivePoint {
+	var yOne G1BaseField
+	yOne.SetOne()
+
+	var xZero G1BaseField
+	xZero.SetZero()
+
+	var zZero G1BaseField
+	zZero.SetZero()
+
+	f.X = xZero
+	f.Y = yOne
+	f.Z = zZero
+
+	return f
+}
+
+func (p *G1ProjectivePoint) Eq(pCompare *G1ProjectivePoint) bool {
+	// Cast *PointBN254 to *C.BN254_projective_t
+	// The unsafe.Pointer cast is necessary because Go doesn't allow direct casts
+	// between different pointer types.
+	// It'S your responsibility to ensure that the types are compatible.
+	pC := (*C.BN254_projective_t)(unsafe.Pointer(p))
+	pCompareC := (*C.BN254_projective_t)(unsafe.Pointer(pCompare))
+
+	// Call the C function
+	// The C function doesn't keep any references to the data,
+	// so it'S fine if the Go garbage collector moves or deletes the data later.
+	return bool(C.eq_bn254(pC, pCompareC))
+}
+
+func (p *G1ProjectivePoint) IsOnCurve() bool {
+	point := (*C.BN254_projective_t)(unsafe.Pointer(p))
+	res := C.projective_is_on_curve_bn254(point)
+
+	return bool(res)
+}
+
+func (p *G1ProjectivePoint) Random() *G1ProjectivePoint {
+	outC := (*C.BN254_projective_t)(unsafe.Pointer(p))
+	C.random_projective_bn254(outC)
+
+	return p
+}
+
+func (p *G1ProjectivePoint) StripZ() *G1PointAffine {
+	return &G1PointAffine{
+		X: p.X,
+		Y: p.Y,
+	}
+}
+
+func (p *G1ProjectivePoint) FromLimbs(x, y, z *[]uint32) *G1ProjectivePoint {
+	var _x G1BaseField
+	var _y G1BaseField
+	var _z G1BaseField
+
+	_x.FromLimbs(GetFixedLimbs(x))
+	_y.FromLimbs(GetFixedLimbs(y))
+	_z.FromLimbs(GetFixedLimbs(z))
+
+	p.X = _x
+	p.Y = _y
+	p.Z = _z
+
+	return p
+}
+
+/*
+ * PointAffineNoInfinityBN254
+ */
+
+type G1PointAffine struct {
+	X, Y G1BaseField
+}
+
+func (p *G1PointAffine) FromProjective(projective *G1ProjectivePoint) *G1PointAffine {
+	in := (*C.BN254_projective_t)(unsafe.Pointer(projective))
+	out := (*C.BN254_affine_t)(unsafe.Pointer(p))
+
+	C.projective_to_affine_bn254(out,in)
+
+	return p
+}
+
+func (p *G1PointAffine) ToProjective() *G1ProjectivePoint {
+	var Z G1BaseField
+	Z.SetOne()
+
+	return &G1ProjectivePoint{
+		X: p.X,
+		Y: p.Y,
+		Z: Z,
+	}
+}
+
+func (p *G1PointAffine) FromLimbs(X, Y *[]uint32) *G1PointAffine {
+	var _x G1BaseField
+	var _y G1BaseField
+
+	_x.FromLimbs(GetFixedLimbs(X))
+	_y.FromLimbs(GetFixedLimbs(Y))
+
+	return p
+}
+
+/*
+ * Multiplication
+ */
+
+func MultiplyVec(a []G1ProjectivePoint, b []G1ScalarField, deviceID int) {
+	if len(a) != len(b) {
+		panic("a and b have different lengths")
+	}
+
+	pointsC := (*C.BN254_projective_t)(unsafe.Pointer(&a[0]))
+	scalarsC := (*C.BN254_scalar_t)(unsafe.Pointer(&b[0]))
+	deviceIdC := C.size_t(deviceID)
+	nElementsC := C.size_t(len(a))
+
+	C.vec_mod_mult_point_bn254(pointsC, scalarsC, nElementsC, deviceIdC)
+}
+
+func MultiplyScalar(a []G1ScalarField, b []G1ScalarField, deviceID int) {
+	if len(a) != len(b) {
+		panic("a and b have different lengths")
+	}
+
+	aC := (*C.BN254_scalar_t)(unsafe.Pointer(&a[0]))
+	bC := (*C.BN254_scalar_t)(unsafe.Pointer(&b[0]))
+	deviceIdC := C.size_t(deviceID)
+	nElementsC := C.size_t(len(a))
+
+	C.vec_mod_mult_scalar_bn254(aC, bC, nElementsC, deviceIdC)
+}
+
+// Multiply a matrix by a scalar:
+//
+//	`a` - flattenned matrix;
+//	`b` - vector to multiply `a` by;
+func MultiplyMatrix(a []G1ScalarField, b []G1ScalarField, deviceID int) {
+	c := make([]G1ScalarField, len(b))
+	for i := range c {
+		var p G1ScalarField
+		p.SetZero()
+
+		c[i] = p
+	}
+
+	aC := (*C.BN254_scalar_t)(unsafe.Pointer(&a[0]))
+	bC := (*C.BN254_scalar_t)(unsafe.Pointer(&b[0]))
+	cC := (*C.BN254_scalar_t)(unsafe.Pointer(&c[0]))
+	deviceIdC := C.size_t(deviceID)
+	nElementsC := C.size_t(len(a))
+
+	C.matrix_vec_mod_mult_bn254(aC, bC, cC, nElementsC, deviceIdC)
+}
+
+/*
+ * Utils
+ */
+
+func GetFixedLimbs(slice *[]uint32) [BASE_SIZE]uint32 {
+	if len(*slice) <= BASE_SIZE {
+		limbs := [BASE_SIZE]uint32{}
+		copy(limbs[:len(*slice)], *slice)
+		return limbs
+	}
+
+	panic("slice has too many elements")
+}

goicicle/curves/bn254/g1_test.go

@@ -0,0 +1,198 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bn254
+
+import (
+	"encoding/binary"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestNewFieldBN254One(t *testing.T) {
+	var oneField G1BaseField
+	oneField.SetOne()
+
+	rawOneField := [8]uint32([8]uint32{0x1, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0})
+
+	assert.Equal(t, oneField.S, rawOneField)
+}
+
+func TestNewFieldBN254Zero(t *testing.T) {
+	var zeroField G1BaseField
+	zeroField.SetZero()
+
+	rawZeroField := [8]uint32([8]uint32{0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0})
+
+	assert.Equal(t, zeroField.S, rawZeroField)
+}
+
+func TestFieldBN254ToBytesLe(t *testing.T) {
+	var p G1ProjectivePoint
+	p.Random()
+
+	expected := make([]byte, len(p.X.S)*4) // each uint32 takes 4 bytes
+	for i, v := range p.X.S {
+		binary.LittleEndian.PutUint32(expected[i*4:], v)
+	}
+
+	assert.Equal(t, p.X.ToBytesLe(), expected)
+	assert.Equal(t, len(p.X.ToBytesLe()), 32)
+}
+
+func TestNewPointBN254Zero(t *testing.T) {
+	var pointZero G1ProjectivePoint
+	pointZero.SetZero()
+
+	var baseOne G1BaseField
+	baseOne.SetOne()
+
+	var zeroSanity G1BaseField
+	zeroSanity.SetZero()
+
+	assert.Equal(t, pointZero.X, zeroSanity)
+	assert.Equal(t, pointZero.Y, baseOne)
+	assert.Equal(t, pointZero.Z, zeroSanity)
+}
+
+func TestFromProjectiveToAffine(t *testing.T) {
+	var projective G1ProjectivePoint
+	var affine G1PointAffine
+
+	projective.Random()
+
+	affine.FromProjective(&projective)
+	var projective2 G1ProjectivePoint
+	projective2.FromAffine(&affine)
+
+	assert.True(t, projective.IsOnCurve())
+	assert.True(t, projective2.IsOnCurve())
+	assert.True(t, projective.Eq(&projective2))
+}
+
+func TestBN254Eq(t *testing.T) {
+	var p1 G1ProjectivePoint
+	p1.Random()
+	var p2 G1ProjectivePoint
+	p2.Random()
+
+	assert.Equal(t, p1.Eq(&p1), true)
+	assert.Equal(t, p1.Eq(&p2), false)
+}
+
+func TestBN254StripZ(t *testing.T) {
+	var p1 G1ProjectivePoint
+	p1.Random()
+
+	p2ZLess := p1.StripZ()
+
+	assert.IsType(t, G1PointAffine{}, *p2ZLess)
+	assert.Equal(t, p1.X, p2ZLess.X)
+	assert.Equal(t, p1.Y, p2ZLess.Y)
+}
+
+func TestPointBN254fromLimbs(t *testing.T) {
+	var p G1ProjectivePoint
+	p.Random()
+
+	x := p.X.Limbs()
+	y := p.Y.Limbs()
+	z := p.Z.Limbs()
+
+	xSlice := x[:]
+	ySlice := y[:]
+	zSlice := z[:]
+
+	var pFromLimbs G1ProjectivePoint
+	pFromLimbs.FromLimbs(&xSlice, &ySlice, &zSlice)
+
+	assert.Equal(t, pFromLimbs, p)
+}
+
+func TestNewPointAffineNoInfinityBN254Zero(t *testing.T) {
+	var zeroP G1PointAffine
+
+	var zeroSanity G1BaseField
+	zeroSanity.SetZero()
+
+	assert.Equal(t, zeroP.X, zeroSanity)
+	assert.Equal(t, zeroP.Y, zeroSanity)
+}
+
+func TestPointAffineNoInfinityBN254FromLimbs(t *testing.T) {
+	// Initialize your test values
+	x := [8]uint32{1, 2, 3, 4, 5, 6, 7, 8}
+	y := [8]uint32{9, 10, 11, 12, 13, 14, 15, 16}
+	xSlice := x[:]
+	ySlice := y[:]
+
+	// Execute your function
+	var result G1PointAffine
+	result.FromLimbs(&xSlice, &ySlice)
+
+	var xBase G1BaseField
+	var yBase G1BaseField
+	xBase.FromLimbs(x)
+	yBase.FromLimbs(y)
+
+	// Define your expected result
+	expected := &G1PointAffine{
+		X: xBase,
+		Y: yBase,
+	}
+
+	// Test if result is as expected
+	assert.Equal(t, result, expected)
+}
+
+func TestGetFixedLimbs(t *testing.T) {
+	t.Run("case of valid input of length less than 8", func(t *testing.T) {
+		slice := []uint32{1, 2, 3, 4, 5, 6, 7}
+		expected := [8]uint32{1, 2, 3, 4, 5, 6, 7, 0}
+
+		result := GetFixedLimbs(&slice)
+		assert.Equal(t, result, expected)
+	})
+
+	t.Run("case of valid input of length 8", func(t *testing.T) {
+		slice := []uint32{1, 2, 3, 4, 5, 6, 7, 8}
+		expected := [8]uint32{1, 2, 3, 4, 5, 6, 7, 8}
+
+		result := GetFixedLimbs(&slice)
+		assert.Equal(t, result, expected)
+	})
+
+	t.Run("case of empty input", func(t *testing.T) {
+		slice := []uint32{}
+		expected := [8]uint32{0, 0, 0, 0, 0, 0, 0, 0}
+
+		result := GetFixedLimbs(&slice)
+		assert.Equal(t, result, expected)
+	})
+
+	t.Run("case of input length greater than 8", func(t *testing.T) {
+		slice := []uint32{1, 2, 3, 4, 5, 6, 7, 8, 9}
+
+		defer func() {
+			if r := recover(); r == nil {
+				t.Errorf("the code did not panic")
+			}
+		}()
+
+		GetFixedLimbs(&slice)
+	})
+}

goicicle/curves/bn254/g2.go

@@ -0,0 +1,113 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bn254
+
+import (
+	"encoding/binary"
+	"unsafe"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbn254
+// #include "projective.h"
+// #include "ve_mod_mult.h"
+import "C"
+
+// G2 extension field
+
+type G2Element [4]uint64
+
+type ExtentionField struct {
+	A0, A1 G2Element
+}
+
+type G2PointAffine struct {
+	X, Y ExtentionField
+}
+
+type G2Point struct {
+	X, Y, Z ExtentionField
+}
+
+
+func (p *G2Point) Random() *G2Point {
+	outC := (*C.BN254_g2_projective_t)(unsafe.Pointer(p))
+	C.random_g2_projective_bn254(outC)
+
+	return p
+}
+
+func (p *G2Point) FromAffine(affine *G2PointAffine) *G2Point {
+	out := (*C.BN254_g2_projective_t)(unsafe.Pointer(p))
+	in := (*C.BN254_g2_affine_t)(unsafe.Pointer(affine))
+
+	C.g2_projective_from_affine_bn254(out, in)
+
+	return p
+}
+
+func (p *G2Point) Eq(pCompare *G2Point) bool {
+	// Cast *PointBN254 to *C.BN254_projective_t
+	// The unsafe.Pointer cast is necessary because Go doesn't allow direct casts
+	// between different pointer types.
+	// It's your responsibility to ensure that the types are compatible.
+	pC := (*C.BN254_g2_projective_t)(unsafe.Pointer(p))
+	pCompareC := (*C.BN254_g2_projective_t)(unsafe.Pointer(pCompare))
+
+	// Call the C function
+	// The C function doesn't keep any references to the data,
+	// so it's fine if the Go garbage collector moves or deletes the data later.
+	return bool(C.eq_g2_bn254(pC, pCompareC))
+}
+
+func (f *G2Element) ToBytesLe() []byte {
+	var bytes []byte
+	for _, val := range f {
+		buf := make([]byte, 8) // 8 bytes because uint64 is 64-bit
+		binary.LittleEndian.PutUint64(buf, val)
+		bytes = append(bytes, buf...)
+	}
+	return bytes
+}
+
+func (p *G2PointAffine) ToProjective() G2Point {
+	return G2Point{
+		X: p.X,
+		Y: p.Y,
+		Z: ExtentionField{
+			A0: G2Element{1, 0, 0, 0},
+			A1: G2Element{0, 0, 0, 0},
+		},
+	}
+}
+
+func (p *G2PointAffine) FromProjective(projective *G2Point) *G2PointAffine {
+	out := (*C.BN254_g2_affine_t)(unsafe.Pointer(p))
+	in := (*C.BN254_g2_projective_t)(unsafe.Pointer(projective))
+
+	C.g2_projective_to_affine_bn254(out, in)
+
+	return p
+}
+
+func (p *G2Point) IsOnCurve() bool {
+	// Directly copy memory from the C struct to the Go struct
+	point := (*C.BN254_g2_projective_t)(unsafe.Pointer(p))
+	res := C.g2_projective_is_on_curve_bn254(point)
+
+	return bool(res)
+}

goicicle/curves/bn254/g2_test.go

@@ -0,0 +1,76 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bn254
+
+import (
+	"github.com/stretchr/testify/assert"
+	"testing"
+)
+
+func TestG2Eqg2(t *testing.T) {
+	var point G2Point
+
+	point.Random()
+
+	assert.True(t, point.Eq(&point))
+}
+
+func TestG2FromProjectiveToAffine(t *testing.T) {
+	var projective G2Point
+	var affine G2PointAffine
+
+	projective.Random()
+
+	affine.FromProjective(&projective)
+	var projective2 G2Point
+	projective2.FromAffine(&affine)
+
+	
+	assert.True(t, projective.IsOnCurve())
+	assert.True(t, projective2.IsOnCurve())
+	assert.True(t, projective.Eq(&projective2))
+}
+
+func TestG2Eqg2NotEqual(t *testing.T) {
+	var point G2Point
+	point.Random()
+
+	var point2 G2Point
+	point2.Random()
+
+	assert.False(t, point.Eq(&point2))
+}
+
+func TestG2ToBytes(t *testing.T) {
+	element := G2Element{0x6546098ea84b6298, 0x4a384533d1f68aca, 0xaa0666972d771336, 0x1569e4a34321993}
+	bytes := element.ToBytesLe()
+
+	assert.Equal(t, bytes, []byte{0x98, 0x62, 0x4b, 0xa8, 0x8e, 0x9, 0x46, 0x65, 0xca, 0x8a, 0xf6, 0xd1, 0x33, 0x45, 0x38, 0x4a, 0x36, 0x13, 0x77, 0x2d, 0x97, 0x66, 0x6, 0xaa, 0x93, 0x19, 0x32, 0x34, 0x4a, 0x9e, 0x56, 0x1})
+}
+
+func TestG2ShouldConvertToProjective(t *testing.T) {
+	var pointProjective G2Point
+	var pointAffine G2PointAffine
+
+	pointProjective.Random()
+	pointAffine.FromProjective(&pointProjective)
+
+	proj := pointAffine.ToProjective()
+
+	assert.True(t, proj.IsOnCurve())
+	assert.True(t, pointProjective.Eq(&proj))
+}

goicicle/curves/bn254/include/msm.h

@@ -0,0 +1,94 @@
+
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+#include <cuda.h>
+#include <cuda_runtime.h>
+#include <stdbool.h>
+// msm.h
+
+#ifndef _BN254_MSM_H
+#define _BN254_MSM_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+// Incomplete declaration of BN254 projective and affine structs
+typedef struct BN254_projective_t BN254_projective_t;
+typedef struct BN254_g2_projective_t BN254_g2_projective_t;
+typedef struct BN254_affine_t BN254_affine_t;
+typedef struct BN254_g2_affine_t BN254_g2_affine_t;
+typedef struct BN254_scalar_t BN254_scalar_t;
+typedef cudaStream_t CudaStream_t;
+
+int msm_cuda_bn254(
+  BN254_projective_t* out, BN254_affine_t* points, BN254_scalar_t* scalars, size_t count, size_t device_id);
+
+int msm_batch_cuda_bn254(
+  BN254_projective_t* out,
+  BN254_affine_t* points,
+  BN254_scalar_t* scalars,
+  size_t batch_size,
+  size_t msm_size,
+  size_t device_id);
+
+int commit_cuda_bn254(
+  BN254_projective_t* d_out,
+  BN254_scalar_t* d_scalars,
+  BN254_affine_t* d_points,
+  size_t count,
+  unsigned large_bucket_factor,
+  size_t device_id);
+
+int commit_batch_cuda_bn254(
+  BN254_projective_t* d_out,
+  BN254_scalar_t* d_scalars,
+  BN254_affine_t* d_points,
+  size_t count,
+  size_t batch_size,
+  size_t device_id);
+
+int msm_g2_cuda_bn254(
+  BN254_g2_projective_t* out, BN254_g2_affine_t* points, BN254_scalar_t* scalars, size_t count, size_t device_id);
+int msm_batch_g2_cuda_bn254(
+  BN254_g2_projective_t* out,
+  BN254_g2_affine_t* points,
+  BN254_scalar_t* scalars,
+  size_t batch_size,
+  size_t msm_size,
+  size_t device_id);
+int commit_g2_cuda_bn254(
+  BN254_g2_projective_t* d_out,
+  BN254_scalar_t* d_scalars,
+  BN254_g2_affine_t* d_points,
+  size_t count,
+  unsigned large_bucket_factor,
+  size_t device_id);
+int commit_batch_g2_cuda_bn254(
+  BN254_g2_projective_t* d_out,
+  BN254_scalar_t* d_scalars,
+  BN254_g2_affine_t* d_points,
+  size_t count,
+  size_t batch_size,
+  size_t device_id,
+  cudaStream_t stream);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _BN254_MSM_H */

goicicle/curves/bn254/include/ntt.h

@@ -0,0 +1,193 @@
+
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+#include <cuda.h>
+#include <stdbool.h>
+// ntt.h
+
+#ifndef _BN254_NTT_H
+#define _BN254_NTT_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+// Incomplete declaration of BN254 projective and affine structs
+typedef struct BN254_projective_t BN254_projective_t;
+typedef struct BN254_affine_t BN254_affine_t;
+typedef struct BN254_scalar_t BN254_scalar_t;
+
+typedef struct BN254_g2_projective_t BN254_g2_projective_t;
+typedef struct BN254_g2_affine_t BN254_g2_affine_t;
+
+int ntt_cuda_bn254(BN254_scalar_t* arr, uint32_t n, bool inverse, size_t device_id);
+int ntt_batch_cuda_bn254(BN254_scalar_t* arr, uint32_t arr_size, uint32_t batch_size, bool inverse, size_t device_id);
+
+int ecntt_cuda_bn254(BN254_projective_t* arr, uint32_t n, bool inverse, size_t device_id);
+int ecntt_batch_cuda_bn254(
+  BN254_projective_t* arr, uint32_t arr_size, uint32_t batch_size, bool inverse, size_t device_id);
+
+BN254_scalar_t*
+build_domain_cuda_bn254(uint32_t domain_size, uint32_t logn, bool inverse, size_t device_id, size_t stream);
+int interpolate_scalars_cuda_bn254(
+  BN254_scalar_t* d_out,
+  BN254_scalar_t* d_evaluations,
+  BN254_scalar_t* d_domain,
+  unsigned n,
+  unsigned device_id,
+  size_t stream);
+int interpolate_scalars_batch_cuda_bn254(
+  BN254_scalar_t* d_out,
+  BN254_scalar_t* d_evaluations,
+  BN254_scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int interpolate_points_cuda_bn254(
+  BN254_projective_t* d_out,
+  BN254_projective_t* d_evaluations,
+  BN254_scalar_t* d_domain,
+  unsigned n,
+  size_t device_id,
+  size_t stream);
+int interpolate_points_batch_cuda_bn254(
+  BN254_projective_t* d_out,
+  BN254_projective_t* d_evaluations,
+  BN254_scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int interpolate_scalars_on_coset_cuda_bn254(
+  BN254_scalar_t* d_out,
+  BN254_scalar_t* d_evaluations,
+  BN254_scalar_t* d_domain,
+  unsigned n,
+  BN254_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int interpolate_scalars_batch_on_coset_cuda_bn254(
+  BN254_scalar_t* d_out,
+  BN254_scalar_t* d_evaluations,
+  BN254_scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  BN254_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int evaluate_scalars_cuda_bn254(
+  BN254_scalar_t* d_out,
+  BN254_scalar_t* d_coefficients,
+  BN254_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned device_id,
+  size_t stream);
+int evaluate_scalars_batch_cuda_bn254(
+  BN254_scalar_t* d_out,
+  BN254_scalar_t* d_coefficients,
+  BN254_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_cuda_bn254(
+  BN254_projective_t* d_out,
+  BN254_projective_t* d_coefficients,
+  BN254_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_batch_cuda_bn254(
+  BN254_projective_t* d_out,
+  BN254_projective_t* d_coefficients,
+  BN254_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int evaluate_scalars_on_coset_cuda_bn254(
+  BN254_scalar_t* d_out,
+  BN254_scalar_t* d_coefficients,
+  BN254_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  BN254_scalar_t* coset_powers,
+  unsigned device_id,
+  size_t stream);
+int evaluate_scalars_on_coset_batch_cuda_bn254(
+  BN254_scalar_t* d_out,
+  BN254_scalar_t* d_coefficients,
+  BN254_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  BN254_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_on_coset_cuda_bn254(
+  BN254_projective_t* d_out,
+  BN254_projective_t* d_coefficients,
+  BN254_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  BN254_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_on_coset_batch_cuda_bn254(
+  BN254_projective_t* d_out,
+  BN254_projective_t* d_coefficients,
+  BN254_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  BN254_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int reverse_order_scalars_cuda_bn254(BN254_scalar_t* arr, int n, size_t device_id, size_t stream);
+int reverse_order_scalars_batch_cuda_bn254(BN254_scalar_t* arr, int n, int batch_size, size_t device_id, size_t stream);
+int reverse_order_points_cuda_bn254(BN254_projective_t* arr, int n, size_t device_id, size_t stream);
+int reverse_order_points_batch_cuda_bn254(
+  BN254_projective_t* arr, int n, int batch_size, size_t device_id, size_t stream);
+int add_scalars_cuda_bn254(
+  BN254_scalar_t* d_out, BN254_scalar_t* d_in1, BN254_scalar_t* d_in2, unsigned n, size_t stream);
+int sub_scalars_cuda_bn254(
+  BN254_scalar_t* d_out, BN254_scalar_t* d_in1, BN254_scalar_t* d_in2, unsigned n, size_t stream);
+int to_montgomery_scalars_cuda_bn254(BN254_scalar_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_scalars_cuda_bn254(BN254_scalar_t* d_inout, unsigned n, size_t stream);
+
+// points g1
+int to_montgomery_proj_points_cuda_bn254(BN254_projective_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_proj_points_cuda_bn254(BN254_projective_t* d_inout, unsigned n, size_t stream);
+int to_montgomery_aff_points_cuda_bn254(BN254_affine_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_aff_points_cuda_bn254(BN254_affine_t* d_inout, unsigned n, size_t stream);
+
+// points g2
+int to_montgomery_proj_points_g2_cuda_bn254(BN254_g2_projective_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_proj_points_g2_cuda_bn254(BN254_g2_projective_t* d_inout, unsigned n, size_t stream);
+int to_montgomery_aff_points_g2_cuda_bn254(BN254_g2_affine_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_aff_points_g2_cuda_bn254(BN254_g2_affine_t* d_inout, unsigned n, size_t stream);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _BN254_NTT_H */

goicicle/curves/bn254/include/projective.h

@@ -0,0 +1,50 @@
+
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+#include <cuda.h>
+#include <stdbool.h>
+// projective.h
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+typedef struct BN254_projective_t BN254_projective_t;
+typedef struct BN254_g2_projective_t BN254_g2_projective_t;
+typedef struct BN254_affine_t BN254_affine_t;
+typedef struct BN254_g2_affine_t BN254_g2_affine_t;
+typedef struct BN254_scalar_t BN254_scalar_t;
+
+bool projective_is_on_curve_bn254(BN254_projective_t* point1);
+
+int random_scalar_bn254(BN254_scalar_t* out); 
+int random_projective_bn254(BN254_projective_t* out);
+BN254_projective_t* projective_zero_bn254();
+int projective_to_affine_bn254(BN254_affine_t* out, BN254_projective_t* point1);
+int projective_from_affine_bn254(BN254_projective_t* out, BN254_affine_t* point1);
+
+int random_g2_projective_bn254(BN254_g2_projective_t* out);
+int g2_projective_to_affine_bn254(BN254_g2_affine_t* out, BN254_g2_projective_t* point1);
+int g2_projective_from_affine_bn254(BN254_g2_projective_t* out, BN254_g2_affine_t* point1);
+bool g2_projective_is_on_curve_bn254(BN254_g2_projective_t* point1);
+
+bool eq_bn254(BN254_projective_t* point1, BN254_projective_t* point2);
+bool eq_g2_bn254(BN254_g2_projective_t* point1, BN254_g2_projective_t* point2);
+
+#ifdef __cplusplus
+}
+#endif

goicicle/curves/bn254/include/ve_mod_mult.h

@@ -0,0 +1,45 @@
+
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+#include <cuda.h>
+#include <stdbool.h>
+// ve_mod_mult.h
+
+#ifndef _BN254_VEC_MULT_H
+#define _BN254_VEC_MULT_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+typedef struct BN254_projective_t BN254_projective_t;
+typedef struct BN254_scalar_t BN254_scalar_t;
+
+int32_t
+vec_mod_mult_point_bn254(BN254_projective_t* inout, BN254_scalar_t* scalar_vec, size_t n_elments, size_t device_id);
+int32_t
+vec_mod_mult_scalar_bn254(BN254_scalar_t* inout, BN254_scalar_t* scalar_vec, size_t n_elments, size_t device_id);
+int32_t vec_mod_mult_device_scalar_bn254(
+  BN254_scalar_t* inout, BN254_scalar_t* scalar_vec, size_t n_elements, size_t device_id);
+int32_t matrix_vec_mod_mult_bn254(
+  BN254_scalar_t* matrix_flattened, BN254_scalar_t* input, BN254_scalar_t* output, size_t n_elments, size_t device_id);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _BN254_VEC_MULT_H */

goicicle/curves/bn254/msm.go

@@ -0,0 +1,208 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bn254
+
+import (
+	"errors"
+	"fmt"
+	"unsafe"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbn254
+// #include "msm.h"
+import "C"
+
+func Msm(out *G1ProjectivePoint, points []G1PointAffine, scalars []G1ScalarField, device_id int) (*G1ProjectivePoint, error) {
+	if len(points) != len(scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	pointsC := (*C.BN254_affine_t)(unsafe.Pointer(&points[0]))
+	scalarsC := (*C.BN254_scalar_t)(unsafe.Pointer(&scalars[0]))
+	outC := (*C.BN254_projective_t)(unsafe.Pointer(out))
+	ret := C.msm_cuda_bn254(outC, pointsC, scalarsC, C.size_t(len(points)), C.size_t(device_id))
+
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_cuda_bn254 returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func MsmG2(out *G2Point, points []G2PointAffine, scalars []G1ScalarField, device_id int) (*G2Point, error) {
+	if len(points) != len(scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	pointsC := (*C.BN254_g2_affine_t)(unsafe.Pointer(&points[0]))
+	scalarsC := (*C.BN254_scalar_t)(unsafe.Pointer(&scalars[0]))
+	outC := (*C.BN254_g2_projective_t)(unsafe.Pointer(out))
+
+	ret := C.msm_g2_cuda_bn254(outC, pointsC, scalarsC, C.size_t(len(points)), C.size_t(device_id))
+
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_g2_cuda_bn254 returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func MsmBatch(points *[]G1PointAffine, scalars *[]G1ScalarField, batchSize, deviceId int) ([]*G1ProjectivePoint, error) {
+	// Check for nil pointers
+	if points == nil || scalars == nil {
+		return nil, errors.New("points or scalars is nil")
+	}
+
+	if len(*points) != len(*scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	// Check for empty slices
+	if len(*points) == 0 || len(*scalars) == 0 {
+		return nil, errors.New("points or scalars is empty")
+	}
+
+	// Check for zero batchSize
+	if batchSize <= 0 {
+		return nil, errors.New("error on: batchSize must be greater than zero")
+	}
+
+	out := make([]*G1ProjectivePoint, batchSize)
+
+	for i := 0; i < len(out); i++ {
+		var p G1ProjectivePoint
+		p.SetZero()
+
+		out[i] = &p
+	}
+
+	outC := (*C.BN254_projective_t)(unsafe.Pointer(&out[0]))
+	pointsC := (*C.BN254_affine_t)(unsafe.Pointer(&(*points)[0]))
+	scalarsC := (*C.BN254_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+	msmSizeC := C.size_t(len(*points) / batchSize)
+	deviceIdC := C.size_t(deviceId)
+	batchSizeC := C.size_t(batchSize)
+
+	ret := C.msm_batch_cuda_bn254(outC, pointsC, scalarsC, batchSizeC, msmSizeC, deviceIdC)
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_batch_cuda_bn254 returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func MsmG2Batch(points *[]G2PointAffine, scalars *[]G1ScalarField, batchSize, deviceId int) ([]*G2Point, error) {
+	// Check for nil pointers
+	if points == nil || scalars == nil {
+		return nil, errors.New("points or scalars is nil")
+	}
+
+	if len(*points) != len(*scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	// Check for empty slices
+	if len(*points) == 0 || len(*scalars) == 0 {
+		return nil, errors.New("points or scalars is empty")
+	}
+
+	// Check for zero batchSize
+	if batchSize <= 0 {
+		return nil, errors.New("error on: batchSize must be greater than zero")
+	}
+
+	out := make([]*G2Point, batchSize)
+
+	outC := (*C.BN254_g2_projective_t)(unsafe.Pointer(&out[0]))
+	pointsC := (*C.BN254_g2_affine_t)(unsafe.Pointer(&(*points)[0]))
+	scalarsC := (*C.BN254_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+	msmSizeC := C.size_t(len(*points) / batchSize)
+	deviceIdC := C.size_t(deviceId)
+	batchSizeC := C.size_t(batchSize)
+
+	ret := C.msm_batch_g2_cuda_bn254(outC, pointsC, scalarsC, batchSizeC, msmSizeC, deviceIdC)
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_batch_cuda_bn254 returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func Commit(d_out, d_scalars, d_points unsafe.Pointer, count, bucketFactor int) int {
+	d_outC := (*C.BN254_projective_t)(d_out)
+	scalarsC := (*C.BN254_scalar_t)(d_scalars)
+	pointsC := (*C.BN254_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	largeBucketFactorC := C.uint(bucketFactor)
+
+	ret := C.commit_cuda_bn254(d_outC, scalarsC, pointsC, countC, largeBucketFactorC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}
+
+func CommitG2(d_out, d_scalars, d_points unsafe.Pointer, count, bucketFactor int) int {
+	d_outC := (*C.BN254_g2_projective_t)(d_out)
+	scalarsC := (*C.BN254_scalar_t)(d_scalars)
+	pointsC := (*C.BN254_g2_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	largeBucketFactorC := C.uint(bucketFactor)
+
+	ret := C.commit_g2_cuda_bn254(d_outC, scalarsC, pointsC, countC, largeBucketFactorC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}
+
+func CommitBatch(d_out, d_scalars, d_points unsafe.Pointer, count, batch_size int) int {
+	d_outC := (*C.BN254_projective_t)(d_out)
+	scalarsC := (*C.BN254_scalar_t)(d_scalars)
+	pointsC := (*C.BN254_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	batch_sizeC := (C.size_t)(batch_size)
+
+	ret := C.commit_batch_cuda_bn254(d_outC, scalarsC, pointsC, countC, batch_sizeC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}
+
+func CommitG2Batch(d_out, d_scalars, d_points unsafe.Pointer, count, batch_size int) int {
+	d_outC := (*C.BN254_g2_projective_t)(d_out)
+	scalarsC := (*C.BN254_scalar_t)(d_scalars)
+	pointsC := (*C.BN254_g2_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	batch_sizeC := (C.size_t)(batch_size)
+
+	ret := C.msm_batch_g2_cuda_bn254(d_outC, pointsC, scalarsC, countC, batch_sizeC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}

goicicle/curves/bn254/msm_test.go

@@ -0,0 +1,355 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bn254
+
+import (
+	"fmt"
+	"math"
+	"testing"
+	"time"
+	"unsafe"
+
+	"github.com/ingonyama-zk/icicle/goicicle"
+	"github.com/stretchr/testify/assert"
+)
+
+func GeneratePoints(count int) []G1PointAffine {
+	// Declare a slice of integers
+	var points []G1PointAffine
+
+	// populate the slice
+	for i := 0; i < 10; i++ {
+		var pointProjective G1ProjectivePoint
+		pointProjective.Random()
+
+		var pointAffine G1PointAffine
+		pointAffine.FromProjective(&pointProjective)
+
+		points = append(points, pointAffine)
+	}
+
+	log2_10 := math.Log2(10)
+	log2Count := math.Log2(float64(count))
+	log2Size := int(math.Ceil(log2Count - log2_10))
+
+	for i := 0; i < log2Size; i++ {
+		points = append(points, points...)
+	}
+
+	return points[:count]
+}
+
+func GeneratePointsProj(count int) []G1ProjectivePoint {
+	// Declare a slice of integers
+	var points []G1ProjectivePoint
+	// Use a loop to populate the slice
+	for i := 0; i < count; i++ {
+		var p G1ProjectivePoint
+		p.Random()
+
+		points = append(points, p)
+	}
+
+	return points
+}
+
+func GenerateScalars(count int, skewed bool) []G1ScalarField {
+	// Declare a slice of integers
+	var scalars []G1ScalarField
+
+	var rand G1ScalarField
+	var zero G1ScalarField
+	var one G1ScalarField
+	var randLarge G1ScalarField
+
+	zero.SetZero()
+	one.SetOne()
+	randLarge.Random()
+
+	if skewed && count > 1_200_000 {
+		for i := 0; i < count-1_200_000; i++ {
+			rand.Random()
+			scalars = append(scalars, rand)
+		}
+
+		for i := 0; i < 600_000; i++ {
+			scalars = append(scalars, randLarge)
+		}
+		for i := 0; i < 400_000; i++ {
+			scalars = append(scalars, zero)
+		}
+		for i := 0; i < 200_000; i++ {
+			scalars = append(scalars, one)
+		}
+	} else {
+		for i := 0; i < count; i++ {
+			rand.Random()
+			scalars = append(scalars, rand)
+		}
+	}
+
+	return scalars[:count]
+}
+
+func TestMSM(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1 << v
+
+		points := GeneratePoints(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, true)
+		fmt.Print("Finished generating scalars\n")
+
+		out := new(G1ProjectivePoint)
+		startTime := time.Now()
+		_, e := Msm(out, points, scalars, 0) // non mont
+		fmt.Printf("icicle MSM took: %d ms\n", time.Since(startTime).Milliseconds())
+
+		assert.Equal(t, e, nil, "error should be nil")
+
+		assert.True(t, out.IsOnCurve())
+	}
+}
+
+func TestCommitMSM(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1<<v - 1
+
+		points := GeneratePoints(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, true)
+		fmt.Print("Finished generating scalars\n")
+
+		out_d, _ := goicicle.CudaMalloc(96)
+
+		pointsBytes := count * 64
+		points_d, _ := goicicle.CudaMalloc(pointsBytes)
+		goicicle.CudaMemCpyHtoD[G1PointAffine](points_d, points, pointsBytes)
+
+		scalarBytes := count * 32
+		scalars_d, _ := goicicle.CudaMalloc(scalarBytes)
+		goicicle.CudaMemCpyHtoD[G1ScalarField](scalars_d, scalars, scalarBytes)
+
+		startTime := time.Now()
+		e := Commit(out_d, scalars_d, points_d, count, 10)
+		fmt.Printf("icicle MSM took: %d ms\n", time.Since(startTime).Milliseconds())
+
+		outHost := make([]G1ProjectivePoint, 1)
+		goicicle.CudaMemCpyDtoH[G1ProjectivePoint](outHost, out_d, 96)
+
+		assert.Equal(t, e, 0, "error should be 0")
+		assert.True(t, outHost[0].IsOnCurve())
+	}
+}
+
+func BenchmarkCommit(b *testing.B) {
+	LOG_MSM_SIZES := []int{20, 21, 22, 23, 24, 25, 26}
+
+	for _, logMsmSize := range LOG_MSM_SIZES {
+		msmSize := 1 << logMsmSize
+		points := GeneratePoints(msmSize)
+		scalars := GenerateScalars(msmSize, false)
+
+		out_d, _ := goicicle.CudaMalloc(96)
+
+		pointsBytes := msmSize * 64
+		points_d, _ := goicicle.CudaMalloc(pointsBytes)
+		goicicle.CudaMemCpyHtoD[G1PointAffine](points_d, points, pointsBytes)
+
+		scalarBytes := msmSize * 32
+		scalars_d, _ := goicicle.CudaMalloc(scalarBytes)
+		goicicle.CudaMemCpyHtoD[G1ScalarField](scalars_d, scalars, scalarBytes)
+
+		b.Run(fmt.Sprintf("MSM %d", logMsmSize), func(b *testing.B) {
+			for n := 0; n < b.N; n++ {
+				e := Commit(out_d, scalars_d, points_d, msmSize, 10)
+
+				if e != 0 {
+					panic("Error occured")
+				}
+			}
+		})
+	}
+}
+
+func TestBenchMSM(t *testing.T) {
+	for _, batchPow2 := range []int{2, 4} {
+		for _, pow2 := range []int{4, 6} {
+			msmSize := 1 << pow2
+			batchSize := 1 << batchPow2
+			count := msmSize * batchSize
+
+			points := GeneratePoints(count)
+			scalars := GenerateScalars(count, false)
+
+			a, e := MsmBatch(&points, &scalars, batchSize, 0)
+
+			if e != nil {
+				t.Errorf("MsmBatchBN254 returned an error: %v", e)
+			}
+
+			if len(a) != batchSize {
+				t.Errorf("Expected length %d, but got %d", batchSize, len(a))
+			}
+		}
+	}
+}
+
+func BenchmarkMSM(b *testing.B) {
+	LOG_MSM_SIZES := []int{20, 21, 22, 23, 24, 25, 26}
+
+	for _, logMsmSize := range LOG_MSM_SIZES {
+		msmSize := 1 << logMsmSize
+		points := GeneratePoints(msmSize)
+		scalars := GenerateScalars(msmSize, false)
+		b.Run(fmt.Sprintf("MSM %d", logMsmSize), func(b *testing.B) {
+			for n := 0; n < b.N; n++ {
+				out := new(G1ProjectivePoint)
+				_, e := Msm(out, points, scalars, 0)
+
+				if e != nil {
+					panic("Error occured")
+				}
+			}
+		})
+	}
+}
+
+// G2
+func GenerateG2Points(count int) []G2PointAffine {
+	// Declare a slice of integers
+	var points []G2PointAffine
+
+	// populate the slice
+	for i := 0; i < 10; i++ {
+		var p G2Point
+		p.Random()
+		var affine G2PointAffine
+		affine.FromProjective(&p)
+
+		points = append(points, affine)
+	}
+
+	log2_10 := math.Log2(10)
+	log2Count := math.Log2(float64(count))
+	log2Size := int(math.Ceil(log2Count - log2_10))
+
+	for i := 0; i < log2Size; i++ {
+		points = append(points, points...)
+	}
+
+	return points[:count]
+}
+
+func TestMsmG2BN254(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1 << v
+		points := GenerateG2Points(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, false)
+		fmt.Print("Finished generating scalars\n")
+
+		out := new(G2Point)
+		_, e := MsmG2(out, points, scalars, 0)
+		assert.Equal(t, e, nil, "error should be nil")
+		assert.True(t, out.IsOnCurve())
+	}
+}
+
+func BenchmarkMsmG2BN254(b *testing.B) {
+	LOG_MSM_SIZES := []int{20, 21, 22, 23, 24, 25, 26}
+
+	for _, logMsmSize := range LOG_MSM_SIZES {
+		msmSize := 1 << logMsmSize
+		points := GenerateG2Points(msmSize)
+		scalars := GenerateScalars(msmSize, false)
+		b.Run(fmt.Sprintf("MSM G2 %d", logMsmSize), func(b *testing.B) {
+			for n := 0; n < b.N; n++ {
+				out := new(G2Point)
+				_, e := MsmG2(out, points, scalars, 0)
+
+				if e != nil {
+					panic("Error occured")
+				}
+			}
+		})
+	}
+}
+
+func TestCommitG2MSM(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1 << v
+
+		points := GenerateG2Points(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, true)
+		fmt.Print("Finished generating scalars\n")
+
+		var sizeCheckG2PointAffine G2PointAffine
+		inputPointsBytes := count * int(unsafe.Sizeof(sizeCheckG2PointAffine))
+
+		var sizeCheckG2Point G2Point
+		out_d, _ := goicicle.CudaMalloc(int(unsafe.Sizeof(sizeCheckG2Point)))
+
+		points_d, _ := goicicle.CudaMalloc(inputPointsBytes)
+		goicicle.CudaMemCpyHtoD[G2PointAffine](points_d, points, inputPointsBytes)
+
+		scalarBytes := count * 32
+		scalars_d, _ := goicicle.CudaMalloc(scalarBytes)
+		goicicle.CudaMemCpyHtoD[G1ScalarField](scalars_d, scalars, scalarBytes)
+
+		startTime := time.Now()
+		e := CommitG2(out_d, scalars_d, points_d, count, 10)
+		fmt.Printf("icicle MSM took: %d ms\n", time.Since(startTime).Milliseconds())
+
+		outHost := make([]G2Point, 1)
+		goicicle.CudaMemCpyDtoH[G2Point](outHost, out_d, int(unsafe.Sizeof(sizeCheckG2Point)))
+
+		assert.Equal(t, e, 0, "error should be 0")
+		assert.Equal(t, len(outHost), 1)
+		result := outHost[0]
+
+		assert.True(t, result.IsOnCurve())
+	}
+}
+
+func TestBatchG2MSM(t *testing.T) {
+	for _, batchPow2 := range []int{2, 4} {
+		for _, pow2 := range []int{4, 6} {
+			msmSize := 1 << pow2
+			batchSize := 1 << batchPow2
+			count := msmSize * batchSize
+
+			points := GenerateG2Points(count)
+			scalars := GenerateScalars(count, false)
+
+			pointsResults, e := MsmG2Batch(&points, &scalars, batchSize, 0)
+
+			if e != nil {
+				t.Errorf("MsmBatchBN254 returned an error: %v", e)
+			}
+
+			if len(pointsResults) != batchSize {
+				t.Errorf("Expected length %d, but got %d", batchSize, len(pointsResults))
+			}
+
+			for _, s := range pointsResults {
+				assert.True(t, s.IsOnCurve())
+			}
+		}
+	}
+}

goicicle/curves/bn254/ntt.go

@@ -0,0 +1,221 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bn254
+
+import (
+	"errors"
+	"fmt"
+	"unsafe"
+
+	"github.com/ingonyama-zk/icicle/goicicle"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbn254
+// #include "ntt.h"
+import "C"
+
+const (
+	NONE = 0
+	DIF  = 1
+	DIT  = 2
+)
+
+func Ntt(scalars *[]G1ScalarField, isInverse bool, deviceId int) uint64 {
+	scalarsC := (*C.BN254_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+
+	ret := C.ntt_cuda_bn254(scalarsC, C.uint32_t(len(*scalars)), C.bool(isInverse), C.size_t(deviceId))
+
+	return uint64(ret)
+}
+
+func NttBatch(scalars *[]G1ScalarField, isInverse bool, batchSize, deviceId int) uint64 {
+	scalarsC := (*C.BN254_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+	isInverseC := C.bool(isInverse)
+	batchSizeC := C.uint32_t(batchSize)
+	deviceIdC := C.size_t(deviceId)
+
+	ret := C.ntt_batch_cuda_bn254(scalarsC, C.uint32_t(len(*scalars)), batchSizeC, isInverseC, deviceIdC)
+
+	return uint64(ret)
+}
+
+func EcNtt(values *[]G1ProjectivePoint, isInverse bool, deviceId int) uint64 {
+	valuesC := (*C.BN254_projective_t)(unsafe.Pointer(&(*values)[0]))
+	deviceIdC := C.size_t(deviceId)
+	isInverseC := C.bool(isInverse)
+	n := C.uint32_t(len(*values))
+
+	ret := C.ecntt_cuda_bn254(valuesC, n, isInverseC, deviceIdC)
+
+	return uint64(ret)
+}
+
+func EcNttBatch(values *[]G1ProjectivePoint, isInverse bool, batchSize, deviceId int) uint64 {
+	valuesC := (*C.BN254_projective_t)(unsafe.Pointer(&(*values)[0]))
+	deviceIdC := C.size_t(deviceId)
+	isInverseC := C.bool(isInverse)
+	n := C.uint32_t(len(*values))
+	batchSizeC := C.uint32_t(batchSize)
+
+	ret := C.ecntt_batch_cuda_bn254(valuesC, n, batchSizeC, isInverseC, deviceIdC)
+
+	return uint64(ret)
+}
+
+func GenerateTwiddles(d_size int, log_d_size int, inverse bool) (up unsafe.Pointer, err error) {
+	domain_size := C.uint32_t(d_size)
+	logn := C.uint32_t(log_d_size)
+	is_inverse := C.bool(inverse)
+
+	dp := C.build_domain_cuda_bn254(domain_size, logn, is_inverse, 0, 0)
+
+	if dp == nil {
+		err = errors.New("nullptr returned from generating twiddles")
+		return unsafe.Pointer(nil), err
+	}
+
+	return unsafe.Pointer(dp), nil
+}
+
+// Reverses d_scalars in-place
+func ReverseScalars(d_scalars unsafe.Pointer, len int) (int, error) {
+	scalarsC := (*C.BN254_scalar_t)(d_scalars)
+	lenC := C.int(len)
+	if success := C.reverse_order_scalars_cuda_bn254(scalarsC, lenC, 0, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func Interpolate(scalars, twiddles, cosetPowers unsafe.Pointer, size int, isCoset bool) unsafe.Pointer {
+	size_d := size * 32
+	dp, err := goicicle.CudaMalloc(size_d)
+
+	if err != nil {
+		return nil
+	}
+
+	d_out := (*C.BN254_scalar_t)(dp)
+	scalarsC := (*C.BN254_scalar_t)(scalars)
+	twiddlesC := (*C.BN254_scalar_t)(twiddles)
+	cosetPowersC := (*C.BN254_scalar_t)(cosetPowers)
+	sizeC := C.uint(size)
+
+	var ret C.int
+	if isCoset {
+		ret = C.interpolate_scalars_on_coset_cuda_bn254(d_out, scalarsC, twiddlesC, sizeC, cosetPowersC, 0, 0)
+	} else {
+		ret = C.interpolate_scalars_cuda_bn254(d_out, scalarsC, twiddlesC, sizeC, 0, 0)
+	}
+	if ret != 0 {
+		fmt.Print("error interpolating")
+	}
+
+	return unsafe.Pointer(d_out)
+}
+
+func Evaluate(scalars_out, scalars, twiddles, coset_powers unsafe.Pointer, scalars_size, twiddles_size int, isCoset bool) int {
+	scalars_outC := (*C.BN254_scalar_t)(scalars_out)
+	scalarsC := (*C.BN254_scalar_t)(scalars)
+	twiddlesC := (*C.BN254_scalar_t)(twiddles)
+	coset_powersC := (*C.BN254_scalar_t)(coset_powers)
+	sizeC := C.uint(scalars_size)
+	twiddlesC_size := C.uint(twiddles_size)
+
+	var ret C.int
+	if isCoset {
+		ret = C.evaluate_scalars_on_coset_cuda_bn254(scalars_outC, scalarsC, twiddlesC, twiddlesC_size, sizeC, coset_powersC, 0, 0)
+	} else {
+		ret = C.evaluate_scalars_cuda_bn254(scalars_outC, scalarsC, twiddlesC, twiddlesC_size, sizeC, 0, 0)
+	}
+
+	if ret != 0 {
+		fmt.Print("error interpolating")
+		return -1
+	}
+
+	return 0
+}
+
+func VecScalarAdd(in1_d, in2_d unsafe.Pointer, size int) int {
+	in1_dC := (*C.BN254_scalar_t)(in1_d)
+	in2_dC := (*C.BN254_scalar_t)(in2_d)
+	sizeC := C.uint(size)
+
+	ret := C.add_scalars_cuda_bn254(in1_dC, in1_dC, in2_dC, sizeC, 0)
+
+	if ret != 0 {
+		fmt.Print("error adding scalar vectors")
+		return -1
+	}
+
+	return 0
+}
+
+func VecScalarSub(in1_d, in2_d unsafe.Pointer, size int) int {
+	in1_dC := (*C.BN254_scalar_t)(in1_d)
+	in2_dC := (*C.BN254_scalar_t)(in2_d)
+	sizeC := C.uint(size)
+
+	ret := C.sub_scalars_cuda_bn254(in1_dC, in1_dC, in2_dC, sizeC, 0)
+
+	if ret != 0 {
+		fmt.Print("error subtracting scalar vectors")
+		return -1
+	}
+
+	return 0
+}
+
+func ToMontgomery(d_scalars unsafe.Pointer, len int) (int, error) {
+	scalarsC := (*C.BN254_scalar_t)(d_scalars)
+	lenC := C.uint(len)
+	if success := C.to_montgomery_scalars_cuda_bn254(scalarsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func FromMontgomery(d_scalars unsafe.Pointer, len int) (int, error) {
+	scalarsC := (*C.BN254_scalar_t)(d_scalars)
+	lenC := C.uint(len)
+	if success := C.from_montgomery_scalars_cuda_bn254(scalarsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func AffinePointFromMontgomery(d_points unsafe.Pointer, len int) (int, error) {
+	pointsC := (*C.BN254_affine_t)(d_points)
+	lenC := C.uint(len)
+
+	if success := C.from_montgomery_aff_points_cuda_bn254(pointsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func G2AffinePointFromMontgomery(d_points unsafe.Pointer, len int) (int, error) {
+	pointsC := (*C.BN254_g2_affine_t)(d_points)
+	lenC := C.uint(len)
+
+	if success := C.from_montgomery_aff_points_g2_cuda_bn254(pointsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}

goicicle/curves/bn254/ntt_test.go

@@ -0,0 +1,148 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bn254
+
+import (
+	"fmt"
+	"github.com/stretchr/testify/assert"
+	"reflect"
+	"testing"
+)
+
+func TestNttBN254Batch(t *testing.T) {
+	count := 1 << 20
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	NttBatch(&nttResult, false, count, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	assert.Equal(t, nttResult, nttResult)
+}
+
+func TestNttBN254CompareToGnarkDIF(t *testing.T) {
+	count := 1 << 2
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	Ntt(&nttResult, false, DIF, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	assert.Equal(t, nttResult, nttResult)
+}
+
+func TestINttBN254CompareToGnarkDIT(t *testing.T) {
+	count := 1 << 3
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	Ntt(&nttResult, true, DIT, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	assert.Equal(t, nttResult, nttResult)
+}
+
+func TestNttBN254(t *testing.T) {
+	count := 1 << 3
+
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	Ntt(&nttResult, false, NONE, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	inttResult := make([]G1ScalarField, len(nttResult))
+	copy(inttResult, nttResult)
+
+	assert.Equal(t, inttResult, nttResult)
+	Ntt(&inttResult, true, NONE, 0)
+	assert.Equal(t, inttResult, scalars)
+}
+
+func TestNttBatchBN254(t *testing.T) {
+	count := 1 << 5
+	batches := 4
+
+	scalars := GenerateScalars(count*batches, false)
+
+	var scalarVecOfVec [][]G1ScalarField = make([][]G1ScalarField, 0)
+
+	for i := 0; i < batches; i++ {
+		start := i * count
+		end := (i + 1) * count
+		batch := make([]G1ScalarField, len(scalars[start:end]))
+		copy(batch, scalars[start:end])
+		scalarVecOfVec = append(scalarVecOfVec, batch)
+	}
+
+	nttBatchResult := make([]G1ScalarField, len(scalars))
+	copy(nttBatchResult, scalars)
+
+	NttBatch(&nttBatchResult, false, count, 0)
+
+	var nttResultVecOfVec [][]G1ScalarField
+
+	for i := 0; i < batches; i++ {
+		// Clone the slice
+		clone := make([]G1ScalarField, len(scalarVecOfVec[i]))
+		copy(clone, scalarVecOfVec[i])
+
+		// Add it to the result vector of vectors
+		nttResultVecOfVec = append(nttResultVecOfVec, clone)
+
+		// Call the ntt_bn254 function
+		Ntt(&nttResultVecOfVec[i], false, NONE, 0)
+	}
+
+	assert.NotEqual(t, nttBatchResult, scalars)
+
+	// Check that the ntt of each vec of scalars is equal to the intt of the specific batch
+	for i := 0; i < batches; i++ {
+		if !reflect.DeepEqual(nttResultVecOfVec[i], nttBatchResult[i*count:((i+1)*count)]) {
+			t.Errorf("ntt of vec of scalars not equal to intt of specific batch")
+		}
+	}
+}
+
+func BenchmarkNTT(b *testing.B) {
+	LOG_NTT_SIZES := []int{12, 15, 20, 21, 22, 23, 24, 25, 26}
+
+	for _, logNTTSize := range LOG_NTT_SIZES {
+		nttSize := 1 << logNTTSize
+		b.Run(fmt.Sprintf("NTT %d", logNTTSize), func(b *testing.B) {
+			scalars := GenerateScalars(nttSize, false)
+
+			nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+			copy(nttResult, scalars)
+			for n := 0; n < b.N; n++ {
+				Ntt(&nttResult, false, NONE, 0)
+			}
+		})
+	}
+}

goicicle/curves/bn254/utils.go

@@ -0,0 +1,48 @@
+package bn254
+
+import (
+	"encoding/binary"
+	"fmt"
+	"log"
+	"regexp"
+	"runtime"
+	"time"
+)
+
+// Function to convert [8]uint32 to [4]uint64
+func ConvertUint32ArrToUint64Arr(arr32 [8]uint32) [4]uint64 {
+	var arr64 [4]uint64
+	for i := 0; i < len(arr32); i += 2 {
+		arr64[i/2] = (uint64(arr32[i]) << 32) | uint64(arr32[i+1])
+	}
+	return arr64
+}
+
+func ConvertUint64ArrToUint32Arr(arr64 [4]uint64) [8]uint32 {
+	var arr32 [8]uint32
+	for i, v := range arr64 {
+		b := make([]byte, 8)
+		binary.LittleEndian.PutUint64(b, v)
+
+		arr32[i*2] = binary.LittleEndian.Uint32(b[0:4])
+		arr32[i*2+1] = binary.LittleEndian.Uint32(b[4:8])
+	}
+
+	return arr32
+}
+
+func TimeTrack(start time.Time) {
+	elapsed := time.Since(start)
+
+	// Skip this function, and fetch the PC and file for its parent.
+	pc, _, _, _ := runtime.Caller(1)
+
+	// Retrieve a function object this functions parent.
+	funcObj := runtime.FuncForPC(pc)
+
+	// Regex to extract just the function name (and not the module path).
+	runtimeFunc := regexp.MustCompile(`^.*\.(.*)$`)
+	name := runtimeFunc.ReplaceAllString(funcObj.Name(), "$1")
+
+	log.Println(fmt.Sprintf("%s took %s", name, elapsed))
+}

goicicle/curves/bn254/utils_test.go

@@ -0,0 +1,81 @@
+package bn254
+
+import (
+	"testing"
+)
+
+func TestConvertUint32ArrToUint64Arr(t *testing.T) {
+	testCases := []struct {
+		name  string
+		input [8]uint32
+		want  [4]uint64
+	}{
+		{
+			name:  "Test with incremental array",
+			input: [8]uint32{1, 2, 3, 4, 5, 6, 7, 8},
+			want:  [4]uint64{4294967298, 12884901892, 21474836486, 30064771080},
+		},
+		{
+			name:  "Test with all zeros",
+			input: [8]uint32{0, 0, 0, 0, 0, 0, 0, 0},
+			want:  [4]uint64{0, 0, 0, 0},
+		},
+		{
+			name:  "Test with maximum uint32 values",
+			input: [8]uint32{4294967295, 4294967295, 4294967295, 4294967295, 4294967295, 4294967295, 4294967295, 4294967295},
+			want:  [4]uint64{18446744073709551615, 18446744073709551615, 18446744073709551615, 18446744073709551615},
+		},
+		{
+			name:  "Test with alternating min and max uint32 values",
+			input: [8]uint32{0, 4294967295, 0, 4294967295, 0, 4294967295, 0, 4294967295},
+			want:  [4]uint64{4294967295, 4294967295, 4294967295, 4294967295},
+		},
+		{
+			name:  "Test with alternating max and min uint32 values",
+			input: [8]uint32{4294967295, 0, 4294967295, 0, 4294967295, 0, 4294967295, 0},
+			want:  [4]uint64{18446744069414584320, 18446744069414584320, 18446744069414584320, 18446744069414584320},
+		},
+	}
+
+	for _, tc := range testCases {
+		t.Run(tc.name, func(t *testing.T) {
+			got := ConvertUint32ArrToUint64Arr(tc.input)
+			if got != tc.want {
+				t.Errorf("got %v, want %v", got, tc.want)
+			}
+		})
+	}
+}
+
+func TestConvertUint64ArrToUint32Arr(t *testing.T) {
+	testCases := []struct {
+		name     string
+		input    [4]uint64
+		expected [8]uint32
+	}{
+		{
+			name:     "test one",
+			input:    [4]uint64{1, 2, 3, 4},
+			expected: [8]uint32{1, 0, 2, 0, 3, 0, 4, 0},
+		},
+		{
+			name:     "test two",
+			input:    [4]uint64{100, 200, 300, 400},
+			expected: [8]uint32{100, 0, 200, 0, 300, 0, 400, 0},
+		},
+		{
+			name:     "test three",
+			input:    [4]uint64{1000, 2000, 3000, 4000},
+			expected: [8]uint32{1000, 0, 2000, 0, 3000, 0, 4000, 0},
+		},
+	}
+
+	for _, tc := range testCases {
+		t.Run(tc.name, func(t *testing.T) {
+			got := ConvertUint64ArrToUint32Arr(tc.input)
+			if got != tc.expected {
+				t.Errorf("got %v, want %v", got, tc.expected)
+			}
+		})
+	}
+}

goicicle/curves/bn254/vec_mod.go

@@ -0,0 +1,41 @@
+// Copyright 2023 Ingonyama
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Code generated by Ingonyama DO NOT EDIT
+
+package bn254
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ -lbn254
+// #include "ve_mod_mult.h"
+import "C"
+import (
+	"fmt"
+	"unsafe"
+)
+
+func VecScalarMulMod(scalarVec1, scalarVec2 unsafe.Pointer, size int) int {
+	scalarVec1C := (*C.BN254_scalar_t)(scalarVec1)
+	scalarVec2C := (*C.BN254_scalar_t)(scalarVec2)
+	sizeC := C.size_t(size)
+
+	ret := C.vec_mod_mult_device_scalar_bn254(scalarVec1C, scalarVec2C, sizeC, 0)
+
+	if ret != 0 {
+		fmt.Print("error multiplying scalar vectors")
+		return -1
+	}
+
+	return 0
+}

goicicle/goicicle.go

@@ -0,0 +1,49 @@
+package goicicle
+
+// This file implements CUDA driver context management
+
+// #cgo CFLAGS: -I /usr/loca/cuda/include
+// #cgo LDFLAGS: -L/usr/local/cuda/lib64 -lcudart
+/*
+#include <cuda.h>
+#include <cuda_runtime.h>
+*/
+import "C"
+
+import (
+	"errors"
+	"unsafe"
+)
+
+func CudaMalloc(size int) (dp unsafe.Pointer, err error) {
+	var p C.void
+	dp = unsafe.Pointer(&p)
+	if err := C.cudaMalloc(&dp, C.size_t(size)); err != 0 {
+		return nil, errors.New("could not create memory space")
+	}
+	return dp, nil
+}
+
+func CudaFree(dp unsafe.Pointer) int {
+	if err := C.cudaFree(dp); err != 0 {
+		return -1
+	}
+	return 0
+}
+
+func CudaMemCpyHtoD[T any](dst_d unsafe.Pointer, src []T, size int) int {
+	src_c := unsafe.Pointer(&src[0])
+	if err := C.cudaMemcpy(dst_d, src_c, C.size_t(size), 1); err != 0 {
+		return -1
+	}
+	return 0
+}
+
+func CudaMemCpyDtoH[T any](dst []T, src_d unsafe.Pointer, size int) int {
+	dst_c := unsafe.Pointer(&dst[0])
+
+	if err := C.cudaMemcpy(dst_c, src_d, C.size_t(size), 2); err != 0 {
+		return -1
+	}
+	return 0
+}

goicicle/setup.sh

@@ -0,0 +1,41 @@
+#!/bin/bash
+
+SUDO=''
+if [ "$EUID" != 0 ]; then 
+  echo "Icicle setup script should be run with root privileges, please run this as root"
+  SUDO='sudo'
+fi
+
+
+TARGET_BN254="libbn254.so"
+TARGET_BLS12_381="libbls12_381.so"
+TARGET_BLS12_377="libbls12_377.so"
+
+MAKE_FAIL=0
+
+$SUDO make $1 || MAKE_FAIL=1
+
+if [ $MAKE_FAIL != 0 ]; then
+    echo "make failed, install dependencies and re-run setup script with root privileges"
+    exit
+fi
+
+TARGET_BN254_PATH=$(dirname "$(find `pwd` -name $TARGET_BN254 -print -quit)")/
+TARGET_BLS12_381_PATH=$(dirname "$(find `pwd` -name $TARGET_BLS12_381 -print -quit)")/
+TARGET_BLS12_377_PATH=$(dirname "$(find `pwd` -name $TARGET_BLS12_377 -print -quit)")/
+
+
+if [[ "$TARGET_BLS12_377_PATH" != "" ]]; then
+    echo "BLS12_377 found @ $TARGET_BLS12_377_PATH"
+    export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:$TARGET_BLS12_377_PATH
+fi
+
+if [[ "$TARGET_BN254_PATH" != "" ]]; then
+    echo "BN254 found @ $TARGET_BN254_PATH"
+    export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:$TARGET_BN254_PATH
+fi
+
+if [[ "$TARGET_BLS12_381_PATH" != "" ]]; then
+    echo "BLS12_381_PATH found @ $TARGET_BLS12_381_PATH"
+    export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:$TARGET_BLS12_381_PATH
+fi

goicicle/templates/curves/curves.go

@@ -0,0 +1,42 @@
+package config
+
+// {{.SharedLib}}
+type Curve struct {
+	PackageName        string
+	CurveNameUpperCase string
+	CurveNameLowerCase string
+	SharedLib          string
+	ScalarSize         int
+	BaseSize           int
+	G2ElementSize      int
+}
+
+var BN_254 = Curve{
+	PackageName:        "bn254",
+	CurveNameUpperCase: "BN254",
+	CurveNameLowerCase: "bn254",
+	SharedLib:          "-lbn254",
+	ScalarSize:         8,
+	BaseSize:           8,
+	G2ElementSize:      4,
+}
+
+var BLS_12_377 = Curve{
+	PackageName:        "bls12377",
+	CurveNameUpperCase: "BLS12_377",
+	CurveNameLowerCase: "bls12_377",
+	SharedLib:          "-lbls12_377",
+	ScalarSize:         8,
+	BaseSize:           12,
+	G2ElementSize:      6,
+}
+
+var BLS_12_381 = Curve{
+	PackageName:        "bls12381",
+	CurveNameUpperCase: "BLS12_381",
+	CurveNameLowerCase: "bls12_381",
+	SharedLib:          "-lbls12_381",
+	ScalarSize:         8,
+	BaseSize:           12,
+	G2ElementSize:      6,
+}

goicicle/templates/curves/g1.go.tmpl

@@ -0,0 +1,306 @@
+import (
+	"unsafe"
+
+	"encoding/binary"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ {{.SharedLib}}
+// #include "projective.h"
+// #include "ve_mod_mult.h"
+import "C"
+
+const SCALAR_SIZE = {{.ScalarSize}}
+const BASE_SIZE = {{.BaseSize}}
+
+type G1ScalarField struct {
+	S [SCALAR_SIZE]uint32
+}
+
+type G1BaseField struct {
+	S [BASE_SIZE]uint32
+}
+
+/*
+ * BaseField Constrctors
+ */
+
+func (f *G1BaseField) SetZero() *G1BaseField {
+	var S [BASE_SIZE]uint32
+	f.S = S
+
+	return f
+}
+
+func (f *G1BaseField) SetOne() *G1BaseField {
+	var S [BASE_SIZE]uint32
+
+	S[0] = 1
+
+	f.S = S
+	return f
+}
+
+func (p *G1ProjectivePoint) FromAffine(affine *G1PointAffine) *G1ProjectivePoint {
+	out := (*C.{{.CurveNameUpperCase}}_projective_t)(unsafe.Pointer(p))
+	in := (*C.{{.CurveNameUpperCase}}_affine_t)(unsafe.Pointer(affine))
+
+	C.projective_from_affine_{{.CurveNameLowerCase}}(out, in)
+
+	return p
+}
+
+func (f *G1BaseField) FromLimbs(limbs [BASE_SIZE]uint32) *G1BaseField {
+	copy(f.S[:], limbs[:])
+
+	return f
+}
+
+/*
+ * BaseField methods
+ */
+
+func (f *G1BaseField) Limbs() [BASE_SIZE]uint32 {
+	return f.S
+}
+
+func (f *G1BaseField) ToBytesLe() []byte {
+	bytes := make([]byte, len(f.S)*4)
+	for i, v := range f.S {
+		binary.LittleEndian.PutUint32(bytes[i*4:], v)
+	}
+
+	return bytes
+}
+
+/*
+ * ScalarField methods
+ */
+
+func (p *G1ScalarField) Random() *G1ScalarField {
+	outC := (*C.{{.CurveNameUpperCase}}_scalar_t)(unsafe.Pointer(p))
+	C.random_scalar_{{.CurveNameLowerCase}}(outC)
+
+	return p
+}
+
+func (f *G1ScalarField) SetZero() *G1ScalarField {
+	var S [SCALAR_SIZE]uint32
+	f.S = S
+
+	return f
+}
+
+func (f *G1ScalarField) SetOne() *G1ScalarField {
+	var S [SCALAR_SIZE]uint32
+	S[0] = 1
+	f.S = S
+
+	return f
+}
+
+func (a *G1ScalarField) Eq(b *G1ScalarField) bool {
+	for i, v := range a.S {
+		if b.S[i] != v {
+			return false
+		}
+	}
+	return true
+}
+
+/*
+ * ScalarField methods
+ */
+
+func (f *G1ScalarField) Limbs() [SCALAR_SIZE]uint32 {
+	return f.S
+}
+
+func (f *G1ScalarField) ToBytesLe() []byte {
+	bytes := make([]byte, len(f.S)*4)
+	for i, v := range f.S {
+		binary.LittleEndian.PutUint32(bytes[i*4:], v)
+	}
+
+	return bytes
+}
+
+/*
+ * Point{{.CurveNameUpperCase}}
+ */
+
+type G1ProjectivePoint struct {
+	X, Y, Z G1BaseField
+}
+
+func (f *G1ProjectivePoint) SetZero() *G1ProjectivePoint {
+	var yOne G1BaseField
+	yOne.SetOne()
+
+	var xZero G1BaseField
+	xZero.SetZero()
+
+	var zZero G1BaseField
+	zZero.SetZero()
+
+	f.X = xZero
+	f.Y = yOne
+	f.Z = zZero
+
+	return f
+}
+
+func (p *G1ProjectivePoint) Eq(pCompare *G1ProjectivePoint) bool {
+	// Cast *Point{{.CurveNameUpperCase}} to *C.{{.CurveNameUpperCase}}_projective_t
+	// The unsafe.Pointer cast is necessary because Go doesn't allow direct casts
+	// between different pointer types.
+	// It'S your responsibility to ensure that the types are compatible.
+	pC := (*C.{{.CurveNameUpperCase}}_projective_t)(unsafe.Pointer(p))
+	pCompareC := (*C.{{.CurveNameUpperCase}}_projective_t)(unsafe.Pointer(pCompare))
+
+	// Call the C function
+	// The C function doesn't keep any references to the data,
+	// so it'S fine if the Go garbage collector moves or deletes the data later.
+	return bool(C.eq_{{.CurveNameLowerCase}}(pC, pCompareC))
+}
+
+func (p *G1ProjectivePoint) IsOnCurve() bool {
+	point := (*C.{{.CurveNameUpperCase}}_projective_t)(unsafe.Pointer(p))
+	res := C.projective_is_on_curve_{{.CurveNameLowerCase}}(point)
+
+	return bool(res)
+}
+
+func (p *G1ProjectivePoint) Random() *G1ProjectivePoint {
+	outC := (*C.{{.CurveNameUpperCase}}_projective_t)(unsafe.Pointer(p))
+	C.random_projective_{{.CurveNameLowerCase}}(outC)
+
+	return p
+}
+
+func (p *G1ProjectivePoint) StripZ() *G1PointAffine {
+	return &G1PointAffine{
+		X: p.X,
+		Y: p.Y,
+	}
+}
+
+func (p *G1ProjectivePoint) FromLimbs(x, y, z *[]uint32) *G1ProjectivePoint {
+	var _x G1BaseField
+	var _y G1BaseField
+	var _z G1BaseField
+
+	_x.FromLimbs(GetFixedLimbs(x))
+	_y.FromLimbs(GetFixedLimbs(y))
+	_z.FromLimbs(GetFixedLimbs(z))
+
+	p.X = _x
+	p.Y = _y
+	p.Z = _z
+
+	return p
+}
+
+/*
+ * PointAffineNoInfinity{{.CurveNameUpperCase}}
+ */
+
+type G1PointAffine struct {
+	X, Y G1BaseField
+}
+
+func (p *G1PointAffine) FromProjective(projective *G1ProjectivePoint) *G1PointAffine {
+	in := (*C.{{.CurveNameUpperCase}}_projective_t)(unsafe.Pointer(projective))
+	out := (*C.{{.CurveNameUpperCase}}_affine_t)(unsafe.Pointer(p))
+
+	C.projective_to_affine_{{.CurveNameLowerCase}}(out,in)
+
+	return p
+}
+
+func (p *G1PointAffine) ToProjective() *G1ProjectivePoint {
+	var Z G1BaseField
+	Z.SetOne()
+
+	return &G1ProjectivePoint{
+		X: p.X,
+		Y: p.Y,
+		Z: Z,
+	}
+}
+
+func (p *G1PointAffine) FromLimbs(X, Y *[]uint32) *G1PointAffine {
+	var _x G1BaseField
+	var _y G1BaseField
+
+	_x.FromLimbs(GetFixedLimbs(X))
+	_y.FromLimbs(GetFixedLimbs(Y))
+
+	return p
+}
+
+/*
+ * Multiplication
+ */
+
+func MultiplyVec(a []G1ProjectivePoint, b []G1ScalarField, deviceID int) {
+	if len(a) != len(b) {
+		panic("a and b have different lengths")
+	}
+
+	pointsC := (*C.{{.CurveNameUpperCase}}_projective_t)(unsafe.Pointer(&a[0]))
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(unsafe.Pointer(&b[0]))
+	deviceIdC := C.size_t(deviceID)
+	nElementsC := C.size_t(len(a))
+
+	C.vec_mod_mult_point_{{.CurveNameLowerCase}}(pointsC, scalarsC, nElementsC, deviceIdC)
+}
+
+func MultiplyScalar(a []G1ScalarField, b []G1ScalarField, deviceID int) {
+	if len(a) != len(b) {
+		panic("a and b have different lengths")
+	}
+
+	aC := (*C.{{.CurveNameUpperCase}}_scalar_t)(unsafe.Pointer(&a[0]))
+	bC := (*C.{{.CurveNameUpperCase}}_scalar_t)(unsafe.Pointer(&b[0]))
+	deviceIdC := C.size_t(deviceID)
+	nElementsC := C.size_t(len(a))
+
+	C.vec_mod_mult_scalar_{{.CurveNameLowerCase}}(aC, bC, nElementsC, deviceIdC)
+}
+
+// Multiply a matrix by a scalar:
+//
+//	`a` - flattenned matrix;
+//	`b` - vector to multiply `a` by;
+func MultiplyMatrix(a []G1ScalarField, b []G1ScalarField, deviceID int) {
+	c := make([]G1ScalarField, len(b))
+	for i := range c {
+		var p G1ScalarField
+		p.SetZero()
+
+		c[i] = p
+	}
+
+	aC := (*C.{{.CurveNameUpperCase}}_scalar_t)(unsafe.Pointer(&a[0]))
+	bC := (*C.{{.CurveNameUpperCase}}_scalar_t)(unsafe.Pointer(&b[0]))
+	cC := (*C.{{.CurveNameUpperCase}}_scalar_t)(unsafe.Pointer(&c[0]))
+	deviceIdC := C.size_t(deviceID)
+	nElementsC := C.size_t(len(a))
+
+	C.matrix_vec_mod_mult_{{.CurveNameLowerCase}}(aC, bC, cC, nElementsC, deviceIdC)
+}
+
+/*
+ * Utils
+ */
+
+func GetFixedLimbs(slice *[]uint32) [BASE_SIZE]uint32 {
+	if len(*slice) <= BASE_SIZE {
+		limbs := [BASE_SIZE]uint32{}
+		copy(limbs[:len(*slice)], *slice)
+		return limbs
+	}
+
+	panic("slice has too many elements")
+}

goicicle/templates/curves/g1_test.go.tmpl

@@ -0,0 +1,180 @@
+import (
+	"encoding/binary"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestNewField{{.CurveNameUpperCase}}One(t *testing.T) {
+	var oneField G1BaseField
+	oneField.SetOne()
+
+	rawOneField := [8]uint32([8]uint32{0x1, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0})
+
+	assert.Equal(t, oneField.S, rawOneField)
+}
+
+func TestNewField{{.CurveNameUpperCase}}Zero(t *testing.T) {
+	var zeroField G1BaseField
+	zeroField.SetZero()
+
+	rawZeroField := [8]uint32([8]uint32{0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0})
+
+	assert.Equal(t, zeroField.S, rawZeroField)
+}
+
+func TestField{{.CurveNameUpperCase}}ToBytesLe(t *testing.T) {
+	var p G1ProjectivePoint
+	p.Random()
+
+	expected := make([]byte, len(p.X.S)*4) // each uint32 takes 4 bytes
+	for i, v := range p.X.S {
+		binary.LittleEndian.PutUint32(expected[i*4:], v)
+	}
+
+	assert.Equal(t, p.X.ToBytesLe(), expected)
+	assert.Equal(t, len(p.X.ToBytesLe()), 32)
+}
+
+func TestNewPoint{{.CurveNameUpperCase}}Zero(t *testing.T) {
+	var pointZero G1ProjectivePoint
+	pointZero.SetZero()
+
+	var baseOne G1BaseField
+	baseOne.SetOne()
+
+	var zeroSanity G1BaseField
+	zeroSanity.SetZero()
+
+	assert.Equal(t, pointZero.X, zeroSanity)
+	assert.Equal(t, pointZero.Y, baseOne)
+	assert.Equal(t, pointZero.Z, zeroSanity)
+}
+
+func TestFromProjectiveToAffine(t *testing.T) {
+	var projective G1ProjectivePoint
+	var affine G1PointAffine
+
+	projective.Random()
+
+	affine.FromProjective(&projective)
+	var projective2 G1ProjectivePoint
+	projective2.FromAffine(&affine)
+
+	assert.True(t, projective.IsOnCurve())
+	assert.True(t, projective2.IsOnCurve())
+	assert.True(t, projective.Eq(&projective2))
+}
+
+func Test{{.CurveNameUpperCase}}Eq(t *testing.T) {
+	var p1 G1ProjectivePoint
+	p1.Random()
+	var p2 G1ProjectivePoint
+	p2.Random()
+
+	assert.Equal(t, p1.Eq(&p1), true)
+	assert.Equal(t, p1.Eq(&p2), false)
+}
+
+func Test{{.CurveNameUpperCase}}StripZ(t *testing.T) {
+	var p1 G1ProjectivePoint
+	p1.Random()
+
+	p2ZLess := p1.StripZ()
+
+	assert.IsType(t, G1PointAffine{}, *p2ZLess)
+	assert.Equal(t, p1.X, p2ZLess.X)
+	assert.Equal(t, p1.Y, p2ZLess.Y)
+}
+
+func TestPoint{{.CurveNameUpperCase}}fromLimbs(t *testing.T) {
+	var p G1ProjectivePoint
+	p.Random()
+
+	x := p.X.Limbs()
+	y := p.Y.Limbs()
+	z := p.Z.Limbs()
+
+	xSlice := x[:]
+	ySlice := y[:]
+	zSlice := z[:]
+
+	var pFromLimbs G1ProjectivePoint
+	pFromLimbs.FromLimbs(&xSlice, &ySlice, &zSlice)
+
+	assert.Equal(t, pFromLimbs, p)
+}
+
+func TestNewPointAffineNoInfinity{{.CurveNameUpperCase}}Zero(t *testing.T) {
+	var zeroP G1PointAffine
+
+	var zeroSanity G1BaseField
+	zeroSanity.SetZero()
+
+	assert.Equal(t, zeroP.X, zeroSanity)
+	assert.Equal(t, zeroP.Y, zeroSanity)
+}
+
+func TestPointAffineNoInfinity{{.CurveNameUpperCase}}FromLimbs(t *testing.T) {
+	// Initialize your test values
+	x := [8]uint32{1, 2, 3, 4, 5, 6, 7, 8}
+	y := [8]uint32{9, 10, 11, 12, 13, 14, 15, 16}
+	xSlice := x[:]
+	ySlice := y[:]
+
+	// Execute your function
+	var result G1PointAffine
+	result.FromLimbs(&xSlice, &ySlice)
+
+	var xBase G1BaseField
+	var yBase G1BaseField
+	xBase.FromLimbs(x)
+	yBase.FromLimbs(y)
+
+	// Define your expected result
+	expected := &G1PointAffine{
+		X: xBase,
+		Y: yBase,
+	}
+
+	// Test if result is as expected
+	assert.Equal(t, result, expected)
+}
+
+func TestGetFixedLimbs(t *testing.T) {
+	t.Run("case of valid input of length less than 8", func(t *testing.T) {
+		slice := []uint32{1, 2, 3, 4, 5, 6, 7}
+		expected := [8]uint32{1, 2, 3, 4, 5, 6, 7, 0}
+
+		result := GetFixedLimbs(&slice)
+		assert.Equal(t, result, expected)
+	})
+
+	t.Run("case of valid input of length 8", func(t *testing.T) {
+		slice := []uint32{1, 2, 3, 4, 5, 6, 7, 8}
+		expected := [8]uint32{1, 2, 3, 4, 5, 6, 7, 8}
+
+		result := GetFixedLimbs(&slice)
+		assert.Equal(t, result, expected)
+	})
+
+	t.Run("case of empty input", func(t *testing.T) {
+		slice := []uint32{}
+		expected := [8]uint32{0, 0, 0, 0, 0, 0, 0, 0}
+
+		result := GetFixedLimbs(&slice)
+		assert.Equal(t, result, expected)
+	})
+
+	t.Run("case of input length greater than 8", func(t *testing.T) {
+		slice := []uint32{1, 2, 3, 4, 5, 6, 7, 8, 9}
+
+		defer func() {
+			if r := recover(); r == nil {
+				t.Errorf("the code did not panic")
+			}
+		}()
+
+		GetFixedLimbs(&slice)
+	})
+}

goicicle/templates/curves/g2.go.tmpl

@@ -0,0 +1,95 @@
+import (
+	"encoding/binary"
+	"unsafe"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ {{.SharedLib}}
+// #include "projective.h"
+// #include "ve_mod_mult.h"
+import "C"
+
+// G2 extension field
+
+type G2Element [{{.G2ElementSize}}]uint64
+
+type ExtentionField struct {
+	A0, A1 G2Element
+}
+
+type G2PointAffine struct {
+	X, Y ExtentionField
+}
+
+type G2Point struct {
+	X, Y, Z ExtentionField
+}
+
+
+func (p *G2Point) Random() *G2Point {
+	outC := (*C.{{.CurveNameUpperCase}}_g2_projective_t)(unsafe.Pointer(p))
+	C.random_g2_projective_{{.CurveNameLowerCase}}(outC)
+
+	return p
+}
+
+func (p *G2Point) FromAffine(affine *G2PointAffine) *G2Point {
+	out := (*C.{{.CurveNameUpperCase}}_g2_projective_t)(unsafe.Pointer(p))
+	in := (*C.{{.CurveNameUpperCase}}_g2_affine_t)(unsafe.Pointer(affine))
+
+	C.g2_projective_from_affine_{{.CurveNameLowerCase}}(out, in)
+
+	return p
+}
+
+func (p *G2Point) Eq(pCompare *G2Point) bool {
+	// Cast *Point{{.CurveNameUpperCase}} to *C.{{.CurveNameUpperCase}}_projective_t
+	// The unsafe.Pointer cast is necessary because Go doesn't allow direct casts
+	// between different pointer types.
+	// It's your responsibility to ensure that the types are compatible.
+	pC := (*C.{{.CurveNameUpperCase}}_g2_projective_t)(unsafe.Pointer(p))
+	pCompareC := (*C.{{.CurveNameUpperCase}}_g2_projective_t)(unsafe.Pointer(pCompare))
+
+	// Call the C function
+	// The C function doesn't keep any references to the data,
+	// so it's fine if the Go garbage collector moves or deletes the data later.
+	return bool(C.eq_g2_{{.CurveNameLowerCase}}(pC, pCompareC))
+}
+
+func (f *G2Element) ToBytesLe() []byte {
+	var bytes []byte
+	for _, val := range f {
+		buf := make([]byte, 8) // 8 bytes because uint64 is 64-bit
+		binary.LittleEndian.PutUint64(buf, val)
+		bytes = append(bytes, buf...)
+	}
+	return bytes
+}
+
+func (p *G2PointAffine) ToProjective() G2Point {
+	return G2Point{
+		X: p.X,
+		Y: p.Y,
+		Z: ExtentionField{
+			A0: G2Element{1, 0, 0, 0},
+			A1: G2Element{0, 0, 0, 0},
+		},
+	}
+}
+
+func (p *G2PointAffine) FromProjective(projective *G2Point) *G2PointAffine {
+	out := (*C.{{.CurveNameUpperCase}}_g2_affine_t)(unsafe.Pointer(p))
+	in := (*C.{{.CurveNameUpperCase}}_g2_projective_t)(unsafe.Pointer(projective))
+
+	C.g2_projective_to_affine_{{.CurveNameLowerCase}}(out, in)
+
+	return p
+}
+
+func (p *G2Point) IsOnCurve() bool {
+	// Directly copy memory from the C struct to the Go struct
+	point := (*C.{{.CurveNameUpperCase}}_g2_projective_t)(unsafe.Pointer(p))
+	res := C.g2_projective_is_on_curve_{{.CurveNameLowerCase}}(point)
+
+	return bool(res)
+}

goicicle/templates/curves/g2_test.go.tmpl

@@ -0,0 +1,58 @@
+import (
+	"github.com/stretchr/testify/assert"
+	"testing"
+)
+
+func TestG2Eqg2(t *testing.T) {
+	var point G2Point
+
+	point.Random()
+
+	assert.True(t, point.Eq(&point))
+}
+
+func TestG2FromProjectiveToAffine(t *testing.T) {
+	var projective G2Point
+	var affine G2PointAffine
+
+	projective.Random()
+
+	affine.FromProjective(&projective)
+	var projective2 G2Point
+	projective2.FromAffine(&affine)
+
+	
+	assert.True(t, projective.IsOnCurve())
+	assert.True(t, projective2.IsOnCurve())
+	assert.True(t, projective.Eq(&projective2))
+}
+
+func TestG2Eqg2NotEqual(t *testing.T) {
+	var point G2Point
+	point.Random()
+
+	var point2 G2Point
+	point2.Random()
+
+	assert.False(t, point.Eq(&point2))
+}
+
+func TestG2ToBytes(t *testing.T) {
+	element := G2Element{0x6546098ea84b6298, 0x4a384533d1f68aca, 0xaa0666972d771336, 0x1569e4a34321993}
+	bytes := element.ToBytesLe()
+
+	assert.Equal(t, bytes, []byte{0x98, 0x62, 0x4b, 0xa8, 0x8e, 0x9, 0x46, 0x65, 0xca, 0x8a, 0xf6, 0xd1, 0x33, 0x45, 0x38, 0x4a, 0x36, 0x13, 0x77, 0x2d, 0x97, 0x66, 0x6, 0xaa, 0x93, 0x19, 0x32, 0x34, 0x4a, 0x9e, 0x56, 0x1})
+}
+
+func TestG2ShouldConvertToProjective(t *testing.T) {
+	var pointProjective G2Point
+	var pointAffine G2PointAffine
+
+	pointProjective.Random()
+	pointAffine.FromProjective(&pointProjective)
+
+	proj := pointAffine.ToProjective()
+
+	assert.True(t, proj.IsOnCurve())
+	assert.True(t, pointProjective.Eq(&proj))
+}

goicicle/templates/hfiles/msm.h.tmpl

@@ -0,0 +1,84 @@
+#include <cuda.h>
+#include <cuda_runtime.h>
+#include <stdbool.h>
+// msm.h
+
+#ifndef _{{.CurveNameUpperCase}}_MSM_H
+#define _{{.CurveNameUpperCase}}_MSM_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+// Incomplete declaration of {{.CurveNameUpperCase}} projective and affine structs
+typedef struct {{.CurveNameUpperCase}}_projective_t {{.CurveNameUpperCase}}_projective_t;
+typedef struct {{.CurveNameUpperCase}}_g2_projective_t {{.CurveNameUpperCase}}_g2_projective_t;
+typedef struct {{.CurveNameUpperCase}}_affine_t {{.CurveNameUpperCase}}_affine_t;
+typedef struct {{.CurveNameUpperCase}}_g2_affine_t {{.CurveNameUpperCase}}_g2_affine_t;
+typedef struct {{.CurveNameUpperCase}}_scalar_t {{.CurveNameUpperCase}}_scalar_t;
+typedef cudaStream_t CudaStream_t;
+
+int msm_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_projective_t* out, {{.CurveNameUpperCase}}_affine_t* points, {{.CurveNameUpperCase}}_scalar_t* scalars, size_t count, size_t device_id);
+
+int msm_batch_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_projective_t* out,
+  {{.CurveNameUpperCase}}_affine_t* points,
+  {{.CurveNameUpperCase}}_scalar_t* scalars,
+  size_t batch_size,
+  size_t msm_size,
+  size_t device_id);
+
+int commit_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_projective_t* d_out,
+  {{.CurveNameUpperCase}}_scalar_t* d_scalars,
+  {{.CurveNameUpperCase}}_affine_t* d_points,
+  size_t count,
+  unsigned large_bucket_factor,
+  size_t device_id);
+
+int commit_batch_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_projective_t* d_out,
+  {{.CurveNameUpperCase}}_scalar_t* d_scalars,
+  {{.CurveNameUpperCase}}_affine_t* d_points,
+  size_t count,
+  size_t batch_size,
+  size_t device_id);
+
+int msm_g2_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_g2_projective_t* out,
+  {{.CurveNameUpperCase}}_g2_affine_t* points,
+  {{.CurveNameUpperCase}}_scalar_t* scalars,
+  size_t count,
+  size_t device_id);
+
+int msm_batch_g2_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_g2_projective_t* out,
+  {{.CurveNameUpperCase}}_g2_affine_t* points,
+  {{.CurveNameUpperCase}}_scalar_t* scalars,
+  size_t batch_size,
+  size_t msm_size,
+  size_t device_id);
+
+int commit_g2_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_g2_projective_t* d_out,
+  {{.CurveNameUpperCase}}_scalar_t* d_scalars,
+  {{.CurveNameUpperCase}}_g2_affine_t* d_points,
+  size_t count,
+  unsigned large_bucket_factor,
+  size_t device_id);
+
+int commit_batch_g2_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_g2_projective_t* d_out,
+  {{.CurveNameUpperCase}}_scalar_t* d_scalars,
+  {{.CurveNameUpperCase}}_g2_affine_t* d_points,
+  size_t count,
+  size_t batch_size,
+  size_t device_id,
+  cudaStream_t stream);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _{{.CurveNameUpperCase}}_MSM_H */

goicicle/templates/hfiles/ntt.h.tmpl

@@ -0,0 +1,181 @@
+#include <cuda.h>
+#include <stdbool.h>
+// ntt.h
+
+#ifndef _{{.CurveNameUpperCase}}_NTT_H
+#define _{{.CurveNameUpperCase}}_NTT_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+// Incomplete declaration of {{.CurveNameUpperCase}} projective and affine structs
+typedef struct {{.CurveNameUpperCase}}_projective_t {{.CurveNameUpperCase}}_projective_t;
+typedef struct {{.CurveNameUpperCase}}_affine_t {{.CurveNameUpperCase}}_affine_t;
+typedef struct {{.CurveNameUpperCase}}_scalar_t {{.CurveNameUpperCase}}_scalar_t;
+
+typedef struct {{.CurveNameUpperCase}}_g2_projective_t {{.CurveNameUpperCase}}_g2_projective_t;
+typedef struct {{.CurveNameUpperCase}}_g2_affine_t {{.CurveNameUpperCase}}_g2_affine_t;
+
+int ntt_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_scalar_t* arr, uint32_t n, bool inverse, size_t device_id);
+int ntt_batch_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* arr, uint32_t arr_size, uint32_t batch_size, bool inverse, size_t device_id);
+
+int ecntt_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_projective_t* arr, uint32_t n, bool inverse, size_t device_id);
+int ecntt_batch_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_projective_t* arr, uint32_t arr_size, uint32_t batch_size, bool inverse, size_t device_id);
+
+{{.CurveNameUpperCase}}_scalar_t* 
+build_domain_cuda_{{.CurveNameLowerCase}}(uint32_t domain_size, uint32_t logn, bool inverse, size_t device_id, size_t stream);
+
+int interpolate_scalars_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* d_out,
+  {{.CurveNameUpperCase}}_scalar_t* d_evaluations,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned n,
+  unsigned device_id,
+  size_t stream);
+int interpolate_scalars_batch_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* d_out,
+  {{.CurveNameUpperCase}}_scalar_t* d_evaluations,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int interpolate_points_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_projective_t* d_out,
+  {{.CurveNameUpperCase}}_projective_t* d_evaluations,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned n,
+  size_t device_id,
+  size_t stream);
+int interpolate_points_batch_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_projective_t* d_out,
+  {{.CurveNameUpperCase}}_projective_t* d_evaluations,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int interpolate_scalars_on_coset_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* d_out,
+  {{.CurveNameUpperCase}}_scalar_t* d_evaluations,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned n,
+  {{.CurveNameUpperCase}}_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int interpolate_scalars_batch_on_coset_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* d_out,
+  {{.CurveNameUpperCase}}_scalar_t* d_evaluations,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  {{.CurveNameUpperCase}}_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+
+int evaluate_scalars_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* d_out,
+  {{.CurveNameUpperCase}}_scalar_t* d_coefficients,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned device_id,
+  size_t stream);
+int evaluate_scalars_batch_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* d_out,
+  {{.CurveNameUpperCase}}_scalar_t* d_coefficients,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_projective_t* d_out,
+  {{.CurveNameUpperCase}}_projective_t* d_coefficients,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_batch_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_projective_t* d_out,
+  {{.CurveNameUpperCase}}_projective_t* d_coefficients,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id,
+  size_t stream);
+int evaluate_scalars_on_coset_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* d_out,
+  {{.CurveNameUpperCase}}_scalar_t* d_coefficients,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  {{.CurveNameUpperCase}}_scalar_t* coset_powers,
+  unsigned device_id,
+  size_t stream);
+int evaluate_scalars_on_coset_batch_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* d_out,
+  {{.CurveNameUpperCase}}_scalar_t* d_coefficients,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  {{.CurveNameUpperCase}}_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_on_coset_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_projective_t* d_out,
+  {{.CurveNameUpperCase}}_projective_t* d_coefficients,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  {{.CurveNameUpperCase}}_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+int evaluate_points_on_coset_batch_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_projective_t* d_out,
+  {{.CurveNameUpperCase}}_projective_t* d_coefficients,
+  {{.CurveNameUpperCase}}_scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  {{.CurveNameUpperCase}}_scalar_t* coset_powers,
+  size_t device_id,
+  size_t stream);
+
+int reverse_order_scalars_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_scalar_t* arr, int n, size_t device_id, size_t stream);
+int reverse_order_scalars_batch_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* arr, int n, int batch_size, size_t device_id, size_t stream);
+int reverse_order_points_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_projective_t* arr, int n, size_t device_id, size_t stream);
+int reverse_order_points_batch_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_projective_t* arr, int n, int batch_size, size_t device_id, size_t stream);
+int add_scalars_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* d_out, {{.CurveNameUpperCase}}_scalar_t* d_in1, {{.CurveNameUpperCase}}_scalar_t* d_in2, unsigned n, size_t stream);
+int sub_scalars_cuda_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* d_out, {{.CurveNameUpperCase}}_scalar_t* d_in1, {{.CurveNameUpperCase}}_scalar_t* d_in2, unsigned n, size_t stream);
+int to_montgomery_scalars_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_scalar_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_scalars_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_scalar_t* d_inout, unsigned n, size_t stream);
+
+// points g1
+int to_montgomery_proj_points_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_projective_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_proj_points_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_projective_t* d_inout, unsigned n, size_t stream);
+int to_montgomery_aff_points_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_affine_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_aff_points_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_affine_t* d_inout, unsigned n, size_t stream);
+
+// points g2
+int to_montgomery_proj_points_g2_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_g2_projective_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_proj_points_g2_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_g2_projective_t* d_inout, unsigned n, size_t stream);
+int to_montgomery_aff_points_g2_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_g2_affine_t* d_inout, unsigned n, size_t stream);
+int from_montgomery_aff_points_g2_cuda_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_g2_affine_t* d_inout, unsigned n, size_t stream);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _{{.CurveNameUpperCase}}_NTT_H */

goicicle/templates/hfiles/projective.h.tmpl

@@ -0,0 +1,33 @@
+#include <cuda.h>
+#include <stdbool.h>
+// projective.h
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+typedef struct {{.CurveNameUpperCase}}_projective_t {{.CurveNameUpperCase}}_projective_t;
+typedef struct {{.CurveNameUpperCase}}_g2_projective_t {{.CurveNameUpperCase}}_g2_projective_t;
+typedef struct {{.CurveNameUpperCase}}_affine_t {{.CurveNameUpperCase}}_affine_t;
+typedef struct {{.CurveNameUpperCase}}_g2_affine_t {{.CurveNameUpperCase}}_g2_affine_t;
+typedef struct {{.CurveNameUpperCase}}_scalar_t {{.CurveNameUpperCase}}_scalar_t;
+
+bool projective_is_on_curve_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_projective_t* point1);
+
+int random_scalar_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_scalar_t* out); 
+int random_projective_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_projective_t* out);
+{{.CurveNameUpperCase}}_projective_t* projective_zero_{{.CurveNameLowerCase}}();
+int projective_to_affine_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_affine_t* out, {{.CurveNameUpperCase}}_projective_t* point1);
+int projective_from_affine_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_projective_t* out, {{.CurveNameUpperCase}}_affine_t* point1);
+
+int random_g2_projective_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_g2_projective_t* out);
+int g2_projective_to_affine_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_g2_affine_t* out, {{.CurveNameUpperCase}}_g2_projective_t* point1);
+int g2_projective_from_affine_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_g2_projective_t* out, {{.CurveNameUpperCase}}_g2_affine_t* point1);
+bool g2_projective_is_on_curve_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_g2_projective_t* point1);
+
+bool eq_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_projective_t* point1, {{.CurveNameUpperCase}}_projective_t* point2);
+bool eq_g2_{{.CurveNameLowerCase}}({{.CurveNameUpperCase}}_g2_projective_t* point1, {{.CurveNameUpperCase}}_g2_projective_t* point2);
+
+#ifdef __cplusplus
+}
+#endif

goicicle/templates/hfiles/ve_mod_mult.h.tmpl

@@ -0,0 +1,32 @@
+#include <stdbool.h>
+#include <cuda.h>
+// ve_mod_mult.h
+
+#ifndef _{{.CurveNameUpperCase}}_VEC_MULT_H
+#define _{{.CurveNameUpperCase}}_VEC_MULT_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+typedef struct {{.CurveNameUpperCase}}_projective_t {{.CurveNameUpperCase}}_projective_t;
+typedef struct {{.CurveNameUpperCase}}_scalar_t {{.CurveNameUpperCase}}_scalar_t;
+
+int32_t vec_mod_mult_point_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_projective_t* inout, {{.CurveNameUpperCase}}_scalar_t* scalar_vec, size_t n_elments, size_t device_id);
+int32_t vec_mod_mult_scalar_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* inout, {{.CurveNameUpperCase}}_scalar_t* scalar_vec, size_t n_elments, size_t device_id);
+int32_t vec_mod_mult_device_scalar_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* inout, {{.CurveNameUpperCase}}_scalar_t* scalar_vec, size_t n_elements, size_t device_id);
+int32_t matrix_vec_mod_mult_{{.CurveNameLowerCase}}(
+  {{.CurveNameUpperCase}}_scalar_t* matrix_flattened,
+  {{.CurveNameUpperCase}}_scalar_t* input,
+  {{.CurveNameUpperCase}}_scalar_t* output,
+  size_t n_elments,
+  size_t device_id);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* _{{.CurveNameUpperCase}}_VEC_MULT_H */

goicicle/templates/main.go

@@ -0,0 +1,246 @@
+package main
+
+import (
+	"fmt"
+	"os"
+	"path/filepath"
+
+	"github.com/consensys/bavard"
+	config "github.com/ingonyama-zk/icicle/goicicle/templates/curves"
+)
+
+const (
+	copyrightHolder = "Ingonyama"
+	generatedBy     = "Ingonyama"
+	copyrightYear   = 2023
+	baseDir         = "../curves/"
+)
+
+var bgen = bavard.NewBatchGenerator(copyrightHolder, copyrightYear, generatedBy)
+
+func genMainFiles() {
+	bn254_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bn254", "g1.go"), Templates: []string{"g1.go.tmpl"}},
+	}
+
+	bls12377_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12377", "g1.go"), Templates: []string{"g1.go.tmpl"}},
+	}
+
+	bls12381_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12381", "g1.go"), Templates: []string{"g1.go.tmpl"}},
+	}
+
+	assertNoError(bgen.Generate(config.BLS_12_377, config.BLS_12_377.PackageName, "./curves/", bls12377_entries...))
+	assertNoError(bgen.Generate(config.BN_254, config.BN_254.PackageName, "./curves/", bn254_entries...))
+	assertNoError(bgen.Generate(config.BLS_12_381, config.BLS_12_381.PackageName, "./curves/", bls12381_entries...))
+
+	bn254_g2_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bn254", "g2.go"), Templates: []string{"g2.go.tmpl"}},
+	}
+
+	bls12377_g2_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12377", "g2.go"), Templates: []string{"g2.go.tmpl"}},
+	}
+
+	bls12381_g2_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12381", "g2.go"), Templates: []string{"g2.go.tmpl"}},
+	}
+
+	assertNoError(bgen.Generate(config.BLS_12_377, config.BLS_12_377.PackageName, "./curves/", bls12377_g2_entries...))
+	assertNoError(bgen.Generate(config.BN_254, config.BN_254.PackageName, "./curves/", bn254_g2_entries...))
+	assertNoError(bgen.Generate(config.BLS_12_381, config.BLS_12_381.PackageName, "./curves/", bls12381_g2_entries...))
+	bn254_msm_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bn254", "msm.go"), Templates: []string{"msm.go.tmpl"}},
+	}
+
+	bls12377_msm_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12377", "msm.go"), Templates: []string{"msm.go.tmpl"}},
+	}
+
+	bls12381_msm_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12381", "msm.go"), Templates: []string{"msm.go.tmpl"}},
+	}
+
+	assertNoError(bgen.Generate(config.BLS_12_377, config.BLS_12_377.PackageName, "./msm/", bls12377_msm_entries...))
+	assertNoError(bgen.Generate(config.BN_254, config.BN_254.PackageName, "./msm/", bn254_msm_entries...))
+	assertNoError(bgen.Generate(config.BLS_12_381, config.BLS_12_381.PackageName, "./msm/", bls12381_msm_entries...))
+
+	bn254_ntt_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bn254", "ntt.go"), Templates: []string{"ntt.go.tmpl"}},
+	}
+
+	bls12377_ntt_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12377", "ntt.go"), Templates: []string{"ntt.go.tmpl"}},
+	}
+
+	bls12381_ntt_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12381", "ntt.go"), Templates: []string{"ntt.go.tmpl"}},
+	}
+
+	assertNoError(bgen.Generate(config.BLS_12_377, config.BLS_12_377.PackageName, "./ntt/", bls12377_ntt_entries...))
+	assertNoError(bgen.Generate(config.BN_254, config.BN_254.PackageName, "./ntt/", bn254_ntt_entries...))
+	assertNoError(bgen.Generate(config.BLS_12_381, config.BLS_12_381.PackageName, "./ntt/", bls12381_ntt_entries...))
+
+	bn254_vec_mod_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bn254", "vec_mod.go"), Templates: []string{"vec_mod.go.tmpl"}},
+	}
+
+	bls12377_vec_mod_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12377", "vec_mod.go"), Templates: []string{"vec_mod.go.tmpl"}},
+	}
+
+	bls12381_vec_mod_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12381", "vec_mod.go"), Templates: []string{"vec_mod.go.tmpl"}},
+	}
+
+	assertNoError(bgen.Generate(config.BLS_12_377, config.BLS_12_377.PackageName, "./ops/", bls12377_vec_mod_entries...))
+	assertNoError(bgen.Generate(config.BN_254, config.BN_254.PackageName, "./ops/", bn254_vec_mod_entries...))
+	assertNoError(bgen.Generate(config.BLS_12_381, config.BLS_12_381.PackageName, "./ops/", bls12381_vec_mod_entries...))
+
+	h_msm_bn254 := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bn254", "include", "msm.h"), Templates: []string{"msm.h.tmpl"}},
+	}
+
+	h_msm_bls12377 := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12377", "include", "msm.h"), Templates: []string{"msm.h.tmpl"}},
+	}
+
+	h_msm_bls12381 := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12381", "include", "msm.h"), Templates: []string{"msm.h.tmpl"}},
+	}
+
+	assertNoError(bgen.Generate(config.BLS_12_377, config.BLS_12_377.PackageName, "./hfiles/", h_msm_bls12377...))
+	assertNoError(bgen.Generate(config.BN_254, config.BN_254.PackageName, "./hfiles/", h_msm_bn254...))
+	assertNoError(bgen.Generate(config.BLS_12_381, config.BLS_12_381.PackageName, "./hfiles/", h_msm_bls12381...))
+
+	h_ntt_bn254 := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bn254", "include", "ntt.h"), Templates: []string{"ntt.h.tmpl"}},
+	}
+
+	h_ntt_bls12377 := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12377", "include", "ntt.h"), Templates: []string{"ntt.h.tmpl"}},
+	}
+
+	h_ntt_bls12381 := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12381", "include", "ntt.h"), Templates: []string{"ntt.h.tmpl"}},
+	}
+
+	assertNoError(bgen.Generate(config.BLS_12_377, config.BLS_12_377.PackageName, "./hfiles/", h_ntt_bls12377...))
+	assertNoError(bgen.Generate(config.BN_254, config.BN_254.PackageName, "./hfiles/", h_ntt_bn254...))
+	assertNoError(bgen.Generate(config.BLS_12_381, config.BLS_12_381.PackageName, "./hfiles/", h_ntt_bls12381...))
+
+	ve_mod_mult_h_bn254 := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bn254", "include", "ve_mod_mult.h"), Templates: []string{"ve_mod_mult.h.tmpl"}},
+	}
+
+	ve_mod_mult_h_bls12377 := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12377", "include", "ve_mod_mult.h"), Templates: []string{"ve_mod_mult.h.tmpl"}},
+	}
+
+	ve_mod_mult_ht_bls12381 := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12381", "include", "ve_mod_mult.h"), Templates: []string{"ve_mod_mult.h.tmpl"}},
+	}
+
+	assertNoError(bgen.Generate(config.BLS_12_377, config.BLS_12_377.PackageName, "./hfiles/", ve_mod_mult_h_bls12377...))
+	assertNoError(bgen.Generate(config.BN_254, config.BN_254.PackageName, "./hfiles/", ve_mod_mult_h_bn254...))
+	assertNoError(bgen.Generate(config.BLS_12_381, config.BLS_12_381.PackageName, "./hfiles/", ve_mod_mult_ht_bls12381...))
+
+	projective_bn254 := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bn254", "include", "projective.h"), Templates: []string{"projective.h.tmpl"}},
+	}
+
+	projective_bls12377 := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12377", "include", "projective.h"), Templates: []string{"projective.h.tmpl"}},
+	}
+
+	projective_bls12381 := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12381", "include", "projective.h"), Templates: []string{"projective.h.tmpl"}},
+	}
+
+	assertNoError(bgen.Generate(config.BLS_12_377, config.BLS_12_377.PackageName, "./hfiles/", projective_bls12377...))
+	assertNoError(bgen.Generate(config.BN_254, config.BN_254.PackageName, "./hfiles/", projective_bn254...))
+	assertNoError(bgen.Generate(config.BLS_12_381, config.BLS_12_381.PackageName, "./hfiles/", projective_bls12381...))
+}
+
+func genTestFiles() {
+	// G1 TESTS
+	bn254_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bn254", "g1_test.go"), Templates: []string{"g1_test.go.tmpl"}},
+	}
+
+	bls12377_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12377", "g1_test.go"), Templates: []string{"g1_test.go.tmpl"}},
+	}
+
+	bls12381_entries := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12381", "g1_test.go"), Templates: []string{"g1_test.go.tmpl"}},
+	}
+
+	assertNoError(bgen.Generate(config.BLS_12_377, config.BLS_12_377.PackageName, "./curves/", bls12377_entries...))
+	assertNoError(bgen.Generate(config.BN_254, config.BN_254.PackageName, "./curves/", bn254_entries...))
+	assertNoError(bgen.Generate(config.BLS_12_381, config.BLS_12_381.PackageName, "./curves/", bls12381_entries...))
+
+	// G2 TESTS
+	bn254_entries_g2_test := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bn254", "g2_test.go"), Templates: []string{"g2_test.go.tmpl"}},
+	}
+
+	bls12377_entries_g2_test := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12377", "g2_test.go"), Templates: []string{"g2_test.go.tmpl"}},
+	}
+
+	bls12381_entries_g2_test := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12381", "g2_test.go"), Templates: []string{"g2_test.go.tmpl"}},
+	}
+
+	assertNoError(bgen.Generate(config.BLS_12_377, config.BLS_12_377.PackageName, "./curves/", bls12377_entries_g2_test...))
+	assertNoError(bgen.Generate(config.BN_254, config.BN_254.PackageName, "./curves/", bn254_entries_g2_test...))
+	assertNoError(bgen.Generate(config.BLS_12_381, config.BLS_12_381.PackageName, "./curves/", bls12381_entries_g2_test...))
+
+	// MSM TEST
+	bn254_entries_msm_test := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bn254", "msm_test.go"), Templates: []string{"msm_test.go.tmpl"}},
+	}
+
+	bls12377_entries_msm_test := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12377", "msm_test.go"), Templates: []string{"msm_test.go.tmpl"}},
+	}
+
+	bls12381_entries_msm_test := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12381", "msm_test.go"), Templates: []string{"msm_test.go.tmpl"}},
+	}
+
+	assertNoError(bgen.Generate(config.BLS_12_377, config.BLS_12_377.PackageName, "./msm/", bls12377_entries_msm_test...))
+	assertNoError(bgen.Generate(config.BN_254, config.BN_254.PackageName, "./msm/", bn254_entries_msm_test...))
+	assertNoError(bgen.Generate(config.BLS_12_381, config.BLS_12_381.PackageName, "./msm/", bls12381_entries_msm_test...))
+
+	// FFT TEST
+	bn254_entries_fft_test := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bn254", "ntt_test.go"), Templates: []string{"ntt_test.go.tmpl"}},
+	}
+
+	bls12377_entries_fft_test := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12377", "ntt_test.go"), Templates: []string{"ntt_test.go.tmpl"}},
+	}
+
+	bls12381_entries_fft_test := []bavard.Entry{
+		{File: filepath.Join(baseDir, "bls12381", "ntt_test.go"), Templates: []string{"ntt_test.go.tmpl"}},
+	}
+
+	assertNoError(bgen.Generate(config.BLS_12_377, config.BLS_12_377.PackageName, "./ntt/", bls12377_entries_fft_test...))
+	assertNoError(bgen.Generate(config.BN_254, config.BN_254.PackageName, "./ntt/", bn254_entries_fft_test...))
+	assertNoError(bgen.Generate(config.BLS_12_381, config.BLS_12_381.PackageName, "./ntt/", bls12381_entries_fft_test...))
+}
+
+func main() {
+	genMainFiles()
+	genTestFiles()
+}
+
+func assertNoError(err error) {
+	if err != nil {
+		fmt.Printf("\n%s\n", err.Error())
+		os.Exit(-1)
+	}
+}

goicicle/templates/msm/msm.go.tmpl

@@ -0,0 +1,190 @@
+import (
+	"errors"
+	"fmt"
+	"unsafe"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ {{.SharedLib}}
+// #include "msm.h"
+import "C"
+
+func Msm(out *G1ProjectivePoint, points []G1PointAffine, scalars []G1ScalarField, device_id int) (*G1ProjectivePoint, error) {
+	if len(points) != len(scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	pointsC := (*C.{{.CurveNameUpperCase}}_affine_t)(unsafe.Pointer(&points[0]))
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(unsafe.Pointer(&scalars[0]))
+	outC := (*C.{{.CurveNameUpperCase}}_projective_t)(unsafe.Pointer(out))
+	ret := C.msm_cuda_{{.CurveNameLowerCase}}(outC, pointsC, scalarsC, C.size_t(len(points)), C.size_t(device_id))
+
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_cuda_{{.CurveNameLowerCase}} returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func MsmG2(out *G2Point, points []G2PointAffine, scalars []G1ScalarField, device_id int) (*G2Point, error) {
+	if len(points) != len(scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	pointsC := (*C.{{.CurveNameUpperCase}}_g2_affine_t)(unsafe.Pointer(&points[0]))
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(unsafe.Pointer(&scalars[0]))
+	outC := (*C.{{.CurveNameUpperCase}}_g2_projective_t)(unsafe.Pointer(out))
+
+	ret := C.msm_g2_cuda_{{.CurveNameLowerCase}}(outC, pointsC, scalarsC, C.size_t(len(points)), C.size_t(device_id))
+
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_g2_cuda_{{.CurveNameLowerCase}} returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func MsmBatch(points *[]G1PointAffine, scalars *[]G1ScalarField, batchSize, deviceId int) ([]*G1ProjectivePoint, error) {
+	// Check for nil pointers
+	if points == nil || scalars == nil {
+		return nil, errors.New("points or scalars is nil")
+	}
+
+	if len(*points) != len(*scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	// Check for empty slices
+	if len(*points) == 0 || len(*scalars) == 0 {
+		return nil, errors.New("points or scalars is empty")
+	}
+
+	// Check for zero batchSize
+	if batchSize <= 0 {
+		return nil, errors.New("error on: batchSize must be greater than zero")
+	}
+
+	out := make([]*G1ProjectivePoint, batchSize)
+
+	for i := 0; i < len(out); i++ {
+		var p G1ProjectivePoint
+		p.SetZero()
+
+		out[i] = &p
+	}
+
+	outC := (*C.{{.CurveNameUpperCase}}_projective_t)(unsafe.Pointer(&out[0]))
+	pointsC := (*C.{{.CurveNameUpperCase}}_affine_t)(unsafe.Pointer(&(*points)[0]))
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+	msmSizeC := C.size_t(len(*points) / batchSize)
+	deviceIdC := C.size_t(deviceId)
+	batchSizeC := C.size_t(batchSize)
+
+	ret := C.msm_batch_cuda_{{.CurveNameLowerCase}}(outC, pointsC, scalarsC, batchSizeC, msmSizeC, deviceIdC)
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_batch_cuda_{{.CurveNameLowerCase}} returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func MsmG2Batch(points *[]G2PointAffine, scalars *[]G1ScalarField, batchSize, deviceId int) ([]*G2Point, error) {
+	// Check for nil pointers
+	if points == nil || scalars == nil {
+		return nil, errors.New("points or scalars is nil")
+	}
+
+	if len(*points) != len(*scalars) {
+		return nil, errors.New("error on: len(points) != len(scalars)")
+	}
+
+	// Check for empty slices
+	if len(*points) == 0 || len(*scalars) == 0 {
+		return nil, errors.New("points or scalars is empty")
+	}
+
+	// Check for zero batchSize
+	if batchSize <= 0 {
+		return nil, errors.New("error on: batchSize must be greater than zero")
+	}
+
+	out := make([]*G2Point, batchSize)
+
+	outC := (*C.{{.CurveNameUpperCase}}_g2_projective_t)(unsafe.Pointer(&out[0]))
+	pointsC := (*C.{{.CurveNameUpperCase}}_g2_affine_t)(unsafe.Pointer(&(*points)[0]))
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+	msmSizeC := C.size_t(len(*points) / batchSize)
+	deviceIdC := C.size_t(deviceId)
+	batchSizeC := C.size_t(batchSize)
+
+	ret := C.msm_batch_g2_cuda_{{.CurveNameLowerCase}}(outC, pointsC, scalarsC, batchSizeC, msmSizeC, deviceIdC)
+	if ret != 0 {
+		return nil, fmt.Errorf("msm_batch_cuda_{{.CurveNameLowerCase}} returned error code: %d", ret)
+	}
+
+	return out, nil
+}
+
+func Commit(d_out, d_scalars, d_points unsafe.Pointer, count, bucketFactor int) int {
+	d_outC := (*C.{{.CurveNameUpperCase}}_projective_t)(d_out)
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(d_scalars)
+	pointsC := (*C.{{.CurveNameUpperCase}}_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	largeBucketFactorC := C.uint(bucketFactor)
+
+	ret := C.commit_cuda_{{.CurveNameLowerCase}}(d_outC, scalarsC, pointsC, countC, largeBucketFactorC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}
+
+func CommitG2(d_out, d_scalars, d_points unsafe.Pointer, count, bucketFactor int) int {
+	d_outC := (*C.{{.CurveNameUpperCase}}_g2_projective_t)(d_out)
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(d_scalars)
+	pointsC := (*C.{{.CurveNameUpperCase}}_g2_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	largeBucketFactorC := C.uint(bucketFactor)
+
+	ret := C.commit_g2_cuda_{{.CurveNameLowerCase}}(d_outC, scalarsC, pointsC, countC, largeBucketFactorC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}
+
+func CommitBatch(d_out, d_scalars, d_points unsafe.Pointer, count, batch_size int) int {
+	d_outC := (*C.{{.CurveNameUpperCase}}_projective_t)(d_out)
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(d_scalars)
+	pointsC := (*C.{{.CurveNameUpperCase}}_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	batch_sizeC := (C.size_t)(batch_size)
+
+	ret := C.commit_batch_cuda_{{.CurveNameLowerCase}}(d_outC, scalarsC, pointsC, countC, batch_sizeC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}
+
+func CommitG2Batch(d_out, d_scalars, d_points unsafe.Pointer, count, batch_size int) int {
+	d_outC := (*C.{{.CurveNameUpperCase}}_g2_projective_t)(d_out)
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(d_scalars)
+	pointsC := (*C.{{.CurveNameUpperCase}}_g2_affine_t)(d_points)
+	countC := (C.size_t)(count)
+	batch_sizeC := (C.size_t)(batch_size)
+
+	ret := C.msm_batch_g2_cuda_{{.CurveNameLowerCase}}(d_outC, pointsC, scalarsC, countC, batch_sizeC, 0)
+
+	if ret != 0 {
+		return -1
+	}
+
+	return 0
+}

goicicle/templates/msm/msm_test.go.tmpl

@@ -0,0 +1,337 @@
+import (
+	"fmt"
+	"math"
+	"testing"
+	"time"
+	"unsafe"
+
+	"github.com/ingonyama-zk/icicle/goicicle"
+	"github.com/stretchr/testify/assert"
+)
+
+func GeneratePoints(count int) []G1PointAffine {
+	// Declare a slice of integers
+	var points []G1PointAffine
+
+	// populate the slice
+	for i := 0; i < 10; i++ {
+		var pointProjective G1ProjectivePoint
+		pointProjective.Random()
+
+		var pointAffine G1PointAffine
+		pointAffine.FromProjective(&pointProjective)
+
+		points = append(points, pointAffine)
+	}
+
+	log2_10 := math.Log2(10)
+	log2Count := math.Log2(float64(count))
+	log2Size := int(math.Ceil(log2Count - log2_10))
+
+	for i := 0; i < log2Size; i++ {
+		points = append(points, points...)
+	}
+
+	return points[:count]
+}
+
+func GeneratePointsProj(count int) []G1ProjectivePoint {
+	// Declare a slice of integers
+	var points []G1ProjectivePoint
+	// Use a loop to populate the slice
+	for i := 0; i < count; i++ {
+		var p G1ProjectivePoint
+		p.Random()
+
+		points = append(points, p)
+	}
+
+	return points
+}
+
+func GenerateScalars(count int, skewed bool) []G1ScalarField {
+	// Declare a slice of integers
+	var scalars []G1ScalarField
+
+	var rand G1ScalarField
+	var zero G1ScalarField
+	var one G1ScalarField
+	var randLarge G1ScalarField
+
+	zero.SetZero()
+	one.SetOne()
+	randLarge.Random()
+
+	if skewed && count > 1_200_000 {
+		for i := 0; i < count-1_200_000; i++ {
+			rand.Random()
+			scalars = append(scalars, rand)
+		}
+
+		for i := 0; i < 600_000; i++ {
+			scalars = append(scalars, randLarge)
+		}
+		for i := 0; i < 400_000; i++ {
+			scalars = append(scalars, zero)
+		}
+		for i := 0; i < 200_000; i++ {
+			scalars = append(scalars, one)
+		}
+	} else {
+		for i := 0; i < count; i++ {
+			rand.Random()
+			scalars = append(scalars, rand)
+		}
+	}
+
+	return scalars[:count]
+}
+
+func TestMSM(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1 << v
+
+		points := GeneratePoints(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, true)
+		fmt.Print("Finished generating scalars\n")
+
+		out := new(G1ProjectivePoint)
+		startTime := time.Now()
+		_, e := Msm(out, points, scalars, 0) // non mont
+		fmt.Printf("icicle MSM took: %d ms\n", time.Since(startTime).Milliseconds())
+
+		assert.Equal(t, e, nil, "error should be nil")
+
+		assert.True(t, out.IsOnCurve())
+	}
+}
+
+func TestCommitMSM(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1<<v - 1
+
+		points := GeneratePoints(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, true)
+		fmt.Print("Finished generating scalars\n")
+
+		out_d, _ := goicicle.CudaMalloc(96)
+
+		pointsBytes := count * 64
+		points_d, _ := goicicle.CudaMalloc(pointsBytes)
+		goicicle.CudaMemCpyHtoD[G1PointAffine](points_d, points, pointsBytes)
+
+		scalarBytes := count * 32
+		scalars_d, _ := goicicle.CudaMalloc(scalarBytes)
+		goicicle.CudaMemCpyHtoD[G1ScalarField](scalars_d, scalars, scalarBytes)
+
+		startTime := time.Now()
+		e := Commit(out_d, scalars_d, points_d, count, 10)
+		fmt.Printf("icicle MSM took: %d ms\n", time.Since(startTime).Milliseconds())
+
+		outHost := make([]G1ProjectivePoint, 1)
+		goicicle.CudaMemCpyDtoH[G1ProjectivePoint](outHost, out_d, 96)
+
+		assert.Equal(t, e, 0, "error should be 0")
+		assert.True(t, outHost[0].IsOnCurve())
+	}
+}
+
+func BenchmarkCommit(b *testing.B) {
+	LOG_MSM_SIZES := []int{20, 21, 22, 23, 24, 25, 26}
+
+	for _, logMsmSize := range LOG_MSM_SIZES {
+		msmSize := 1 << logMsmSize
+		points := GeneratePoints(msmSize)
+		scalars := GenerateScalars(msmSize, false)
+
+		out_d, _ := goicicle.CudaMalloc(96)
+
+		pointsBytes := msmSize * 64
+		points_d, _ := goicicle.CudaMalloc(pointsBytes)
+		goicicle.CudaMemCpyHtoD[G1PointAffine](points_d, points, pointsBytes)
+
+		scalarBytes := msmSize * 32
+		scalars_d, _ := goicicle.CudaMalloc(scalarBytes)
+		goicicle.CudaMemCpyHtoD[G1ScalarField](scalars_d, scalars, scalarBytes)
+
+		b.Run(fmt.Sprintf("MSM %d", logMsmSize), func(b *testing.B) {
+			for n := 0; n < b.N; n++ {
+				e := Commit(out_d, scalars_d, points_d, msmSize, 10)
+
+				if e != 0 {
+					panic("Error occured")
+				}
+			}
+		})
+	}
+}
+
+func TestBenchMSM(t *testing.T) {
+	for _, batchPow2 := range []int{2, 4} {
+		for _, pow2 := range []int{4, 6} {
+			msmSize := 1 << pow2
+			batchSize := 1 << batchPow2
+			count := msmSize * batchSize
+
+			points := GeneratePoints(count)
+			scalars := GenerateScalars(count, false)
+
+			a, e := MsmBatch(&points, &scalars, batchSize, 0)
+
+			if e != nil {
+				t.Errorf("MsmBatch{{.CurveNameUpperCase}} returned an error: %v", e)
+			}
+
+			if len(a) != batchSize {
+				t.Errorf("Expected length %d, but got %d", batchSize, len(a))
+			}
+		}
+	}
+}
+
+func BenchmarkMSM(b *testing.B) {
+	LOG_MSM_SIZES := []int{20, 21, 22, 23, 24, 25, 26}
+
+	for _, logMsmSize := range LOG_MSM_SIZES {
+		msmSize := 1 << logMsmSize
+		points := GeneratePoints(msmSize)
+		scalars := GenerateScalars(msmSize, false)
+		b.Run(fmt.Sprintf("MSM %d", logMsmSize), func(b *testing.B) {
+			for n := 0; n < b.N; n++ {
+				out := new(G1ProjectivePoint)
+				_, e := Msm(out, points, scalars, 0)
+
+				if e != nil {
+					panic("Error occured")
+				}
+			}
+		})
+	}
+}
+
+// G2
+func GenerateG2Points(count int) []G2PointAffine {
+	// Declare a slice of integers
+	var points []G2PointAffine
+
+	// populate the slice
+	for i := 0; i < 10; i++ {
+		var p G2Point
+		p.Random()
+		var affine G2PointAffine
+		affine.FromProjective(&p)
+
+		points = append(points, affine)
+	}
+
+	log2_10 := math.Log2(10)
+	log2Count := math.Log2(float64(count))
+	log2Size := int(math.Ceil(log2Count - log2_10))
+
+	for i := 0; i < log2Size; i++ {
+		points = append(points, points...)
+	}
+
+	return points[:count]
+}
+
+func TestMsmG2{{.CurveNameUpperCase}}(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1 << v
+		points := GenerateG2Points(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, false)
+		fmt.Print("Finished generating scalars\n")
+
+		out := new(G2Point)
+		_, e := MsmG2(out, points, scalars, 0)
+		assert.Equal(t, e, nil, "error should be nil")
+		assert.True(t, out.IsOnCurve())
+	}
+}
+
+func BenchmarkMsmG2{{.CurveNameUpperCase}}(b *testing.B) {
+	LOG_MSM_SIZES := []int{20, 21, 22, 23, 24, 25, 26}
+
+	for _, logMsmSize := range LOG_MSM_SIZES {
+		msmSize := 1 << logMsmSize
+		points := GenerateG2Points(msmSize)
+		scalars := GenerateScalars(msmSize, false)
+		b.Run(fmt.Sprintf("MSM G2 %d", logMsmSize), func(b *testing.B) {
+			for n := 0; n < b.N; n++ {
+				out := new(G2Point)
+				_, e := MsmG2(out, points, scalars, 0)
+
+				if e != nil {
+					panic("Error occured")
+				}
+			}
+		})
+	}
+}
+
+func TestCommitG2MSM(t *testing.T) {
+	for _, v := range []int{24} {
+		count := 1 << v
+
+		points := GenerateG2Points(count)
+		fmt.Print("Finished generating points\n")
+		scalars := GenerateScalars(count, true)
+		fmt.Print("Finished generating scalars\n")
+
+		var sizeCheckG2PointAffine G2PointAffine
+		inputPointsBytes := count * int(unsafe.Sizeof(sizeCheckG2PointAffine))
+
+		var sizeCheckG2Point G2Point
+		out_d, _ := goicicle.CudaMalloc(int(unsafe.Sizeof(sizeCheckG2Point)))
+
+		points_d, _ := goicicle.CudaMalloc(inputPointsBytes)
+		goicicle.CudaMemCpyHtoD[G2PointAffine](points_d, points, inputPointsBytes)
+
+		scalarBytes := count * 32
+		scalars_d, _ := goicicle.CudaMalloc(scalarBytes)
+		goicicle.CudaMemCpyHtoD[G1ScalarField](scalars_d, scalars, scalarBytes)
+
+		startTime := time.Now()
+		e := CommitG2(out_d, scalars_d, points_d, count, 10)
+		fmt.Printf("icicle MSM took: %d ms\n", time.Since(startTime).Milliseconds())
+
+		outHost := make([]G2Point, 1)
+		goicicle.CudaMemCpyDtoH[G2Point](outHost, out_d, int(unsafe.Sizeof(sizeCheckG2Point)))
+
+		assert.Equal(t, e, 0, "error should be 0")
+		assert.Equal(t, len(outHost), 1)
+		result := outHost[0]
+
+		assert.True(t, result.IsOnCurve())
+	}
+}
+
+func TestBatchG2MSM(t *testing.T) {
+	for _, batchPow2 := range []int{2, 4} {
+		for _, pow2 := range []int{4, 6} {
+			msmSize := 1 << pow2
+			batchSize := 1 << batchPow2
+			count := msmSize * batchSize
+
+			points := GenerateG2Points(count)
+			scalars := GenerateScalars(count, false)
+
+			pointsResults, e := MsmG2Batch(&points, &scalars, batchSize, 0)
+
+			if e != nil {
+				t.Errorf("MsmBatch{{.CurveNameUpperCase}} returned an error: %v", e)
+			}
+
+			if len(pointsResults) != batchSize {
+				t.Errorf("Expected length %d, but got %d", batchSize, len(pointsResults))
+			}
+
+			for _, s := range pointsResults {
+				assert.True(t, s.IsOnCurve())
+			}
+		}
+	}
+}

goicicle/templates/ntt/ntt.go.tmpl

@@ -0,0 +1,203 @@
+import (
+	"errors"
+	"fmt"
+	"unsafe"
+
+	"github.com/ingonyama-zk/icicle/goicicle"
+)
+
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ {{.SharedLib}}
+// #include "ntt.h"
+import "C"
+
+const (
+	NONE = 0
+	DIF  = 1
+	DIT  = 2
+)
+
+func Ntt(scalars *[]G1ScalarField, isInverse bool, deviceId int) uint64 {
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+
+	ret := C.ntt_cuda_{{.CurveNameLowerCase}}(scalarsC, C.uint32_t(len(*scalars)), C.bool(isInverse), C.size_t(deviceId))
+
+	return uint64(ret)
+}
+
+func NttBatch(scalars *[]G1ScalarField, isInverse bool, batchSize, deviceId int) uint64 {
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(unsafe.Pointer(&(*scalars)[0]))
+	isInverseC := C.bool(isInverse)
+	batchSizeC := C.uint32_t(batchSize)
+	deviceIdC := C.size_t(deviceId)
+
+	ret := C.ntt_batch_cuda_{{.CurveNameLowerCase}}(scalarsC, C.uint32_t(len(*scalars)), batchSizeC, isInverseC, deviceIdC)
+
+	return uint64(ret)
+}
+
+func EcNtt(values *[]G1ProjectivePoint, isInverse bool, deviceId int) uint64 {
+	valuesC := (*C.{{.CurveNameUpperCase}}_projective_t)(unsafe.Pointer(&(*values)[0]))
+	deviceIdC := C.size_t(deviceId)
+	isInverseC := C.bool(isInverse)
+	n := C.uint32_t(len(*values))
+
+	ret := C.ecntt_cuda_{{.CurveNameLowerCase}}(valuesC, n, isInverseC, deviceIdC)
+
+	return uint64(ret)
+}
+
+func EcNttBatch(values *[]G1ProjectivePoint, isInverse bool, batchSize, deviceId int) uint64 {
+	valuesC := (*C.{{.CurveNameUpperCase}}_projective_t)(unsafe.Pointer(&(*values)[0]))
+	deviceIdC := C.size_t(deviceId)
+	isInverseC := C.bool(isInverse)
+	n := C.uint32_t(len(*values))
+	batchSizeC := C.uint32_t(batchSize)
+
+	ret := C.ecntt_batch_cuda_{{.CurveNameLowerCase}}(valuesC, n, batchSizeC, isInverseC, deviceIdC)
+
+	return uint64(ret)
+}
+
+func GenerateTwiddles(d_size int, log_d_size int, inverse bool) (up unsafe.Pointer, err error) {
+	domain_size := C.uint32_t(d_size)
+	logn := C.uint32_t(log_d_size)
+	is_inverse := C.bool(inverse)
+
+	dp := C.build_domain_cuda_{{.CurveNameLowerCase}}(domain_size, logn, is_inverse, 0, 0)
+
+	if dp == nil {
+		err = errors.New("nullptr returned from generating twiddles")
+		return unsafe.Pointer(nil), err
+	}
+
+	return unsafe.Pointer(dp), nil
+}
+
+// Reverses d_scalars in-place
+func ReverseScalars(d_scalars unsafe.Pointer, len int) (int, error) {
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(d_scalars)
+	lenC := C.int(len)
+	if success := C.reverse_order_scalars_cuda_{{.CurveNameLowerCase}}(scalarsC, lenC, 0, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func Interpolate(scalars, twiddles, cosetPowers unsafe.Pointer, size int, isCoset bool) unsafe.Pointer {
+	size_d := size * 32
+	dp, err := goicicle.CudaMalloc(size_d)
+
+	if err != nil {
+		return nil
+	}
+
+	d_out := (*C.{{.CurveNameUpperCase}}_scalar_t)(dp)
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(scalars)
+	twiddlesC := (*C.{{.CurveNameUpperCase}}_scalar_t)(twiddles)
+	cosetPowersC := (*C.{{.CurveNameUpperCase}}_scalar_t)(cosetPowers)
+	sizeC := C.uint(size)
+
+	var ret C.int
+	if isCoset {
+		ret = C.interpolate_scalars_on_coset_cuda_{{.CurveNameLowerCase}}(d_out, scalarsC, twiddlesC, sizeC, cosetPowersC, 0, 0)
+	} else {
+		ret = C.interpolate_scalars_cuda_{{.CurveNameLowerCase}}(d_out, scalarsC, twiddlesC, sizeC, 0, 0)
+	}
+	if ret != 0 {
+		fmt.Print("error interpolating")
+	}
+
+	return unsafe.Pointer(d_out)
+}
+
+func Evaluate(scalars_out, scalars, twiddles, coset_powers unsafe.Pointer, scalars_size, twiddles_size int, isCoset bool) int {
+	scalars_outC := (*C.{{.CurveNameUpperCase}}_scalar_t)(scalars_out)
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(scalars)
+	twiddlesC := (*C.{{.CurveNameUpperCase}}_scalar_t)(twiddles)
+	coset_powersC := (*C.{{.CurveNameUpperCase}}_scalar_t)(coset_powers)
+	sizeC := C.uint(scalars_size)
+	twiddlesC_size := C.uint(twiddles_size)
+
+	var ret C.int
+	if isCoset {
+		ret = C.evaluate_scalars_on_coset_cuda_{{.CurveNameLowerCase}}(scalars_outC, scalarsC, twiddlesC, twiddlesC_size, sizeC, coset_powersC, 0, 0)
+	} else {
+		ret = C.evaluate_scalars_cuda_{{.CurveNameLowerCase}}(scalars_outC, scalarsC, twiddlesC, twiddlesC_size, sizeC, 0, 0)
+	}
+
+	if ret != 0 {
+		fmt.Print("error interpolating")
+		return -1
+	}
+
+	return 0
+}
+
+func VecScalarAdd(in1_d, in2_d unsafe.Pointer, size int) int {
+	in1_dC := (*C.{{.CurveNameUpperCase}}_scalar_t)(in1_d)
+	in2_dC := (*C.{{.CurveNameUpperCase}}_scalar_t)(in2_d)
+	sizeC := C.uint(size)
+
+	ret := C.add_scalars_cuda_{{.CurveNameLowerCase}}(in1_dC, in1_dC, in2_dC, sizeC, 0)
+
+	if ret != 0 {
+		fmt.Print("error adding scalar vectors")
+		return -1
+	}
+
+	return 0
+}
+
+func VecScalarSub(in1_d, in2_d unsafe.Pointer, size int) int {
+	in1_dC := (*C.{{.CurveNameUpperCase}}_scalar_t)(in1_d)
+	in2_dC := (*C.{{.CurveNameUpperCase}}_scalar_t)(in2_d)
+	sizeC := C.uint(size)
+
+	ret := C.sub_scalars_cuda_{{.CurveNameLowerCase}}(in1_dC, in1_dC, in2_dC, sizeC, 0)
+
+	if ret != 0 {
+		fmt.Print("error subtracting scalar vectors")
+		return -1
+	}
+
+	return 0
+}
+
+func ToMontgomery(d_scalars unsafe.Pointer, len int) (int, error) {
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(d_scalars)
+	lenC := C.uint(len)
+	if success := C.to_montgomery_scalars_cuda_{{.CurveNameLowerCase}}(scalarsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func FromMontgomery(d_scalars unsafe.Pointer, len int) (int, error) {
+	scalarsC := (*C.{{.CurveNameUpperCase}}_scalar_t)(d_scalars)
+	lenC := C.uint(len)
+	if success := C.from_montgomery_scalars_cuda_{{.CurveNameLowerCase}}(scalarsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func AffinePointFromMontgomery(d_points unsafe.Pointer, len int) (int, error) {
+	pointsC := (*C.{{.CurveNameUpperCase}}_affine_t)(d_points)
+	lenC := C.uint(len)
+
+	if success := C.from_montgomery_aff_points_cuda_{{.CurveNameLowerCase}}(pointsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}
+
+func G2AffinePointFromMontgomery(d_points unsafe.Pointer, len int) (int, error) {
+	pointsC := (*C.{{.CurveNameUpperCase}}_g2_affine_t)(d_points)
+	lenC := C.uint(len)
+
+	if success := C.from_montgomery_aff_points_g2_cuda_{{.CurveNameLowerCase}}(pointsC, lenC, 0); success != 0 {
+		return -1, errors.New("reversing failed")
+	}
+	return 0, nil
+}

goicicle/templates/ntt/ntt_test.go.tmpl

@@ -0,0 +1,130 @@
+import (
+	"fmt"
+	"github.com/stretchr/testify/assert"
+	"reflect"
+	"testing"
+)
+
+func TestNtt{{.CurveNameUpperCase}}Batch(t *testing.T) {
+	count := 1 << 20
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	NttBatch(&nttResult, false, count, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	assert.Equal(t, nttResult, nttResult)
+}
+
+func TestNtt{{.CurveNameUpperCase}}CompareToGnarkDIF(t *testing.T) {
+	count := 1 << 2
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	Ntt(&nttResult, false, DIF, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	assert.Equal(t, nttResult, nttResult)
+}
+
+func TestINtt{{.CurveNameUpperCase}}CompareToGnarkDIT(t *testing.T) {
+	count := 1 << 3
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	Ntt(&nttResult, true, DIT, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	assert.Equal(t, nttResult, nttResult)
+}
+
+func TestNtt{{.CurveNameUpperCase}}(t *testing.T) {
+	count := 1 << 3
+
+	scalars := GenerateScalars(count, false)
+
+	nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+	copy(nttResult, scalars)
+
+	assert.Equal(t, nttResult, scalars)
+	Ntt(&nttResult, false, NONE, 0)
+	assert.NotEqual(t, nttResult, scalars)
+
+	inttResult := make([]G1ScalarField, len(nttResult))
+	copy(inttResult, nttResult)
+
+	assert.Equal(t, inttResult, nttResult)
+	Ntt(&inttResult, true, NONE, 0)
+	assert.Equal(t, inttResult, scalars)
+}
+
+func TestNttBatch{{.CurveNameUpperCase}}(t *testing.T) {
+	count := 1 << 5
+	batches := 4
+
+	scalars := GenerateScalars(count*batches, false)
+
+	var scalarVecOfVec [][]G1ScalarField = make([][]G1ScalarField, 0)
+
+	for i := 0; i < batches; i++ {
+		start := i * count
+		end := (i + 1) * count
+		batch := make([]G1ScalarField, len(scalars[start:end]))
+		copy(batch, scalars[start:end])
+		scalarVecOfVec = append(scalarVecOfVec, batch)
+	}
+
+	nttBatchResult := make([]G1ScalarField, len(scalars))
+	copy(nttBatchResult, scalars)
+
+	NttBatch(&nttBatchResult, false, count, 0)
+
+	var nttResultVecOfVec [][]G1ScalarField
+
+	for i := 0; i < batches; i++ {
+		// Clone the slice
+		clone := make([]G1ScalarField, len(scalarVecOfVec[i]))
+		copy(clone, scalarVecOfVec[i])
+
+		// Add it to the result vector of vectors
+		nttResultVecOfVec = append(nttResultVecOfVec, clone)
+
+		// Call the ntt_{{.CurveNameLowerCase}} function
+		Ntt(&nttResultVecOfVec[i], false, NONE, 0)
+	}
+
+	assert.NotEqual(t, nttBatchResult, scalars)
+
+	// Check that the ntt of each vec of scalars is equal to the intt of the specific batch
+	for i := 0; i < batches; i++ {
+		if !reflect.DeepEqual(nttResultVecOfVec[i], nttBatchResult[i*count:((i+1)*count)]) {
+			t.Errorf("ntt of vec of scalars not equal to intt of specific batch")
+		}
+	}
+}
+
+func BenchmarkNTT(b *testing.B) {
+	LOG_NTT_SIZES := []int{12, 15, 20, 21, 22, 23, 24, 25, 26}
+
+	for _, logNTTSize := range LOG_NTT_SIZES {
+		nttSize := 1 << logNTTSize
+		b.Run(fmt.Sprintf("NTT %d", logNTTSize), func(b *testing.B) {
+			scalars := GenerateScalars(nttSize, false)
+
+			nttResult := make([]G1ScalarField, len(scalars)) // Make a new slice with the same length
+			copy(nttResult, scalars)
+			for n := 0; n < b.N; n++ {
+				Ntt(&nttResult, false, NONE, 0)
+			}
+		})
+	}
+}

goicicle/templates/ops/vec_mod.go.tmpl

@@ -0,0 +1,23 @@
+// #cgo CFLAGS: -I./include/
+// #cgo LDFLAGS: -L${SRCDIR}/../../ {{.SharedLib}}
+// #include "ve_mod_mult.h"
+import "C"
+import (
+	"fmt"
+	"unsafe"
+)
+
+func VecScalarMulMod(scalarVec1, scalarVec2 unsafe.Pointer, size int) int {
+	scalarVec1C := (*C.{{.CurveNameUpperCase}}_scalar_t)(scalarVec1)
+	scalarVec2C := (*C.{{.CurveNameUpperCase}}_scalar_t)(scalarVec2)
+	sizeC := C.size_t(size)
+
+	ret := C.vec_mod_mult_device_scalar_{{.CurveNameLowerCase}}(scalarVec1C, scalarVec2C, sizeC, 0)
+
+	if ret != 0 {
+		fmt.Print("error multiplying scalar vectors")
+		return -1
+	}
+
+	return 0
+}

icicle/CMakeLists.txt

@@ -1,8 +1,8 @@
-cmake_minimum_required(VERSION 3.16)
+cmake_minimum_required(VERSION 3.18)
 
 # GoogleTest requires at least C++14
 set(CMAKE_CXX_STANDARD 17)
-set(CMAKE_CUDA_STANDARD 14)
+set(CMAKE_CUDA_STANDARD 17)
 set(CMAKE_CUDA_STANDARD_REQUIRED TRUE)
 set(CMAKE_CXX_STANDARD_REQUIRED TRUE)
 # add the target cuda architectures
@@ -22,6 +22,10 @@ FetchContent_Declare(
   URL https://github.com/google/googletest/archive/refs/tags/v1.13.0.zip
 )
 # For Windows: Prevent overriding the parent project's compiler/linker settings
+
+# boosting lib
+include_directories("/home/miner/include/boost_1_80_0")
+
 set(gtest_force_shared_crt ON CACHE BOOL "" FORCE)
 FetchContent_MakeAvailable(googletest)
 

icicle/appUtils/msm/msm.cuh

@@ -3,19 +3,46 @@
 #pragma once
 
 template <typename S, typename P, typename A>
-void bucket_method_msm(unsigned bitsize, unsigned c, S *scalars, A *points, unsigned size, P* final_result, bool on_device, cudaStream_t stream);
+void bucket_method_msm(
+  unsigned bitsize,
+  unsigned c,
+  S* scalars,
+  A* points,
+  unsigned size,
+  P* final_result,
+  bool on_device,
+  bool big_triangle,
+  cudaStream_t stream);
 
 template <typename S, typename P, typename A>
-void batched_bucket_method_msm(unsigned bitsize, unsigned c, S *scalars, A *points, unsigned batch_size, unsigned msm_size, P* final_results, bool on_device, cudaStream_t stream);
+void batched_bucket_method_msm(
+  unsigned bitsize,
+  unsigned c,
+  S* scalars,
+  A* points,
+  unsigned batch_size,
+  unsigned msm_size,
+  P* final_results,
+  bool on_device,
+  cudaStream_t stream);
 
 template <typename S, typename P, typename A>
-void batched_large_msm(S* scalars, A* points, unsigned batch_size, unsigned msm_size, P* result, bool on_device, cudaStream_t stream);
+void batched_large_msm(
+  S* scalars, A* points, unsigned batch_size, unsigned msm_size, P* result, bool on_device, cudaStream_t stream);
 
 template <typename S, typename P, typename A>
-void large_msm(S* scalars, A* points, unsigned size, P* result, bool on_device, cudaStream_t stream);
+void large_msm(
+  S* scalars,
+  A* points,
+  unsigned size,
+  P* result,
+  bool on_device,
+  bool big_triangle,
+  unsigned large_bucket_factor,
+  cudaStream_t stream);
 
 template <typename S, typename P, typename A>
-void short_msm(S *h_scalars, A *h_points, unsigned size, P* h_final_result, cudaStream_t stream);
+void short_msm(S* h_scalars, A* h_points, unsigned size, P* h_final_result, cudaStream_t stream);
 
 template <typename A, typename S, typename P>
 void reference_msm(S* scalars, A* a_points, unsigned size);

icicle/appUtils/msm/tests/msm_test.cu

@@ -1,113 +1,115 @@
-#include <iostream>
-#include <chrono>
-#include <vector>
-#include "msm.cu"
-#include "../../utils/cuda_utils.cuh"
-#include "../../primitives/projective.cuh"
 #include "../../primitives/field.cuh"
-#include "../../curves/bls12_381/curve_config.cuh"
+#include "../../primitives/projective.cuh"
+#include "../../utils/cuda_utils.cuh"
+#include "msm.cu"
+#include <chrono>
+#include <iostream>
+#include <vector>
+// #include "../../curves/bls12_377/curve_config.cuh"
+#include "../../curves/bn254/curve_config.cuh"
 
-using namespace BLS12_381;
+// using namespace BLS12_377;
+using namespace BN254;
 
-class Dummy_Scalar {
-  public:
-    static constexpr unsigned NBITS = 32;
+class Dummy_Scalar
+{
+public:
+  static constexpr unsigned NBITS = 32;
 
-    unsigned x;
+  unsigned x;
+  unsigned p = 10;
+  // unsigned p = 1<<30;
 
-    friend HOST_INLINE std::ostream& operator<<(std::ostream& os, const Dummy_Scalar& scalar) {
-      os << scalar.x;
-      return os;
-    }
+  static HOST_DEVICE_INLINE Dummy_Scalar zero() { return {0}; }
 
-    HOST_DEVICE_INLINE unsigned get_scalar_digit(unsigned digit_num, unsigned digit_width) {
-      return (x>>(digit_num*digit_width))&((1<<digit_width)-1);
-    }
+  static HOST_DEVICE_INLINE Dummy_Scalar one() { return {1}; }
 
-    friend HOST_DEVICE_INLINE Dummy_Scalar operator+(Dummy_Scalar p1, const Dummy_Scalar& p2) {   
-      return {p1.x+p2.x};
-    }
+  friend HOST_INLINE std::ostream& operator<<(std::ostream& os, const Dummy_Scalar& scalar)
+  {
+    os << scalar.x;
+    return os;
+  }
 
-    friend HOST_DEVICE_INLINE bool operator==(const Dummy_Scalar& p1, const Dummy_Scalar& p2) {
-      return (p1.x == p2.x);
-    }
+  HOST_DEVICE_INLINE unsigned get_scalar_digit(unsigned digit_num, unsigned digit_width)
+  {
+    return (x >> (digit_num * digit_width)) & ((1 << digit_width) - 1);
+  }
 
-    friend HOST_DEVICE_INLINE bool operator==(const Dummy_Scalar& p1, const unsigned p2) {
-      return (p1.x == p2);
-    }
+  friend HOST_DEVICE_INLINE Dummy_Scalar operator+(Dummy_Scalar p1, const Dummy_Scalar& p2)
+  {
+    return {(p1.x + p2.x) % p1.p};
+  }
 
-    // static HOST_DEVICE_INLINE Dummy_Scalar neg(const Dummy_Scalar &scalar) { 
-    //   return {Dummy_Scalar::neg(point.x)};
-    // }
-    static HOST_INLINE Dummy_Scalar rand_host() {
-      return {(unsigned)rand()};
-    }
+  friend HOST_DEVICE_INLINE bool operator==(const Dummy_Scalar& p1, const Dummy_Scalar& p2) { return (p1.x == p2.x); }
+
+  friend HOST_DEVICE_INLINE bool operator==(const Dummy_Scalar& p1, const unsigned p2) { return (p1.x == p2); }
+
+  static HOST_DEVICE_INLINE Dummy_Scalar neg(const Dummy_Scalar& scalar) { return {scalar.p - scalar.x}; }
+  static HOST_INLINE Dummy_Scalar rand_host()
+  {
+    return {(unsigned)rand() % 10};
+    // return {(unsigned)rand()};
+  }
 };
 
-class Dummy_Projective {
+class Dummy_Projective
+{
+public:
+  Dummy_Scalar x;
 
-  public:
-    Dummy_Scalar x;
+  static HOST_DEVICE_INLINE Dummy_Projective zero() { return {0}; }
 
-    static HOST_DEVICE_INLINE Dummy_Projective zero() {
-      return {0};
+  static HOST_DEVICE_INLINE Dummy_Projective one() { return {1}; }
+
+  static HOST_DEVICE_INLINE Dummy_Projective to_affine(const Dummy_Projective& point) { return {point.x}; }
+
+  static HOST_DEVICE_INLINE Dummy_Projective from_affine(const Dummy_Projective& point) { return {point.x}; }
+
+  static HOST_DEVICE_INLINE Dummy_Projective neg(const Dummy_Projective& point) { return {Dummy_Scalar::neg(point.x)}; }
+
+  friend HOST_DEVICE_INLINE Dummy_Projective operator+(Dummy_Projective p1, const Dummy_Projective& p2)
+  {
+    return {p1.x + p2.x};
+  }
+
+  // friend HOST_DEVICE_INLINE Dummy_Projective operator-(Dummy_Projective p1, const Dummy_Projective& p2) {
+  //   return p1 + neg(p2);
+  // }
+
+  friend HOST_INLINE std::ostream& operator<<(std::ostream& os, const Dummy_Projective& point)
+  {
+    os << point.x;
+    return os;
+  }
+
+  friend HOST_DEVICE_INLINE Dummy_Projective operator*(Dummy_Scalar scalar, const Dummy_Projective& point)
+  {
+    Dummy_Projective res = zero();
+#ifdef CUDA_ARCH
+#pragma unroll
+#endif
+    for (int i = 0; i < Dummy_Scalar::NBITS; i++) {
+      if (i > 0) { res = res + res; }
+      if (scalar.get_scalar_digit(Dummy_Scalar::NBITS - i - 1, 1)) { res = res + point; }
     }
+    return res;
+  }
 
-    static HOST_DEVICE_INLINE Dummy_Projective to_affine(const Dummy_Projective &point) {
-      return {point.x};
-    }
+  friend HOST_DEVICE_INLINE bool operator==(const Dummy_Projective& p1, const Dummy_Projective& p2)
+  {
+    return (p1.x == p2.x);
+  }
 
-    static HOST_DEVICE_INLINE Dummy_Projective from_affine(const Dummy_Projective &point) {
-      return {point.x};
-    }
+  static HOST_DEVICE_INLINE bool is_zero(const Dummy_Projective& point) { return point.x == 0; }
 
-    // static HOST_DEVICE_INLINE Dummy_Projective neg(const Dummy_Projective &point) { 
-    //   return {Dummy_Scalar::neg(point.x)};
-    // }
-
-    friend HOST_DEVICE_INLINE Dummy_Projective operator+(Dummy_Projective p1, const Dummy_Projective& p2) {   
-      return {p1.x+p2.x};
-    }
-
-    // friend HOST_DEVICE_INLINE Dummy_Projective operator-(Dummy_Projective p1, const Dummy_Projective& p2) {   
-    //   return p1 + neg(p2);
-    // }
-
-    friend HOST_INLINE std::ostream& operator<<(std::ostream& os, const Dummy_Projective& point) {
-      os << point.x;
-      return os;
-    }
-
-    friend HOST_DEVICE_INLINE Dummy_Projective operator*(Dummy_Scalar scalar, const Dummy_Projective& point) {   
-      Dummy_Projective res = zero();
-  #ifdef CUDA_ARCH
-  #pragma unroll
-  #endif
-      for (int i = 0; i < Dummy_Scalar::NBITS; i++) {
-        if (i > 0) {
-          res = res + res;
-        }
-        if (scalar.get_scalar_digit(Dummy_Scalar::NBITS - i - 1, 1)) {
-          res = res + point;
-        }
-      }
-      return res;
-    }
-
-    friend HOST_DEVICE_INLINE bool operator==(const Dummy_Projective& p1, const Dummy_Projective& p2) {
-      return (p1.x == p2.x);
-    }
-
-    static HOST_DEVICE_INLINE bool is_zero(const Dummy_Projective &point) {
-      return point.x == 0;
-    }
-
-    static HOST_INLINE Dummy_Projective rand_host() {
-      return {(unsigned)rand()};
-    }
+  static HOST_INLINE Dummy_Projective rand_host()
+  {
+    return {(unsigned)rand() % 10};
+    // return {(unsigned)rand()};
+  }
 };
 
-//switch between dummy and real:
+// switch between dummy and real:
 
 typedef scalar_t test_scalar;
 typedef projective_t test_projective;
@@ -119,63 +121,100 @@ typedef affine_t test_affine;
 
 int main()
 {
-  unsigned batch_size = 4;
-  unsigned msm_size = 1<<15;
-  unsigned N = batch_size*msm_size;
+  unsigned batch_size = 1;
+  //   unsigned msm_size = 1<<21;
+  unsigned msm_size = 12180757;
+  unsigned N = batch_size * msm_size;
 
-  test_scalar *scalars = new test_scalar[N];
-  test_affine *points = new test_affine[N];
-  
-  for (unsigned i=0;i<N;i++){
-    scalars[i] = (i%msm_size < 10)? test_scalar::rand_host() : scalars[i-10];
-    points[i] = (i%msm_size < 10)? test_projective::to_affine(test_projective::rand_host()): points[i-10];
-    // scalars[i] = test_scalar::rand_host();
+  test_scalar* scalars = new test_scalar[N];
+  test_affine* points = new test_affine[N];
+
+  for (unsigned i = 0; i < N; i++) {
+    // scalars[i] = (i%msm_size < 10)? test_scalar::rand_host() : scalars[i-10];
+    points[i] = (i % msm_size < 10) ? test_projective::to_affine(test_projective::rand_host()) : points[i - 10];
+    scalars[i] = test_scalar::rand_host();
+    // scalars[i] = i < N/2? test_scalar::rand_host() : test_scalar::one();
     // points[i] = test_projective::to_affine(test_projective::rand_host());
   }
-  std::cout<<"finished generating"<<std::endl;
+  std::cout << "finished generating" << std::endl;
 
   // projective_t *short_res = (projective_t*)malloc(sizeof(projective_t));
   // test_projective *large_res = (test_projective*)malloc(sizeof(test_projective));
-  test_projective large_res[batch_size];
-  test_projective batched_large_res[batch_size];
+  test_projective large_res[batch_size * 2];
+  // test_projective batched_large_res[batch_size];
   // fake_point *large_res = (fake_point*)malloc(sizeof(fake_point));
   // fake_point batched_large_res[256];
 
-
   // short_msm<scalar_t, projective_t, affine_t>(scalars, points, N, short_res);
-  for (unsigned i=0;i<batch_size;i++){
-    large_msm<test_scalar, test_projective, test_affine>(scalars+msm_size*i, points+msm_size*i, msm_size, large_res+i, false);
-    // std::cout<<"final result large"<<std::endl;
-    // std::cout<<test_projective::to_affine(*large_res)<<std::endl;
-  }
+  // for (unsigned i=0;i<batch_size;i++){
+  // large_msm<test_scalar, test_projective, test_affine>(scalars+msm_size*i, points+msm_size*i, msm_size, large_res+i,
+  // false); std::cout<<"final result large"<<std::endl; std::cout<<test_projective::to_affine(*large_res)<<std::endl;
+  // }
+
+  test_scalar* scalars_d;
+  test_affine* points_d;
+  test_projective* large_res_d;
+
+  cudaMalloc(&scalars_d, sizeof(test_scalar) * msm_size);
+  cudaMalloc(&points_d, sizeof(test_affine) * msm_size);
+  cudaMalloc(&large_res_d, sizeof(test_projective));
+  cudaMemcpy(scalars_d, scalars, sizeof(test_scalar) * msm_size, cudaMemcpyHostToDevice);
+  cudaMemcpy(points_d, points, sizeof(test_affine) * msm_size, cudaMemcpyHostToDevice);
+
+  std::cout << "finished copying" << std::endl;
+
+  // batched_large_msm<test_scalar, test_projective, test_affine>(scalars, points, batch_size, msm_size,
+  // batched_large_res, false);
+  cudaStream_t stream1;
+  cudaStream_t stream2;
+  cudaStreamCreate(&stream1);
+  cudaStreamCreate(&stream2);
+  auto begin1 = std::chrono::high_resolution_clock::now();
+  large_msm<test_scalar, test_projective, test_affine>(scalars, points, msm_size, large_res, false, true, stream1);
+  auto end1 = std::chrono::high_resolution_clock::now();
+  auto elapsed1 = std::chrono::duration_cast<std::chrono::nanoseconds>(end1 - begin1);
+  printf("Big Triangle : %.3f seconds.\n", elapsed1.count() * 1e-9);
+  // std::cout<<test_projective::to_affine(large_res[0])<<std::endl;
   auto begin = std::chrono::high_resolution_clock::now();
-  batched_large_msm<test_scalar, test_projective, test_affine>(scalars, points, batch_size, msm_size, batched_large_res, false);
-  // large_msm<test_scalar, test_projective, test_affine>(scalars, points, msm_size, large_res, false);
+  large_msm<test_scalar, test_projective, test_affine>(
+    scalars_d, points_d, msm_size, large_res_d, true, false, stream2);
+  // test_reduce_triangle(scalars);
+  // test_reduce_rectangle(scalars);
+  // test_reduce_single(scalars);
+  // test_reduce_var(scalars);
   auto end = std::chrono::high_resolution_clock::now();
   auto elapsed = std::chrono::duration_cast<std::chrono::nanoseconds>(end - begin);
-  printf("Time measured: %.3f seconds.\n", elapsed.count() * 1e-9);
-  std::cout<<test_projective::to_affine(large_res[0])<<std::endl;
+  printf("On Device No Big Triangle: %.3f seconds.\n", elapsed.count() * 1e-9);
+  cudaStreamSynchronize(stream1);
+  cudaStreamSynchronize(stream2);
+  cudaStreamDestroy(stream1);
+  cudaStreamDestroy(stream2);
 
-  // reference_msm<test_affine, test_scalar, test_projective>(scalars, points, msm_size);
+  std::cout << test_projective::to_affine(large_res[0]) << std::endl;
+
+  cudaMemcpy(&large_res[1], large_res_d, sizeof(test_projective), cudaMemcpyDeviceToHost);
+  std::cout << test_projective::to_affine(large_res[1]) << std::endl;
+
+  //   reference_msm<test_affine, test_scalar, test_projective>(scalars, points, msm_size);
+
+  // std::cout<<"final results batched large"<<std::endl;
+  // bool success = true;
+  // for (unsigned i = 0; i < batch_size; i++)
+  // {
+  //   std::cout<<test_projective::to_affine(batched_large_res[i])<<std::endl;
+  //   if (test_projective::to_affine(large_res[i])==test_projective::to_affine(batched_large_res[i])){
+  //     std::cout<<"good"<<std::endl;
+  //   }
+  //   else{
+  //     std::cout<<"miss"<<std::endl;
+  //     std::cout<<test_projective::to_affine(large_res[i])<<std::endl;
+  //     success = false;
+  //   }
+  // }
+  // if (success){
+  //   std::cout<<"success!"<<std::endl;
+  // }
 
-  std::cout<<"final results batched large"<<std::endl;
-  bool success = true;
-  for (unsigned i = 0; i < batch_size; i++)
-  {
-    std::cout<<test_projective::to_affine(batched_large_res[i])<<std::endl;
-    if (test_projective::to_affine(large_res[i])==test_projective::to_affine(batched_large_res[i])){
-      std::cout<<"good"<<std::endl;
-    }
-    else{
-      std::cout<<"miss"<<std::endl;
-      std::cout<<test_projective::to_affine(large_res[i])<<std::endl;
-      success = false;
-    }
-  }
-  if (success){
-    std::cout<<"success!"<<std::endl;
-  }
-  
   // std::cout<<batched_large_res[0]<<std::endl;
   // std::cout<<batched_large_res[1]<<std::endl;
   // std::cout<<projective_t::to_affine(batched_large_res[0])<<std::endl;

icicle/appUtils/ntt/lde.cu

@@ -1,81 +1,125 @@
 #ifndef LDE
 #define LDE
-#include <cuda.h>
-#include "ntt.cuh"
-#include "lde.cuh"
 #include "../vector_manipulation/ve_mod_mult.cuh"
+#include "lde.cuh"
+#include "ntt.cuh"
+#include <cuda.h>
 
+template <typename E, bool SUB>
+__global__ void add_sub_array(E* res, E* in1, E* in2, uint32_t n)
+{
+  int tid = (blockIdx.x * blockDim.x) + threadIdx.x;
+  if (tid < n) { res[tid] = SUB ? in1[tid] - in2[tid] : in1[tid] + in2[tid]; }
+}
+
+template <typename E>
+int sub_polys(E* d_out, E* d_in1, E* d_in2, unsigned n, cudaStream_t stream)
+{
+  uint32_t NUM_THREADS = MAX_THREADS_BATCH;
+  uint32_t NUM_BLOCKS = (n + NUM_THREADS - 1) / NUM_THREADS;
+
+  add_sub_array<E, true><<<NUM_BLOCKS, NUM_THREADS, 0, stream>>>(d_out, d_in1, d_in2, n);
+
+  return 0;
+}
+
+template <typename E>
+int add_polys(E* d_out, E* d_in1, E* d_in2, unsigned n, cudaStream_t stream)
+{
+  uint32_t NUM_THREADS = MAX_THREADS_BATCH;
+  uint32_t NUM_BLOCKS = (n + NUM_THREADS - 1) / NUM_THREADS;
+
+  add_sub_array<E, false><<<NUM_BLOCKS, NUM_THREADS, 0, stream>>>(d_out, d_in1, d_in2, n);
+
+  return 0;
+}
 
 /**
  * Interpolate a batch of polynomials from their evaluations on the same subgroup.
  * Note: this function does not preform any bit-reverse permutations on its inputs or outputs.
- * @param d_out The variable to write coefficients of the resulting polynomials into (the coefficients are in bit-reversed order if the evaluations weren't bit-reversed and vice-versa).
+ * @param d_out The variable to write coefficients of the resulting polynomials into (the coefficients are in
+ * bit-reversed order if the evaluations weren't bit-reversed and vice-versa).
  * @param d_evaluations Input array of evaluations of all polynomials of type E (elements).
  * @param d_domain Domain on which the polynomials are evaluated. Must be a subgroup.
  * @param n Length of `d_domain` array, also equal to the number of evaluations of each polynomial.
  * @param batch_size The size of the batch; the length of `d_evaluations` is `n` * `batch_size`.
  */
-template <typename E, typename S> int interpolate_batch(E * d_out, E * d_evaluations, S * d_domain, unsigned n, unsigned batch_size, cudaStream_t stream) {
-  uint32_t logn = uint32_t(log(n) / log(2));
+template <typename E, typename S>
+int interpolate_batch(
+  E* d_out,
+  E* d_evaluations,
+  S* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  bool coset,
+  S* coset_powers,
+  cudaStream_t stream)
+{
   cudaMemcpyAsync(d_out, d_evaluations, sizeof(E) * n * batch_size, cudaMemcpyDeviceToDevice, stream);
-  
-  int NUM_THREADS = min(n / 2, MAX_THREADS_BATCH);
-  int NUM_BLOCKS = batch_size * max(int((n / 2) / NUM_THREADS), 1);
-  for (uint32_t s = 0; s < logn; s++) //TODO: this loop also can be unrolled
-  {
-    ntt_template_kernel <E, S> <<<NUM_BLOCKS, NUM_THREADS, 0, stream>>>(d_out, n, d_domain, n, NUM_BLOCKS, s, false);
-  }
-
-  NUM_BLOCKS = (n * batch_size + NUM_THREADS - 1) / NUM_THREADS;
-  template_normalize_kernel <E, S> <<<NUM_BLOCKS, NUM_THREADS, 0, stream>>> (d_out, n * batch_size, S::inv_log_size(logn));
-  cudaStreamSynchronize(stream);
+  ntt_inplace_batch_template(d_out, d_domain, n, batch_size, true, coset, coset_powers, stream, true);
   return 0;
 }
 
 /**
  * Interpolate a polynomial from its evaluations on a subgroup.
  * Note: this function does not preform any bit-reverse permutations on its inputs or outputs.
- * @param d_out The variable to write coefficients of the resulting polynomial into (the coefficients are in bit-reversed order if the evaluations weren't bit-reversed and vice-versa).
+ * @param d_out The variable to write coefficients of the resulting polynomial into (the coefficients are in
+ * bit-reversed order if the evaluations weren't bit-reversed and vice-versa).
  * @param d_evaluations Input array of evaluations that have type E (elements).
  * @param d_domain Domain on which the polynomial is evaluated. Must be a subgroup.
  * @param n Length of `d_evaluations` and the size `d_domain` arrays (they should have equal length).
  */
-template <typename E, typename S> int interpolate(E * d_out, E * d_evaluations, S * d_domain, unsigned n, cudaStream_t stream) {
-  return interpolate_batch <E, S> (d_out, d_evaluations, d_domain, n, 1, stream);
+template <typename E, typename S>
+int interpolate(E* d_out, E* d_evaluations, S* d_domain, unsigned n, bool coset, S* coset_powers, cudaStream_t stream)
+{
+  return interpolate_batch<E, S>(d_out, d_evaluations, d_domain, n, 1, coset, coset_powers, stream);
 }
 
-template < typename E > __global__ void fill_array(E * arr, E val, uint32_t n) {
+template <typename E>
+__global__ void fill_array(E* arr, E val, uint32_t n)
+{
   int tid = (blockIdx.x * blockDim.x) + threadIdx.x;
-  if (tid < n) {
-    arr[tid] = val;
-  }
+  if (tid < n) { arr[tid] = val; }
 }
 
 /**
  * Evaluate a batch of polynomials on the same coset.
  * @param d_out The evaluations of the polynomials on coset `u` * `d_domain`.
- * @param d_coefficients Input array of coefficients of all polynomials of type E (elements) to be evaluated in-place on a coset.
+ * @param d_coefficients Input array of coefficients of all polynomials of type E (elements) to be evaluated in-place on
+ * a coset.
  * @param d_domain Domain on which the polynomials are evaluated (see `coset` flag). Must be a subgroup.
  * @param domain_size Length of `d_domain` array, on which the polynomial is computed.
  * @param n The number of coefficients, which might be different from `domain_size`.
  * @param batch_size The size of the batch; the length of `d_coefficients` is `n` * `batch_size`.
  * @param coset The flag that indicates whether to evaluate on a coset. If false, evaluate on a subgroup `d_domain`.
- * @param coset_powers If `coset` is true, a list of powers `[1, u, u^2, ..., u^{n-1}]` where `u` is the generator of the coset.
+ * @param coset_powers If `coset` is true, a list of powers `[1, u, u^2, ..., u^{n-1}]` where `u` is the generator of
+ * the coset.
  */
 template <typename E, typename S>
-int evaluate_batch(E * d_out, E * d_coefficients, S * d_domain, unsigned domain_size, unsigned n, unsigned batch_size, bool coset, S * coset_powers, cudaStream_t stream) {
+int evaluate_batch(
+  E* d_out,
+  E* d_coefficients,
+  S* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  bool coset,
+  S* coset_powers,
+  cudaStream_t stream)
+{
   uint32_t logn = uint32_t(log(domain_size) / log(2));
   if (domain_size > n) {
     // allocate and initialize an array of stream handles to parallelize data copying across batches
-    cudaStream_t *memcpy_streams = (cudaStream_t *) malloc(batch_size * sizeof(cudaStream_t));
-    for (unsigned i = 0; i < batch_size; i++)
-    {
+    cudaStream_t* memcpy_streams = (cudaStream_t*)malloc(batch_size * sizeof(cudaStream_t));
+    for (unsigned i = 0; i < batch_size; i++) {
       cudaStreamCreate(&(memcpy_streams[i]));
 
-      cudaMemcpyAsync(&d_out[i * domain_size], &d_coefficients[i * n], n * sizeof(E), cudaMemcpyDeviceToDevice, memcpy_streams[i]);
+      cudaMemcpyAsync(
+        &d_out[i * domain_size], &d_coefficients[i * n], n * sizeof(E), cudaMemcpyDeviceToDevice, memcpy_streams[i]);
       uint32_t NUM_THREADS = MAX_THREADS_BATCH;
       uint32_t NUM_BLOCKS = (domain_size - n + NUM_THREADS - 1) / NUM_THREADS;
-      fill_array <E> <<<NUM_BLOCKS, NUM_THREADS, 0, memcpy_streams[i]>>> (&d_out[i * domain_size + n], E::zero(), domain_size - n);
+      fill_array<E>
+        <<<NUM_BLOCKS, NUM_THREADS, 0, memcpy_streams[i]>>>(&d_out[i * domain_size + n], E::zero(), domain_size - n);
 
       cudaStreamSynchronize(memcpy_streams[i]);
       cudaStreamDestroy(memcpy_streams[i]);
@@ -83,102 +127,153 @@ int evaluate_batch(E * d_out, E * d_coefficients, S * d_domain, unsigned domain_
   } else
     cudaMemcpyAsync(d_out, d_coefficients, sizeof(E) * domain_size * batch_size, cudaMemcpyDeviceToDevice, stream);
 
-  if (coset)
-    batch_vector_mult(coset_powers, d_out, domain_size, batch_size, stream);
+  if (coset) batch_vector_mult(coset_powers, d_out, domain_size, batch_size, stream);
 
-  int NUM_THREADS = min(domain_size / 2, MAX_THREADS_BATCH);
-  int chunks = max(int((domain_size / 2) / NUM_THREADS), 1);
-  int NUM_BLOCKS = batch_size * chunks;
-  for (uint32_t s = 0; s < logn; s++) //TODO: this loop also can be unrolled
-  {
-    ntt_template_kernel <E, S> <<<NUM_BLOCKS, NUM_THREADS, 0, stream>>>(d_out, domain_size, d_domain, domain_size, batch_size * chunks, logn - s - 1, true);
-  }
-  cudaStreamSynchronize(stream);
+  S* _null = nullptr;
+  ntt_inplace_batch_template(d_out, d_domain, domain_size, batch_size, false, false, _null, stream, true);
   return 0;
 }
 
 /**
  * Evaluate a polynomial on a coset.
- * Note: this function does not preform any bit-reverse permutations on its inputs or outputs, so the order of outputs is bit-reversed.
+ * Note: this function does not preform any bit-reverse permutations on its inputs or outputs, so the order of outputs
+ * is bit-reversed.
  * @param d_out The evaluations of the polynomial on coset `u` * `d_domain`.
  * @param d_coefficients Input array of coefficients of a polynomial of type E (elements).
  * @param d_domain Domain on which the polynomial is evaluated (see `coset` flag). Must be a subgroup.
  * @param domain_size Length of `d_domain` array, on which the polynomial is computed.
  * @param n The number of coefficients, which might be different from `domain_size`.
  * @param coset The flag that indicates whether to evaluate on a coset. If false, evaluate on a subgroup `d_domain`.
- * @param coset_powers If `coset` is true, a list of powers `[1, u, u^2, ..., u^{n-1}]` where `u` is the generator of the coset.
+ * @param coset_powers If `coset` is true, a list of powers `[1, u, u^2, ..., u^{n-1}]` where `u` is the generator of
+ * the coset.
  */
-template <typename E, typename S> 
-int evaluate(E * d_out, E * d_coefficients, S * d_domain, unsigned domain_size, unsigned n, bool coset, S * coset_powers, cudaStream_t stream) {
-  return evaluate_batch <E, S> (d_out, d_coefficients, d_domain, domain_size, n, 1, coset, coset_powers, stream);
+template <typename E, typename S>
+int evaluate(
+  E* d_out,
+  E* d_coefficients,
+  S* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  bool coset,
+  S* coset_powers,
+  cudaStream_t stream)
+{
+  return evaluate_batch<E, S>(d_out, d_coefficients, d_domain, domain_size, n, 1, coset, coset_powers, stream);
 }
 
-template <typename S> 
-int interpolate_scalars(S* d_out, S* d_evaluations, S* d_domain, unsigned n, cudaStream_t stream) {
-  return interpolate(d_out, d_evaluations, d_domain, n, stream);
+template <typename S>
+int interpolate_scalars(S* d_out, S* d_evaluations, S* d_domain, unsigned n, cudaStream_t stream)
+{
+  S* _null = nullptr;
+  return interpolate(d_out, d_evaluations, d_domain, n, false, _null, stream);
 }
 
-template <typename S> 
-int interpolate_scalars_batch(S* d_out, S* d_evaluations, S* d_domain, unsigned n, unsigned batch_size, cudaStream_t stream) {
-  return interpolate_batch(d_out, d_evaluations, d_domain, n, batch_size, stream);
+template <typename S>
+int interpolate_scalars_batch(
+  S* d_out, S* d_evaluations, S* d_domain, unsigned n, unsigned batch_size, cudaStream_t stream)
+{
+  S* _null = nullptr;
+  return interpolate_batch(d_out, d_evaluations, d_domain, n, batch_size, false, _null, stream);
 }
 
-template <typename E, typename S> 
-int interpolate_points(E* d_out, E* d_evaluations, S* d_domain, unsigned n, cudaStream_t stream) {
-  return interpolate(d_out, d_evaluations, d_domain, n, stream);
+template <typename E, typename S>
+int interpolate_points(E* d_out, E* d_evaluations, S* d_domain, unsigned n, cudaStream_t stream)
+{
+  S* _null = nullptr;
+  return interpolate(d_out, d_evaluations, d_domain, n, false, _null, stream);
 }
 
-template <typename E, typename S> 
-int interpolate_points_batch(E* d_out, E* d_evaluations, S* d_domain, unsigned n, unsigned batch_size, cudaStream_t stream) {
-  return interpolate_batch(d_out, d_evaluations, d_domain, n, batch_size, stream);
+template <typename E, typename S>
+int interpolate_points_batch(
+  E* d_out, E* d_evaluations, S* d_domain, unsigned n, unsigned batch_size, cudaStream_t stream)
+{
+  S* _null = nullptr;
+  return interpolate_batch(d_out, d_evaluations, d_domain, n, batch_size, false, _null, stream);
 }
 
-template <typename S> 
-int evaluate_scalars(S* d_out, S* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, cudaStream_t stream) {
+template <typename S>
+int evaluate_scalars(S* d_out, S* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, cudaStream_t stream)
+{
   S* _null = nullptr;
   return evaluate(d_out, d_coefficients, d_domain, domain_size, n, false, _null, stream);
 }
 
-template <typename S> 
-int evaluate_scalars_batch(S* d_out, S* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, unsigned batch_size, cudaStream_t stream) {
+template <typename S>
+int evaluate_scalars_batch(
+  S* d_out, S* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, unsigned batch_size, cudaStream_t stream)
+{
   S* _null = nullptr;
   return evaluate_batch(d_out, d_coefficients, d_domain, domain_size, n, batch_size, false, _null, stream);
 }
 
-template <typename E, typename S> 
-int evaluate_points(E* d_out, E* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, cudaStream_t stream) {
+template <typename E, typename S>
+int evaluate_points(E* d_out, E* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, cudaStream_t stream)
+{
   S* _null = nullptr;
   return evaluate(d_out, d_coefficients, d_domain, domain_size, n, false, _null, stream);
 }
 
-template <typename E, typename S> 
-int evaluate_points_batch(E* d_out, E* d_coefficients, S* d_domain, 
-                          unsigned domain_size, unsigned n, unsigned batch_size, cudaStream_t stream) {
+template <typename E, typename S>
+int evaluate_points_batch(
+  E* d_out, E* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, unsigned batch_size, cudaStream_t stream)
+{
   S* _null = nullptr;
   return evaluate_batch(d_out, d_coefficients, d_domain, domain_size, n, batch_size, false, _null, stream);
 }
 
-template <typename S> 
-int evaluate_scalars_on_coset(S* d_out, S* d_coefficients, S* d_domain, 
-                              unsigned domain_size, unsigned n, S* coset_powers, cudaStream_t stream) {
+template <typename S>
+int interpolate_scalars_on_coset(
+  S* d_out, S* d_evaluations, S* d_domain, unsigned n, S* coset_powers, cudaStream_t stream)
+{
+  return interpolate(d_out, d_evaluations, d_domain, n, true, coset_powers, stream);
+}
+
+template <typename S>
+int interpolate_scalars_on_coset_batch(
+  S* d_out, S* d_evaluations, S* d_domain, unsigned n, unsigned batch_size, S* coset_powers, cudaStream_t stream)
+{
+  return interpolate_batch(d_out, d_evaluations, d_domain, n, batch_size, true, coset_powers, stream);
+}
+
+template <typename S>
+int evaluate_scalars_on_coset(
+  S* d_out, S* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, S* coset_powers, cudaStream_t stream)
+{
   return evaluate(d_out, d_coefficients, d_domain, domain_size, n, true, coset_powers, stream);
 }
 
-template <typename E, typename S> 
-int evaluate_scalars_on_coset_batch(S* d_out, S* d_coefficients, S* d_domain, unsigned domain_size, 
-                                    unsigned n, unsigned batch_size, S* coset_powers, cudaStream_t stream) {
+template <typename E, typename S>
+int evaluate_scalars_on_coset_batch(
+  S* d_out,
+  S* d_coefficients,
+  S* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  S* coset_powers,
+  cudaStream_t stream)
+{
   return evaluate_batch(d_out, d_coefficients, d_domain, domain_size, n, batch_size, true, coset_powers, stream);
 }
 
-template <typename E, typename S> 
-int evaluate_points_on_coset(E* d_out, E* d_coefficients, S* d_domain, 
-                             unsigned domain_size, unsigned n, S* coset_powers, cudaStream_t stream) {
+template <typename E, typename S>
+int evaluate_points_on_coset(
+  E* d_out, E* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, S* coset_powers, cudaStream_t stream)
+{
   return evaluate(d_out, d_coefficients, d_domain, domain_size, n, true, coset_powers, stream);
 }
 
-template <typename E, typename S> 
-int evaluate_points_on_coset_batch(E* d_out, E* d_coefficients, S* d_domain, unsigned domain_size,
-                                   unsigned n, unsigned batch_size, S* coset_powers, cudaStream_t stream) {
+template <typename E, typename S>
+int evaluate_points_on_coset_batch(
+  E* d_out,
+  E* d_coefficients,
+  S* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  S* coset_powers,
+  cudaStream_t stream)
+{
   return evaluate_batch(d_out, d_coefficients, d_domain, domain_size, n, batch_size, true, coset_powers, stream);
 }
 #endif

icicle/appUtils/ntt/lde.cuh

@@ -2,45 +2,62 @@
 #define LDE_H
 #pragma once
 
-template <typename S> 
+template <typename S>
 int interpolate_scalars(S* d_out, S* d_evaluations, S* d_domain, unsigned n, cudaStream_t stream);
 
-template <typename S> 
-int interpolate_scalars_batch(S* d_out, S* d_evaluations, S* d_domain, unsigned n, unsigned batch_size, cudaStream_t stream);
+template <typename S>
+int interpolate_scalars_batch(
+  S* d_out, S* d_evaluations, S* d_domain, unsigned n, unsigned batch_size, cudaStream_t stream);
 
-template <typename E, typename S> 
+template <typename E, typename S>
 int interpolate_points(E* d_out, E* d_evaluations, S* d_domain, unsigned n, cudaStream_t stream);
 
-template <typename E, typename S> 
-int interpolate_points_batch(E* d_out, E* d_evaluations, S* d_domain, unsigned n, unsigned batch_size, cudaStream_t stream);
+template <typename E, typename S>
+int interpolate_points_batch(
+  E* d_out, E* d_evaluations, S* d_domain, unsigned n, unsigned batch_size, cudaStream_t stream);
 
-template <typename S> 
+template <typename S>
 int evaluate_scalars(S* d_out, S* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, cudaStream_t stream);
 
-template <typename S> 
-int evaluate_scalars_batch(S* d_out, S* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, unsigned batch_size, cudaStream_t stream);
+template <typename S>
+int evaluate_scalars_batch(
+  S* d_out, S* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, unsigned batch_size, cudaStream_t stream);
 
-template <typename E, typename S> 
+template <typename E, typename S>
 int evaluate_points(E* d_out, E* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, cudaStream_t stream);
 
-template <typename E, typename S> 
-int evaluate_points_batch(E* d_out, E* d_coefficients, S* d_domain, 
-                          unsigned domain_size, unsigned n, unsigned batch_size, cudaStream_t stream);
+template <typename E, typename S>
+int evaluate_points_batch(
+  E* d_out, E* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, unsigned batch_size, cudaStream_t stream);
 
-template <typename S> 
-int evaluate_scalars_on_coset(S* d_out, S* d_coefficients, S* d_domain, 
-                              unsigned domain_size, unsigned n, S* coset_powers, cudaStream_t stream);
+template <typename S>
+int evaluate_scalars_on_coset(
+  S* d_out, S* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, S* coset_powers, cudaStream_t stream);
 
-template <typename S>                               
-int evaluate_scalars_on_coset_batch(S* d_out, S* d_coefficients, S* d_domain, unsigned domain_size, 
-                                    unsigned n, unsigned batch_size, S* coset_powers, cudaStream_t stream);
+template <typename S>
+int evaluate_scalars_on_coset_batch(
+  S* d_out,
+  S* d_coefficients,
+  S* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  S* coset_powers,
+  cudaStream_t stream);
 
-template <typename E, typename S> 
-int evaluate_points_on_coset(E* d_out, E* d_coefficients, S* d_domain, 
-                             unsigned domain_size, unsigned n, S* coset_powers, cudaStream_t stream);
+template <typename E, typename S>
+int evaluate_points_on_coset(
+  E* d_out, E* d_coefficients, S* d_domain, unsigned domain_size, unsigned n, S* coset_powers, cudaStream_t stream);
 
-template <typename E, typename S> 
-int evaluate_points_on_coset_batch(E* d_out, E* d_coefficients, S* d_domain, unsigned domain_size,
-                                   unsigned n, unsigned batch_size, S* coset_powers, cudaStream_t stream);
+template <typename E, typename S>
+int evaluate_points_on_coset_batch(
+  E* d_out,
+  E* d_coefficients,
+  S* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  S* coset_powers,
+  cudaStream_t stream);
 
 #endif

icicle/appUtils/ntt/ntt.cuh

@@ -2,8 +2,13 @@
 #define NTT
 #pragma once
 
+#include "../../utils/sharedmem.cuh"
+#include "../vector_manipulation/ve_mod_mult.cuh"
+
 const uint32_t MAX_NUM_THREADS = 1024;
-const uint32_t MAX_THREADS_BATCH = 256;
+const uint32_t MAX_THREADS_BATCH = 512;    //TODO: allows 100% occupancy for scalar NTT for sm_86..sm_89
+const uint32_t MAX_SHARED_MEM_ELEMENT_SIZE = 32; //TODO: occupancy calculator, hardcoded for sm_86..sm_89
+const uint32_t MAX_SHARED_MEM = MAX_SHARED_MEM_ELEMENT_SIZE * 1024; 
 
 /**
  * Computes the twiddle factors.  
@@ -37,40 +42,6 @@ const uint32_t MAX_THREADS_BATCH = 256;
   return d_twiddles;
 }
 
-/**
- * Returns the bit reversed order of a number. 
- * for example: on inputs num = 6 (110 in binary) and logn = 3
- * the function should return 3 (011 in binary.)
- * @param num some number with bit representation of size logn.
- * @param logn length of bit representation of `num`.
- * @return bit reveresed order or `num`.
- */
-__device__ __host__ uint32_t reverseBits(uint32_t num, uint32_t logn) {
-  unsigned int reverse_num = 0;
-  for (uint32_t i = 0; i < logn; i++) {
-    if ((num & (1 << i))) reverse_num |= 1 << ((logn - 1) - i);
-  }
-  return reverse_num;
-}
-
-/**
- * Returns the bit reversal ordering of the input array.
- * for example: on input ([a[0],a[1],a[2],a[3]], 4, 2) it returns
- * [a[0],a[3],a[2],a[1]] (elements in indices 3,1 swhich places).
- * @param arr array of some object of type T of size which is a power of 2. 
- * @param n length of `arr`.
- * @param logn log(n).
- * @return A new array which is the bit reversed version of input array. 
- */
-template < typename T > T * template_reverse_order(T * arr, uint32_t n, uint32_t logn) {
-  T * arrReversed = new T[n];
-  for (uint32_t i = 0; i < n; i++) {
-    uint32_t reversed = reverseBits(i, logn);
-    arrReversed[i] = arr[reversed];
-  }
-  return arrReversed;
-}
-
 template < typename T > __global__ void reverse_order_kernel(T* arr, T* arr_reversed, uint32_t n, uint32_t logn, uint32_t batch_size) {
   int threadId = (blockIdx.x * blockDim.x) + threadIdx.x;
   if (threadId < n * batch_size) {
@@ -112,156 +83,146 @@ template < typename T > void reverse_order(T* arr, uint32_t n, uint32_t logn, cu
   reverse_order_batch(arr, n, logn, 1, stream);
 }
 
-/**
- * Cooley-Tukey butterfly kernel. 
- * @param arr array of objects of type E (elements). 
- * @param twiddles array of twiddle factors of type S (scalars). 
- * @param n size of arr. 
- * @param n_twiddles size of omegas.
- * @param m "pair distance" - indicate distance of butterflies inputs.
- * @param i Cooley-Tukey FFT stage number.
- * @param max_thread_num maximal number of threads in stage. 
- */
-template < typename E, typename S > __global__ void template_butterfly_kernel(E * arr, S * twiddles, uint32_t n, uint32_t n_twiddles, uint32_t m, uint32_t i, uint32_t max_thread_num) {
-  int j = (blockIdx.x * blockDim.x) + threadIdx.x;
-  if (j < max_thread_num) {
-    uint32_t g = j * (n / m);
-    uint32_t k = i + j + (m >> 1);
-    E u = arr[i + j];
-    E v = twiddles[g * n_twiddles / n] * arr[k];
-    arr[i + j] = u + v;
-    arr[k] = u - v;
-  }
-}
+
+enum Decimation {
+  NONE = 0,
+  DIF = 1,
+  DIT = 2,
+};
 
 /**
- * Multiply the elements of an input array by a scalar in-place.
- * @param arr input array.
- * @param n size of arr.
- * @param n_inv scalar of type S (scalar).
- */
-template < typename E, typename S > __global__ void template_normalize_kernel(E * arr, uint32_t n, S scalar) {
-  int tid = (blockIdx.x * blockDim.x) + threadIdx.x;
-  if (tid < n) {
-    arr[tid] = scalar * arr[tid];
-  }
-}
-
-/**
- * Cooley-Tukey NTT.
- * NOTE! this function assumes that d_arr and d_twiddles are located in the device memory.
- * @param d_arr input array of type E (elements) allocated on the device memory.
- * @param n length of d_arr.
- * @param logn log(n).
- * @param d_twiddles twiddle factors of type S (scalars) array allocated on the device memory (must be a power of 2).
- * @param n_twiddles length of d_twiddles.
- */
-template < typename E, typename S > void template_ntt_on_device_memory(E * d_arr, uint32_t n, uint32_t logn, S * d_twiddles, uint32_t n_twiddles, cudaStream_t stream) {
-  uint32_t m = 2;
-  // TODO: optimize with separate streams for each iteration
-  for (uint32_t s = 0; s < logn; s++) {
-    for (uint32_t i = 0; i < n; i += m) {
-        uint32_t shifted_m = m >> 1;
-        uint32_t number_of_threads = MAX_NUM_THREADS ^ ((shifted_m ^ MAX_NUM_THREADS) & -(shifted_m < MAX_NUM_THREADS));
-        uint32_t number_of_blocks = shifted_m / MAX_NUM_THREADS + 1;
-        template_butterfly_kernel < E, S > <<< number_of_threads, number_of_blocks, 0, stream >>> (d_arr, d_twiddles, n, n_twiddles, m, i, m >> 1);
-    }
-    m <<= 1;
-  }
-}
-
-/**
- * Cooley-Tukey NTT. 
+ * Cooley-Tuckey NTT.
  * NOTE! this function assumes that d_twiddles are located in the device memory.
- * @param arr input array of type E (elements). 
+ * @param arr input array of type E (elements).
  * @param n length of d_arr.
- * @param d_twiddles twiddle factors of type S (scalars) array allocated on the device memory (must be a power of 2). 
- * @param n_twiddles length of d_twiddles. 
- * @param inverse indicate if the result array should be normalized by n^(-1). 
+ * @param twiddles twiddle factors of type S (scalars) array allocated on the device memory (must be a power of 2).
+ * @param n_twiddles length of twiddles.
+ * @param max_task max count of parallel tasks.
+ * @param s log2(n) loop index.
  */
-template < typename E, typename S > E * ntt_template(E * arr, uint32_t n, S * d_twiddles, uint32_t n_twiddles, bool inverse, cudaStream_t stream) {
-  uint32_t logn = uint32_t(log(n) / log(2));
-  size_t size_E = n * sizeof(E);
-  E * arrReversed = template_reverse_order < E > (arr, n, logn);
-  E * d_arrReversed;
-  cudaMallocAsync( & d_arrReversed, size_E, stream);
-  cudaMemcpyAsync(d_arrReversed, arrReversed, size_E, cudaMemcpyHostToDevice, stream);
-  template_ntt_on_device_memory < E, S > (d_arrReversed, n, logn, d_twiddles, n_twiddles, stream);
-  if (inverse) {
-    int NUM_THREADS = MAX_NUM_THREADS;
-    int NUM_BLOCKS = (n + NUM_THREADS - 1) / NUM_THREADS;
-    template_normalize_kernel < E, S > <<< NUM_THREADS, NUM_BLOCKS, 0, stream >>> (d_arrReversed, n, S::inv_log_size(logn));
-  }
-  cudaMemcpyAsync(arrReversed, d_arrReversed, size_E, cudaMemcpyDeviceToHost, stream);
-  cudaFreeAsync(d_arrReversed, stream);
-  cudaStreamSynchronize(stream);
-  return arrReversed;
-}
+template <typename E, typename S>
+__global__ void ntt_template_kernel_shared_rev(E *__restrict__ arr_g, uint32_t n, const S *__restrict__ r_twiddles, uint32_t n_twiddles, uint32_t max_task, uint32_t ss, uint32_t logn)
+{
+  SharedMemory<E> smem;
+  E *arr = smem.getPointer();
 
-/**
- * Cooley-Tukey (scalar) NTT. 
- * @param arr input array of type E (element). 
- * @param n length of d_arr.
- * @param inverse indicate if the result array should be normalized by n^(-1). 
- */
- template<typename E,typename S> uint32_t ntt_end2end_template(E * arr, uint32_t n, bool inverse, cudaStream_t stream) {
-  uint32_t logn = uint32_t(log(n) / log(2));
-  uint32_t n_twiddles = n; 
-  S * twiddles = new S[n_twiddles];
-  S * d_twiddles;
-  if (inverse){
-    d_twiddles = fill_twiddle_factors_array(n_twiddles, S::omega_inv(logn), stream);
-  } else{
-    d_twiddles = fill_twiddle_factors_array(n_twiddles, S::omega(logn), stream);
-  }
-  E * result = ntt_template < E, S > (arr, n, d_twiddles, n_twiddles, inverse, stream);
-  for(int i = 0; i < n; i++){
-    arr[i] = result[i]; 
-  }
-  cudaFreeAsync(d_twiddles, stream);
-  cudaStreamSynchronize(stream);
-  return 0; // TODO add
-}
+  uint32_t task = blockIdx.x;
+  uint32_t loop_limit = blockDim.x;
+  uint32_t chunks = n / (loop_limit * 2);
+  uint32_t offset = (task / chunks) * n;
+  if (task < max_task)
+  {
+    // flattened loop allows parallel processing
+    uint32_t l = threadIdx.x;
 
+    if (l < loop_limit)
+    {
+#pragma unroll
+      for (; ss < logn; ss++)
+      {
+        int s = logn - ss - 1;
+        bool is_beginning = ss == 0;
+        bool is_end = ss == (logn - 1);
 
-/**
- * Returens the bit reversal ordering of the input array according to the batches *in place*. 
- * The assumption is that arr is divided into N tasks of size n. 
- * Tasks indicates the index of the task (out of N). 
- * @param arr input array of type T.   
- * @param n length of arr.
- * @param logn log(n).
- * @param task log(n).
- */
-template < typename T > __device__ __host__ void reverseOrder_batch(T * arr, uint32_t n, uint32_t logn, uint32_t task) {
-  for (uint32_t i = 0; i < n; i++) {
-    uint32_t reversed = reverseBits(i, logn);
-    if (reversed > i) {
-      T tmp = arr[task * n + i];
-      arr[task * n + i] = arr[task * n + reversed];
-      arr[task * n + reversed] = tmp;
+        uint32_t ntw_i = task % chunks;
+
+        uint32_t n_twiddles_div = n_twiddles >> (s + 1);
+
+        uint32_t shift_s = 1 << s;
+        uint32_t shift2_s = 1 << (s + 1);
+
+        l = ntw_i * loop_limit + l; // to l from chunks to full
+
+        uint32_t j = l & (shift_s - 1);               // Equivalent to: l % (1 << s)
+        uint32_t i = ((l >> s) * shift2_s) & (n - 1); // (..) % n (assuming n is power of 2)
+        uint32_t oij = i + j;
+        uint32_t k = oij + shift_s;
+
+        S tw = r_twiddles[j * n_twiddles_div];
+
+        E u = is_beginning ? arr_g[offset + oij] : arr[oij];
+        E v = is_beginning ? arr_g[offset + k] : arr[k];
+        if (is_end)
+        {
+          arr_g[offset + oij] = u + v;
+          arr_g[offset + k] = tw * (u - v);
+        }
+        else
+        {
+          arr[oij] = u + v;
+          arr[k] = tw *(u - v);
+        }
+
+        __syncthreads();
+      }
     }
   }
 }
 
-
 /**
- * Cooley-Tukey butterfly kernel. 
- * @param arr array of objects of type E (elements). 
- * @param twiddles array of twiddle factors of type S (scalars). 
- * @param n size of arr. 
- * @param n_twiddles size of omegas.
- * @param m "pair distance" - indicate distance of butterflies inputs.
- * @param i Cooley-TUckey FFT stage number.
- * @param offset offset corr. to the specific taks (in batch).  
+ * Cooley-Tuckey NTT.
+ * NOTE! this function assumes that d_twiddles are located in the device memory.
+ * @param arr input array of type E (elements).
+ * @param n length of d_arr.
+ * @param twiddles twiddle factors of type S (scalars) array allocated on the device memory (must be a power of 2).
+ * @param n_twiddles length of twiddles.
+ * @param max_task max count of parallel tasks.
+ * @param s log2(n) loop index.
  */
-template < typename E, typename S > __device__ __host__ void butterfly(E * arrReversed, S * omegas, uint32_t n, uint32_t n_omegas, uint32_t m, uint32_t i, uint32_t j, uint32_t offset) {
-  uint32_t g = j * (n / m);
-  uint32_t k = i + j + (m >> 1);
-  E u = arrReversed[offset + i + j];
-  E v = omegas[g * n_omegas / n] * arrReversed[offset + k];
-  arrReversed[offset + i + j] = u + v;
-  arrReversed[offset + k] = u - v;
+template <typename E, typename S>
+__global__ void ntt_template_kernel_shared(E *__restrict__ arr_g, uint32_t n, const S *__restrict__ r_twiddles, uint32_t n_twiddles, uint32_t max_task, uint32_t s, uint32_t logn)
+{
+  SharedMemory<E> smem;
+  E *arr = smem.getPointer();
+
+  uint32_t task = blockIdx.x;
+  uint32_t loop_limit = blockDim.x;
+  uint32_t chunks = n / (loop_limit * 2);
+  uint32_t offset = (task / chunks) * n;
+  if (task < max_task)
+  {
+    // flattened loop allows parallel processing
+    uint32_t l = threadIdx.x;
+
+    if (l < loop_limit)
+    {
+#pragma unroll
+      for (; s < logn; s++) // TODO: this loop also can be unrolled
+      {
+        uint32_t ntw_i = task % chunks;
+
+        uint32_t n_twiddles_div = n_twiddles >> (s + 1);
+
+        uint32_t shift_s = 1 << s;
+        uint32_t shift2_s = 1 << (s + 1);
+
+        l = ntw_i * loop_limit + l; // to l from chunks to full
+
+        uint32_t j = l & (shift_s - 1);               // Equivalent to: l % (1 << s)
+        uint32_t i = ((l >> s) * shift2_s) & (n - 1); // (..) % n (assuming n is power of 2)
+        uint32_t oij = i + j;
+        uint32_t k = oij + shift_s;
+        S tw = r_twiddles[j * n_twiddles_div];
+
+
+        E u = s == 0 ? arr_g[offset + oij] : arr[oij];
+        E v = s == 0 ? arr_g[offset + k] : arr[k];
+        v = tw * v;
+        if (s == (logn - 1))
+        {
+          arr_g[offset + oij] = u + v;
+          arr_g[offset + k] = u - v;
+        }
+        else
+        {
+          arr[oij] = u + v;
+          arr[k] = u - v;
+        }
+
+        __syncthreads();
+      }
+    }
+  }
 }
 
 /**
@@ -296,39 +257,80 @@ __global__ void ntt_template_kernel(E *arr, uint32_t n, S *twiddles, uint32_t n_
 
       l = ntw_i * blockDim.x + l; //to l from chunks to full
 
-      uint32_t j = l & (shift_s - 1); // Equivalent to: l % (1 << s)
-      uint32_t i = ((l / shift_s) * shift2_s) % n;
+      uint32_t j = l & (shift_s - 1);               // Equivalent to: l % (1 << s)
+      uint32_t i = ((l >> s) * shift2_s) & (n - 1); // (..) % n (assuming n is power of 2)
       uint32_t k = i + j + shift_s;
 
+      S tw = twiddles[j * n_twiddles_div];
+
       uint32_t offset = (task / chunks) * n;
       E u = arr[offset + i + j];
-      E v = rev ? arr[offset + k] : twiddles[j * n_twiddles_div] * arr[offset + k];
+      E v = arr[offset + k];
+      if (!rev) v = tw * v;
       arr[offset + i + j] = u + v;
-      arr[offset + k] = u - v;
-      if (rev)
-        arr[offset + k] = twiddles[j * n_twiddles_div] * arr[offset + k];
+      v = u - v;
+      arr[offset + k] = rev ? tw * v : v;
     }
   }
 }
 
-
 /**
- * Cooley-Tukey NTT.
- * NOTE! this function assumes that d_twiddles are located in the device memory.
- * @param arr input array of type E (elements).
- * @param n length of arr.
- * @param logn log2(n).
- * @param max_task max count of parallel tasks.
+ * NTT/INTT inplace batch
+ * Note: this function does not preform any bit-reverse permutations on its inputs or outputs.
+ * @param d_inout Array for inplace processing
+ * @param d_twiddles 
+ * @param n Length of `d_twiddles` array
+ * @param batch_size The size of the batch; the length of `d_inout` is `n` * `batch_size`.
+ * @param inverse true for iNTT
+ * @param is_coset true for multiplication by coset
+ * @param coset should be array of lenght n - or in case of lesser than n, right-padded with zeroes
+ * @param stream CUDA stream   
+ * @param is_sync_needed do perform sync of the supplied CUDA stream at the end of processing
  */
-template <typename E, typename S>
-__global__ void ntt_template_kernel_rev_ord(E *arr, uint32_t n, uint32_t logn, uint32_t max_task)
+template <typename E, typename S> void ntt_inplace_batch_template(
+  E * d_inout, S * d_twiddles, unsigned n, unsigned batch_size, bool inverse, 
+  bool is_coset, S * coset, cudaStream_t stream, bool is_sync_needed) 
 {
-  int task = (blockIdx.x * blockDim.x) + threadIdx.x;
+  const int logn = int(log(n) / log(2));
+  bool is_shared_mem_enabled = sizeof(E) <= MAX_SHARED_MEM_ELEMENT_SIZE;
+  const int log2_shmem_elems = is_shared_mem_enabled ? int(log(int(MAX_SHARED_MEM / sizeof(E))) / log(2)) : logn;
+  int num_threads = min(min(n / 2, MAX_THREADS_BATCH), 1 << (log2_shmem_elems - 1));
+  const int chunks = max(int((n / 2) / num_threads), 1);
+  const int total_tasks = batch_size * chunks;
+  int num_blocks = total_tasks;
+  const int shared_mem = 2 * num_threads * sizeof(E); // TODO: calculator, as shared mem size may be more efficient less then max to allow more concurrent blocks on SM
+  const int logn_shmem = is_shared_mem_enabled ? int(log(2 * num_threads) / log(2)) : 0; //TODO: shared memory support only for types <= 32 bytes
 
-  if (task < max_task)
-  {
-    reverseOrder_batch<E>(arr, n, logn, task);
-  }
+  // if (inverse) 
+  // {
+  //   if (is_shared_mem_enabled) ntt_template_kernel_shared<<<num_blocks, num_threads, shared_mem, stream>>>(d_inout, 1 << logn_shmem, d_twiddles, n, total_tasks, 0, logn_shmem);
+
+  //   for (int s = logn_shmem; s < logn; s++) // TODO: this loop also can be unrolled
+  //   { 
+  //     ntt_template_kernel <E, S> <<<num_blocks, num_threads, 0, stream>>>(d_inout, n, d_twiddles, n, total_tasks, s, false);
+  //   }
+
+  //   if (is_coset) batch_vector_mult(coset, d_inout, n, batch_size, stream);
+
+  //   num_threads = min(n / 2, MAX_NUM_THREADS);
+  //   num_blocks = (n * batch_size + num_threads - 1) / num_threads;
+  //   // template_normalize_kernel <E, S> <<<num_blocks, num_threads, 0, stream>>> (d_inout, n * batch_size, S::inv_log_size(logn)); 
+  // }
+  // else 
+  // {
+    if (is_coset) batch_vector_mult(coset, d_inout, n, batch_size, stream);
+
+    for (int s = logn - 1; s >= logn_shmem; s--) // TODO: this loop also can be unrolled
+    {
+      ntt_template_kernel<<<num_blocks, num_threads, 0, stream>>>(d_inout, n, d_twiddles, n, total_tasks, s, true);
+    }
+    
+    if (is_shared_mem_enabled) ntt_template_kernel_shared_rev<<<num_blocks, num_threads, shared_mem, stream>>>(d_inout, 1 << logn_shmem, d_twiddles, n, total_tasks, 0, logn_shmem);
+  // }
+  
+  if (!is_sync_needed) return;
+
+  cudaStreamSynchronize(stream);
 }
 
 /**
@@ -343,7 +345,7 @@ __global__ void ntt_template_kernel_rev_ord(E *arr, uint32_t n, uint32_t logn, u
  template <typename E, typename S> uint32_t ntt_end2end_batch_template(E * arr, uint32_t arr_size, uint32_t n, bool inverse, cudaStream_t stream) {
   int batches = int(arr_size / n);
   uint32_t logn = uint32_t(log(n) / log(2));
-  uint32_t n_twiddles = n; // n_twiddles is set to 4096 as BLS12_381::scalar_t::omega() is of that order. 
+  uint32_t n_twiddles = n/2; // n_twiddles is set to 4096 as BLS12_381::scalar_t::omega() is of that order. 
   size_t size_E = arr_size * sizeof(E);
   S * d_twiddles;
   if (inverse){
@@ -356,25 +358,10 @@ __global__ void ntt_template_kernel_rev_ord(E *arr, uint32_t n, uint32_t logn, u
   cudaMemcpyAsync(d_arr, arr, size_E, cudaMemcpyHostToDevice, stream);
   int NUM_THREADS = MAX_THREADS_BATCH;
   int NUM_BLOCKS = (batches + NUM_THREADS - 1) / NUM_THREADS;
-  ntt_template_kernel_rev_ord<E, S><<<NUM_BLOCKS, NUM_THREADS, 0, stream>>>(d_arr, n, logn, batches);
-
-  NUM_THREADS = min(n / 2, MAX_THREADS_BATCH);
-  int chunks = max(int((n / 2) / NUM_THREADS), 1);
-  int total_tasks = batches * chunks;
-  NUM_BLOCKS = total_tasks;
-
-  //TODO: this loop also can be unrolled
-  for (uint32_t s = 0; s < logn; s++)
-  {
-    ntt_template_kernel<E, S><<<NUM_BLOCKS, NUM_THREADS, 0, stream>>>(d_arr, n, d_twiddles, n_twiddles, total_tasks, s, false);
-    cudaStreamSynchronize(stream);
-  }
-  if (inverse == true)
-  {
-    NUM_THREADS = MAX_NUM_THREADS;
-    NUM_BLOCKS = (arr_size + NUM_THREADS - 1) / NUM_THREADS;
-    template_normalize_kernel < E, S > <<< NUM_THREADS, NUM_BLOCKS, 0, stream>>> (d_arr, arr_size, S::inv_log_size(logn));
-  }
+   
+  S* _null = nullptr;
+  ntt_inplace_batch_template(d_arr, d_twiddles, n, batches, inverse, false, _null, stream, false);
+  reverse_order_batch(d_arr, n, logn, batches, stream);
   cudaMemcpyAsync(arr, d_arr, size_E, cudaMemcpyDeviceToHost, stream);
   cudaFreeAsync(d_arr, stream);
   cudaFreeAsync(d_twiddles, stream);
@@ -382,4 +369,14 @@ __global__ void ntt_template_kernel_rev_ord(E *arr, uint32_t n, uint32_t logn, u
   return 0; 
 }
 
+/**
+ * Cooley-Tukey (scalar) NTT. 
+ * @param arr input array of type E (element). 
+ * @param n length of d_arr.
+ * @param inverse indicate if the result array should be normalized by n^(-1). 
+ */
+ template<typename E,typename S> uint32_t ntt_end2end_template(E * arr, uint32_t n, bool inverse, cudaStream_t stream) {
+  return ntt_end2end_batch_template <E, S> (arr, n, n, inverse, stream);
+}
+
 #endif

icicle/appUtils/poseidon/constants.cuh

@@ -1,27 +1,27 @@
 #pragma once
 
-#include <map>
-#include <stdexcept>
-#include <cassert>
-
+#include "constants/constants_11.h"
 #include "constants/constants_2.h"
 #include "constants/constants_4.h"
 #include "constants/constants_8.h"
-#include "constants/constants_11.h"
+#include <cassert>
+#include <map>
+#include <stdexcept>
 
-uint32_t partial_rounds_number_from_arity(const uint32_t arity) {
-    switch (arity) {
-        case 2:
-            return 55;
-        case 4:
-            return 56;
-        case 8:
-            return 57;
-        case 11:
-            return 57;
-        default:
-            throw std::invalid_argument( "unsupported arity" );
-    }
+uint32_t partial_rounds_number_from_arity(const uint32_t arity)
+{
+  switch (arity) {
+  case 2:
+    return 55;
+  case 4:
+    return 56;
+  case 8:
+    return 57;
+  case 11:
+    return 57;
+  default:
+    throw std::invalid_argument("unsupported arity");
+  }
 };
 
 // TO-DO: change to mapping
@@ -29,23 +29,24 @@ const uint32_t FULL_ROUNDS_DEFAULT = 4;
 
 // TO-DO: for now, the constants are only generated in bls12_381
 template <typename S>
-S * load_constants(const uint32_t arity) {
-    unsigned char * constants;
-    switch (arity) {
-        case 2:
-            constants = constants_2;
-            break;
-        case 4:
-            constants = constants_4;
-            break;
-        case 8:
-            constants = constants_8;
-            break;
-        case 11:
-            constants = constants_11;
-            break;
-        default:
-            throw std::invalid_argument( "unsupported arity" );
-    }
-    return reinterpret_cast< S * >(constants);
+S* load_constants(const uint32_t arity)
+{
+  unsigned char* constants;
+  switch (arity) {
+  case 2:
+    constants = constants_2;
+    break;
+  case 4:
+    constants = constants_4;
+    break;
+  case 8:
+    constants = constants_8;
+    break;
+  case 11:
+    constants = constants_11;
+    break;
+  default:
+    throw std::invalid_argument("unsupported arity");
+  }
+  return reinterpret_cast<S*>(constants);
 }

icicle/appUtils/poseidon/poseidon.cu

@@ -1,273 +1,266 @@
 #include "poseidon.cuh"
 
 template <typename S>
-__global__ void prepare_poseidon_states(S * states, size_t number_of_states, S domain_tag, const PoseidonConfiguration<S> config) {
-    int idx = (blockIdx.x * blockDim.x) + threadIdx.x;
-    int state_number = idx / config.t;
-    if (state_number >= number_of_states) {
-        return;
-    }
-    int element_number = idx % config.t;
+__global__ void
+prepare_poseidon_states(S* states, size_t number_of_states, S domain_tag, const PoseidonConfiguration<S> config)
+{
+  int idx = (blockIdx.x * blockDim.x) + threadIdx.x;
+  int state_number = idx / config.t;
+  if (state_number >= number_of_states) { return; }
+  int element_number = idx % config.t;
 
-    S prepared_element;
+  S prepared_element;
 
-    // Domain separation
-    if (element_number == 0) {
-        prepared_element = domain_tag;
-    } else {
-        prepared_element = states[state_number * config.t + element_number - 1];
-    }
+  // Domain separation
+  if (element_number == 0) {
+    prepared_element = domain_tag;
+  } else {
+    prepared_element = states[state_number * config.t + element_number - 1];
+  }
 
-    // Add pre-round constant
-    prepared_element = prepared_element + config.round_constants[element_number];
+  // Add pre-round constant
+  prepared_element = prepared_element + config.round_constants[element_number];
 
-    // Store element in state
-    states[idx] = prepared_element;
+  // Store element in state
+  states[idx] = prepared_element;
 }
 
 template <typename S>
-__device__ __forceinline__ S sbox_alpha_five(S element) {
-    S result = S::sqr(element);
-    result = S::sqr(result);
-    return result * element;
+__device__ __forceinline__ S sbox_alpha_five(S element)
+{
+  S result = S::sqr(element);
+  result = S::sqr(result);
+  return result * element;
 }
 
 template <typename S>
-__device__ S vecs_mul_matrix(S element, S * matrix, int element_number, int vec_number, int size, S * shared_states) {
-    shared_states[threadIdx.x] = element;
-    __syncthreads();
+__device__ S vecs_mul_matrix(S element, S* matrix, int element_number, int vec_number, int size, S* shared_states)
+{
+  shared_states[threadIdx.x] = element;
+  __syncthreads();
 
-    element = S::zero();
-    for (int i = 0; i < size; i++) {
-        element = element + (shared_states[vec_number * size + i] * matrix[i * size + element_number]);
-    }
-    __syncthreads();
-    return element;
+  element = S::zero();
+  for (int i = 0; i < size; i++) {
+    element = element + (shared_states[vec_number * size + i] * matrix[i * size + element_number]);
+  }
+  __syncthreads();
+  return element;
 }
 
 template <typename S>
-__device__ S full_round(S element,
-                        size_t rc_offset,
-                        int local_state_number,
-                        int element_number,
-                        bool multiply_by_mds,
-                        bool add_round_constant,
-                        S * shared_states,
-                        const PoseidonConfiguration<S> config) {
-    element = sbox_alpha_five(element);
-    if (add_round_constant) {
-        element = element + config.round_constants[rc_offset + element_number];
-    }
+__device__ S full_round(
+  S element,
+  size_t rc_offset,
+  int local_state_number,
+  int element_number,
+  bool multiply_by_mds,
+  bool add_round_constant,
+  S* shared_states,
+  const PoseidonConfiguration<S> config)
+{
+  element = sbox_alpha_five(element);
+  if (add_round_constant) { element = element + config.round_constants[rc_offset + element_number]; }
 
-    // Multiply all the states by mds matrix
-    S * matrix = multiply_by_mds ? config.mds_matrix : config.non_sparse_matrix;
-    return vecs_mul_matrix(element, matrix, element_number, local_state_number, config.t, shared_states);
+  // Multiply all the states by mds matrix
+  S* matrix = multiply_by_mds ? config.mds_matrix : config.non_sparse_matrix;
+  return vecs_mul_matrix(element, matrix, element_number, local_state_number, config.t, shared_states);
 }
 
 // Execute full rounds
 template <typename S>
-__global__ void full_rounds(S * states, size_t number_of_states, size_t rc_offset, bool first_half, const PoseidonConfiguration<S> config) {
-    extern __shared__ S shared_states[];
+__global__ void full_rounds(
+  S* states, size_t number_of_states, size_t rc_offset, bool first_half, const PoseidonConfiguration<S> config)
+{
+  extern __shared__ S shared_states[];
 
-    int idx = (blockIdx.x * blockDim.x) + threadIdx.x;
-    int state_number = idx / config.t;
-    if (state_number >= number_of_states) {
-        return;
-    }
-    int local_state_number = threadIdx.x / config.t;
-    int element_number = idx % config.t;
+  int idx = (blockIdx.x * blockDim.x) + threadIdx.x;
+  int state_number = idx / config.t;
+  if (state_number >= number_of_states) { return; }
+  int local_state_number = threadIdx.x / config.t;
+  int element_number = idx % config.t;
 
-    for (int i = 0; i < config.full_rounds_half - 1; i++) {
-        states[idx] = full_round(states[idx],
-                                 rc_offset,
-                                 local_state_number,
-                                 element_number,
-                                 true,
-                                 true,
-                                 shared_states,
-                                 config);
-        rc_offset += config.t;
-    }
+  for (int i = 0; i < config.full_rounds_half - 1; i++) {
+    states[idx] =
+      full_round(states[idx], rc_offset, local_state_number, element_number, true, true, shared_states, config);
+    rc_offset += config.t;
+  }
 
-    states[idx] = full_round(states[idx],
-                             rc_offset,
-                             local_state_number,
-                             element_number,
-                             !first_half,
-                             first_half,
-                             shared_states,
-                             config);
+  states[idx] = full_round(
+    states[idx], rc_offset, local_state_number, element_number, !first_half, first_half, shared_states, config);
 }
 
 template <typename S>
-__device__ S partial_round(S * state,
-                                  size_t rc_offset,
-                                  int round_number,
-                                  const PoseidonConfiguration<S> config) {
-    S element = state[0];
-    element = sbox_alpha_five(element);
-    element = element + config.round_constants[rc_offset];
+__device__ S partial_round(S* state, size_t rc_offset, int round_number, const PoseidonConfiguration<S> config)
+{
+  S element = state[0];
+  element = sbox_alpha_five(element);
+  element = element + config.round_constants[rc_offset];
 
-    S * sparse_matrix = &config.sparse_matrices[(config.t * 2 - 1) * round_number];
+  S* sparse_matrix = &config.sparse_matrices[(config.t * 2 - 1) * round_number];
 
-    state[0] = element * sparse_matrix[0];
-    for (int i = 1; i < config.t; i++) {
-        state[0] = state[0] + (state[i] * sparse_matrix[i]);
-    }
+  state[0] = element * sparse_matrix[0];
+  for (int i = 1; i < config.t; i++) {
+    state[0] = state[0] + (state[i] * sparse_matrix[i]);
+  }
 
-    for (int i = 1; i < config.t; i++) {
-        state[i] = state[i] + (element * sparse_matrix[config.t + i - 1]);
-    }
+  for (int i = 1; i < config.t; i++) {
+    state[i] = state[i] + (element * sparse_matrix[config.t + i - 1]);
+  }
 }
 
 // Execute partial rounds
 template <typename S>
-__global__ void partial_rounds(S * states, size_t number_of_states, size_t rc_offset, const PoseidonConfiguration<S> config) {
-    int idx = (blockIdx.x * blockDim.x) + threadIdx.x;
-    if (idx >= number_of_states) {
-        return;
-    }
+__global__ void
+partial_rounds(S* states, size_t number_of_states, size_t rc_offset, const PoseidonConfiguration<S> config)
+{
+  int idx = (blockIdx.x * blockDim.x) + threadIdx.x;
+  if (idx >= number_of_states) { return; }
 
-    S * state = &states[idx * config.t];
+  S* state = &states[idx * config.t];
 
-    for (int i = 0; i < config.partial_rounds; i++) {
-        partial_round(state, rc_offset, i, config);
-        rc_offset++;
-    }
+  for (int i = 0; i < config.partial_rounds; i++) {
+    partial_round(state, rc_offset, i, config);
+    rc_offset++;
+  }
 }
 
 // These function is just doing copy from the states to the output
 template <typename S>
-__global__ void get_hash_results(S * states, size_t number_of_states, S * out, int t) {
-    int idx = (blockIdx.x * blockDim.x) + threadIdx.x;
-    if (idx >= number_of_states) {
-        return;
-    }
+__global__ void get_hash_results(S* states, size_t number_of_states, S* out, int t)
+{
+  int idx = (blockIdx.x * blockDim.x) + threadIdx.x;
+  if (idx >= number_of_states) { return; }
 
-    out[idx] = states[idx * t + 1];
+  out[idx] = states[idx * t + 1];
 }
 
 template <typename S>
-__host__ void Poseidon<S>::hash_blocks(const S * inp, size_t blocks, S * out, HashType hash_type, cudaStream_t stream) {
-    S * states;
+__host__ void Poseidon<S>::hash_blocks(const S* inp, size_t blocks, S* out, HashType hash_type, cudaStream_t stream)
+{
+  S* states;
 
-    // allocate memory for {blocks} states of {t} scalars each
-    if (cudaMallocAsync(&states, blocks * this->t * sizeof(S), stream) != cudaSuccess) {
-        throw std::runtime_error("Failed memory allocation on the device");
-    }
+  // allocate memory for {blocks} states of {t} scalars each
+  if (cudaMallocAsync(&states, blocks * this->t * sizeof(S), stream) != cudaSuccess) {
+    throw std::runtime_error("Failed memory allocation on the device");
+  }
 
-    // This is where the input matrix of size Arity x NumberOfBlocks is
-    // padded and coppied to device in a T x NumberOfBlocks matrix
-    cudaMemcpy2DAsync(states, this->t * sizeof(S),  // Device pointer and device pitch
-                 inp, (this->t - 1) * sizeof(S),    // Host pointer and pitch
-                 (this->t - 1) * sizeof(S), blocks, // Size of the source matrix (Arity x NumberOfBlocks)
-                 cudaMemcpyHostToDevice, stream);
+  // This is where the input matrix of size Arity x NumberOfBlocks is
+  // padded and coppied to device in a T x NumberOfBlocks matrix
+  cudaMemcpy2DAsync(
+    states, this->t * sizeof(S),       // Device pointer and device pitch
+    inp, (this->t - 1) * sizeof(S),    // Host pointer and pitch
+    (this->t - 1) * sizeof(S), blocks, // Size of the source matrix (Arity x NumberOfBlocks)
+    cudaMemcpyHostToDevice, stream);
 
-    size_t rc_offset = 0;
+  size_t rc_offset = 0;
 
-    // The logic behind this is that 1 thread only works on 1 element
-    // We have {t} elements in each state, and {blocks} states total
-    int number_of_threads = (256 / this->t) * this->t;
-    int hashes_per_block = number_of_threads / this->t;
-    int total_number_of_threads = blocks * this->t;
-    int number_of_blocks = total_number_of_threads / number_of_threads +
-        static_cast<bool>(total_number_of_threads % number_of_threads);
+  // The logic behind this is that 1 thread only works on 1 element
+  // We have {t} elements in each state, and {blocks} states total
+  int number_of_threads = (256 / this->t) * this->t;
+  int hashes_per_block = number_of_threads / this->t;
+  int total_number_of_threads = blocks * this->t;
+  int number_of_blocks =
+    total_number_of_threads / number_of_threads + static_cast<bool>(total_number_of_threads % number_of_threads);
 
-    // The partial rounds operates on the whole state, so we define
-    // the parallelism params for processing a single hash preimage per thread
-    int singlehash_block_size = 128;
-    int number_of_singlehash_blocks = blocks / singlehash_block_size + static_cast<bool>(blocks % singlehash_block_size);
+  // The partial rounds operates on the whole state, so we define
+  // the parallelism params for processing a single hash preimage per thread
+  int singlehash_block_size = 128;
+  int number_of_singlehash_blocks = blocks / singlehash_block_size + static_cast<bool>(blocks % singlehash_block_size);
 
-    // Pick the domain_tag accordinaly
-    S domain_tag;
-    switch (hash_type) {
-        case HashType::ConstInputLen:
-            domain_tag = this->const_input_no_pad_domain_tag;
-            break;
+  // Pick the domain_tag accordinaly
+  S domain_tag;
+  switch (hash_type) {
+  case HashType::ConstInputLen:
+    domain_tag = this->const_input_no_pad_domain_tag;
+    break;
 
-        case HashType::MerkleTree:
-            domain_tag = this->tree_domain_tag;
-    }
+  case HashType::MerkleTree:
+    domain_tag = this->tree_domain_tag;
+  }
 
-    #if !defined(__CUDA_ARCH__) && defined(DEBUG)
-    auto start_time = std::chrono::high_resolution_clock::now();
-    #endif
+#if !defined(__CUDA_ARCH__) && defined(DEBUG)
+  auto start_time = std::chrono::high_resolution_clock::now();
+#endif
 
-    // Domain separation and adding pre-round constants
-    prepare_poseidon_states <<< number_of_blocks, number_of_threads, 0, stream >>> (states, blocks, domain_tag, this->config);
-    rc_offset += this->t;
+  // Domain separation and adding pre-round constants
+  prepare_poseidon_states<<<number_of_blocks, number_of_threads, 0, stream>>>(states, blocks, domain_tag, this->config);
+  rc_offset += this->t;
 
-    #if !defined(__CUDA_ARCH__) && defined(DEBUG)
-    cudaStreamSynchronize(stream);
-    std::cout << "Domain separation: " << rc_offset << std::endl;
-    //print_buffer_from_cuda<S>(states, blocks * this->t);
+#if !defined(__CUDA_ARCH__) && defined(DEBUG)
+  cudaStreamSynchronize(stream);
+  std::cout << "Domain separation: " << rc_offset << std::endl;
+  // print_buffer_from_cuda<S>(states, blocks * this->t);
 
-    auto end_time = std::chrono::high_resolution_clock::now();
-    auto elapsed_time = std::chrono::duration_cast<std::chrono::milliseconds>(end_time - start_time);
-    std::cout << "Elapsed time: " << elapsed_time.count() << " ms" << std::endl;
-    start_time = std::chrono::high_resolution_clock::now();
-    #endif
+  auto end_time = std::chrono::high_resolution_clock::now();
+  auto elapsed_time = std::chrono::duration_cast<std::chrono::milliseconds>(end_time - start_time);
+  std::cout << "Elapsed time: " << elapsed_time.count() << " ms" << std::endl;
+  start_time = std::chrono::high_resolution_clock::now();
+#endif
 
-    // execute half full rounds
-    full_rounds <<< number_of_blocks, number_of_threads, sizeof(S) * hashes_per_block * this->t, stream >>> (states, blocks, rc_offset, true, this->config);
-    rc_offset += this->t * this->config.full_rounds_half;
+  // execute half full rounds
+  full_rounds<<<number_of_blocks, number_of_threads, sizeof(S) * hashes_per_block* this->t, stream>>>(
+    states, blocks, rc_offset, true, this->config);
+  rc_offset += this->t * this->config.full_rounds_half;
 
-    #if !defined(__CUDA_ARCH__) && defined(DEBUG)
-    cudaStreamSynchronize(stream);
-    std::cout << "Full rounds 1. RCOFFSET: " << rc_offset << std::endl;
-    // print_buffer_from_cuda<S>(states, blocks * this->t);
+#if !defined(__CUDA_ARCH__) && defined(DEBUG)
+  cudaStreamSynchronize(stream);
+  std::cout << "Full rounds 1. RCOFFSET: " << rc_offset << std::endl;
+  // print_buffer_from_cuda<S>(states, blocks * this->t);
 
-    end_time = std::chrono::high_resolution_clock::now();
-    elapsed_time = std::chrono::duration_cast<std::chrono::milliseconds>(end_time - start_time);
-    std::cout << "Elapsed time: " << elapsed_time.count() << " ms" << std::endl;
-    start_time = std::chrono::high_resolution_clock::now();
-    #endif
+  end_time = std::chrono::high_resolution_clock::now();
+  elapsed_time = std::chrono::duration_cast<std::chrono::milliseconds>(end_time - start_time);
+  std::cout << "Elapsed time: " << elapsed_time.count() << " ms" << std::endl;
+  start_time = std::chrono::high_resolution_clock::now();
+#endif
 
-    // execute partial rounds
-    partial_rounds <<< number_of_singlehash_blocks, singlehash_block_size, 0, stream >>> (states, blocks, rc_offset, this->config);
-    rc_offset += this->config.partial_rounds;
+  // execute partial rounds
+  partial_rounds<<<number_of_singlehash_blocks, singlehash_block_size, 0, stream>>>(
+    states, blocks, rc_offset, this->config);
+  rc_offset += this->config.partial_rounds;
 
-    #if !defined(__CUDA_ARCH__) && defined(DEBUG)
-    cudaStreamSynchronize(stream);
-    std::cout << "Partial rounds. RCOFFSET: " << rc_offset << std::endl;
-    //print_buffer_from_cuda<S>(states, blocks * this->t);
+#if !defined(__CUDA_ARCH__) && defined(DEBUG)
+  cudaStreamSynchronize(stream);
+  std::cout << "Partial rounds. RCOFFSET: " << rc_offset << std::endl;
+  // print_buffer_from_cuda<S>(states, blocks * this->t);
 
-    end_time = std::chrono::high_resolution_clock::now();
-    elapsed_time = std::chrono::duration_cast<std::chrono::milliseconds>(end_time - start_time);
-    std::cout << "Elapsed time: " << elapsed_time.count() << " ms" << std::endl;
-    start_time = std::chrono::high_resolution_clock::now();
-    #endif
+  end_time = std::chrono::high_resolution_clock::now();
+  elapsed_time = std::chrono::duration_cast<std::chrono::milliseconds>(end_time - start_time);
+  std::cout << "Elapsed time: " << elapsed_time.count() << " ms" << std::endl;
+  start_time = std::chrono::high_resolution_clock::now();
+#endif
 
-    // execute half full rounds
-    full_rounds <<< number_of_blocks, number_of_threads, sizeof(S) * hashes_per_block * this->t, stream >>> (states, blocks, rc_offset, false, this->config);
+  // execute half full rounds
+  full_rounds<<<number_of_blocks, number_of_threads, sizeof(S) * hashes_per_block* this->t, stream>>>(
+    states, blocks, rc_offset, false, this->config);
 
-    #if !defined(__CUDA_ARCH__) && defined(DEBUG)
-    cudaStreamSynchronize(stream);
-    std::cout << "Full rounds 2. RCOFFSET: " << rc_offset << std::endl;
-    //print_buffer_from_cuda<S>(states, blocks * this->t);
-    end_time = std::chrono::high_resolution_clock::now();
-    elapsed_time = std::chrono::duration_cast<std::chrono::milliseconds>(end_time - start_time);
-    std::cout << "Elapsed time: " << elapsed_time.count() << " ms" << std::endl;
-    start_time = std::chrono::high_resolution_clock::now();
-    #endif
+#if !defined(__CUDA_ARCH__) && defined(DEBUG)
+  cudaStreamSynchronize(stream);
+  std::cout << "Full rounds 2. RCOFFSET: " << rc_offset << std::endl;
+  // print_buffer_from_cuda<S>(states, blocks * this->t);
+  end_time = std::chrono::high_resolution_clock::now();
+  elapsed_time = std::chrono::duration_cast<std::chrono::milliseconds>(end_time - start_time);
+  std::cout << "Elapsed time: " << elapsed_time.count() << " ms" << std::endl;
+  start_time = std::chrono::high_resolution_clock::now();
+#endif
 
-    // get output
-    S * out_device;
-    cudaMalloc(&out_device, blocks * sizeof(S));
-    get_hash_results <<< number_of_singlehash_blocks, singlehash_block_size, 0, stream >>> (states, blocks, out_device, this->config.t);
+  // get output
+  S* out_device;
+  cudaMalloc(&out_device, blocks * sizeof(S));
+  get_hash_results<<<number_of_singlehash_blocks, singlehash_block_size, 0, stream>>>(
+    states, blocks, out_device, this->config.t);
 
-    #if !defined(__CUDA_ARCH__) && defined(DEBUG)
-    cudaStreamSynchronize(stream);
-    std::cout << "Get hash results" << std::endl;
-    end_time = std::chrono::high_resolution_clock::now();
-    elapsed_time = std::chrono::duration_cast<std::chrono::milliseconds>(end_time - start_time);
-    std::cout << "Elapsed time: " << elapsed_time.count() << " ms" << std::endl;
-    #endif
-    cudaMemcpyAsync(out, out_device, blocks * sizeof(S), cudaMemcpyDeviceToHost, stream);
-    cudaFreeAsync(out_device, stream);
-    cudaFreeAsync(states, stream);
+#if !defined(__CUDA_ARCH__) && defined(DEBUG)
+  cudaStreamSynchronize(stream);
+  std::cout << "Get hash results" << std::endl;
+  end_time = std::chrono::high_resolution_clock::now();
+  elapsed_time = std::chrono::duration_cast<std::chrono::milliseconds>(end_time - start_time);
+  std::cout << "Elapsed time: " << elapsed_time.count() << " ms" << std::endl;
+#endif
+  cudaMemcpyAsync(out, out_device, blocks * sizeof(S), cudaMemcpyDeviceToHost, stream);
+  cudaFreeAsync(out_device, stream);
+  cudaFreeAsync(states, stream);
 
-    #if !defined(__CUDA_ARCH__) && defined(DEBUG)
-    cudaDeviceReset();
-    #endif
+#if !defined(__CUDA_ARCH__) && defined(DEBUG)
+  cudaDeviceReset();
+#endif
 }

icicle/appUtils/poseidon/poseidon.cuh

@@ -2,19 +2,20 @@
 #include "constants.cuh"
 
 #if !defined(__CUDA_ARCH__) && defined(DEBUG)
-#include <iostream>
-#include <iomanip>
-#include <string>
-#include <sstream>
 #include <chrono>
+#include <iomanip>
+#include <iostream>
+#include <sstream>
+#include <string>
 
 template <typename S>
-__host__ void print_buffer_from_cuda(S * device_ptr, size_t size, size_t t) {
-  S * buffer = static_cast< S * >(malloc(size * sizeof(S)));
+__host__ void print_buffer_from_cuda(S* device_ptr, size_t size, size_t t)
+{
+  S* buffer = static_cast<S*>(malloc(size * sizeof(S)));
   cudaMemcpy(buffer, device_ptr, size * sizeof(S), cudaMemcpyDeviceToHost);
 
   std::cout << "Start print" << std::endl;
-  for(int i = 0; i < size / t; i++) {
+  for (int i = 0; i < size / t; i++) {
     std::cout << "State #" << i << std::endl;
     for (int j = 0; j < t; j++) {
       std::cout << buffer[i * t + j] << std::endl;
@@ -28,136 +29,129 @@ __host__ void print_buffer_from_cuda(S * device_ptr, size_t size, size_t t) {
 
 #ifdef DEBUG
 template <typename S>
-__device__ void print_scalar(S element, int data) {
-    printf("D# %d, T# %d: 0x%08x%08x%08x%08x%08x%08x%08x%08x\n",
-           data,
-           threadIdx.x,
-           element.limbs_storage.limbs[0],
-           element.limbs_storage.limbs[1],
-           element.limbs_storage.limbs[2],
-           element.limbs_storage.limbs[3],
-           element.limbs_storage.limbs[4],
-           element.limbs_storage.limbs[5],
-           element.limbs_storage.limbs[6],
-           element.limbs_storage.limbs[7]
-    );
+__device__ void print_scalar(S element, int data)
+{
+  printf(
+    "D# %d, T# %d: 0x%08x%08x%08x%08x%08x%08x%08x%08x\n", data, threadIdx.x, element.limbs_storage.limbs[0],
+    element.limbs_storage.limbs[1], element.limbs_storage.limbs[2], element.limbs_storage.limbs[3],
+    element.limbs_storage.limbs[4], element.limbs_storage.limbs[5], element.limbs_storage.limbs[6],
+    element.limbs_storage.limbs[7]);
 }
 #endif
 
 template <typename S>
 struct PoseidonConfiguration {
-    uint32_t partial_rounds, full_rounds_half, t;
-    S * round_constants, * mds_matrix, * non_sparse_matrix, *sparse_matrices;
+  uint32_t partial_rounds, full_rounds_half, t;
+  S *round_constants, *mds_matrix, *non_sparse_matrix, *sparse_matrices;
 };
 
 template <typename S>
-class Poseidon {
-  public:
-    uint32_t t;
-    PoseidonConfiguration<S> config;
+class Poseidon
+{
+public:
+  uint32_t t;
+  PoseidonConfiguration<S> config;
 
-    enum HashType {
-        ConstInputLen,
-        MerkleTree,
-    };
+  enum HashType {
+    ConstInputLen,
+    MerkleTree,
+  };
 
-    Poseidon(const uint32_t arity, cudaStream_t stream) {
-        t = arity + 1;
-        this->config.t = t;
-        this->stream = stream;
+  Poseidon(const uint32_t arity, cudaStream_t stream)
+  {
+    t = arity + 1;
+    this->config.t = t;
+    this->stream = stream;
 
-        // Pre-calculate domain tags
-        // Domain tags will vary for different applications of Poseidon
-        uint32_t tree_domain_tag_value = 1;
-        tree_domain_tag_value = (tree_domain_tag_value << arity) - tree_domain_tag_value;
-        tree_domain_tag = S::from(tree_domain_tag_value);
+    // Pre-calculate domain tags
+    // Domain tags will vary for different applications of Poseidon
+    uint32_t tree_domain_tag_value = 1;
+    tree_domain_tag_value = (tree_domain_tag_value << arity) - tree_domain_tag_value;
+    tree_domain_tag = S::from(tree_domain_tag_value);
 
-        const_input_no_pad_domain_tag = S::one();
+    const_input_no_pad_domain_tag = S::one();
 
-        // TO-DO: implement binary shifts for scalar type
-        // const_input_no_pad_domain_tag = S::one() << 64;
-        // const_input_no_pad_domain_tag *= S::from(arity);
+    // TO-DO: implement binary shifts for scalar type
+    // const_input_no_pad_domain_tag = S::one() << 64;
+    // const_input_no_pad_domain_tag *= S::from(arity);
 
-        this->config.full_rounds_half = FULL_ROUNDS_DEFAULT;
-        this->config.partial_rounds = partial_rounds_number_from_arity(arity);
+    this->config.full_rounds_half = FULL_ROUNDS_DEFAULT;
+    this->config.partial_rounds = partial_rounds_number_from_arity(arity);
 
-        uint32_t round_constants_len = t * this->config.full_rounds_half * 2 + this->config.partial_rounds;
-        uint32_t mds_matrix_len = t * t;
-        uint32_t sparse_matrices_len = (t * 2 - 1) * this->config.partial_rounds;
+    uint32_t round_constants_len = t * this->config.full_rounds_half * 2 + this->config.partial_rounds;
+    uint32_t mds_matrix_len = t * t;
+    uint32_t sparse_matrices_len = (t * 2 - 1) * this->config.partial_rounds;
 
-        // All the constants are stored in a single file
-        S * constants = load_constants<S>(arity);
+    // All the constants are stored in a single file
+    S* constants = load_constants<S>(arity);
 
-        S * mds_offset = constants + round_constants_len;
-        S * non_sparse_offset = mds_offset + mds_matrix_len;
-        S * sparse_matrices_offset = non_sparse_offset + mds_matrix_len;
+    S* mds_offset = constants + round_constants_len;
+    S* non_sparse_offset = mds_offset + mds_matrix_len;
+    S* sparse_matrices_offset = non_sparse_offset + mds_matrix_len;
 
-        #if !defined(__CUDA_ARCH__) && defined(DEBUG)
-        std::cout << "P: " << this->config.partial_rounds << " F: " << this->config.full_rounds_half << std::endl;
-        #endif
+#if !defined(__CUDA_ARCH__) && defined(DEBUG)
+    std::cout << "P: " << this->config.partial_rounds << " F: " << this->config.full_rounds_half << std::endl;
+#endif
 
-        // Create streams for copying constants
-        cudaStream_t stream_copy_round_constants, stream_copy_mds_matrix, stream_copy_non_sparse, stream_copy_sparse_matrices;
-        cudaStreamCreate(&stream_copy_round_constants);
-        cudaStreamCreate(&stream_copy_mds_matrix);
-        cudaStreamCreate(&stream_copy_non_sparse);
-        cudaStreamCreate(&stream_copy_sparse_matrices);
-        
-        // Create events for copying constants
-        cudaEvent_t event_copied_round_constants, event_copy_mds_matrix, event_copy_non_sparse, event_copy_sparse_matrices;
-        cudaEventCreateWithFlags(&event_copied_round_constants, cudaEventDisableTiming);
-        cudaEventCreateWithFlags(&event_copy_mds_matrix, cudaEventDisableTiming);
-        cudaEventCreateWithFlags(&event_copy_non_sparse, cudaEventDisableTiming);
-        cudaEventCreateWithFlags(&event_copy_sparse_matrices, cudaEventDisableTiming);
+    // Create streams for copying constants
+    cudaStream_t stream_copy_round_constants, stream_copy_mds_matrix, stream_copy_non_sparse,
+      stream_copy_sparse_matrices;
+    cudaStreamCreate(&stream_copy_round_constants);
+    cudaStreamCreate(&stream_copy_mds_matrix);
+    cudaStreamCreate(&stream_copy_non_sparse);
+    cudaStreamCreate(&stream_copy_sparse_matrices);
 
-        // Malloc memory for copying constants
-        cudaMallocAsync(&this->config.round_constants, sizeof(S) * round_constants_len, stream_copy_round_constants);
-        cudaMallocAsync(&this->config.mds_matrix, sizeof(S) * mds_matrix_len, stream_copy_mds_matrix);
-        cudaMallocAsync(&this->config.non_sparse_matrix, sizeof(S) * mds_matrix_len, stream_copy_non_sparse);
-        cudaMallocAsync(&this->config.sparse_matrices, sizeof(S) * sparse_matrices_len, stream_copy_sparse_matrices);
+    // Create events for copying constants
+    cudaEvent_t event_copied_round_constants, event_copy_mds_matrix, event_copy_non_sparse, event_copy_sparse_matrices;
+    cudaEventCreateWithFlags(&event_copied_round_constants, cudaEventDisableTiming);
+    cudaEventCreateWithFlags(&event_copy_mds_matrix, cudaEventDisableTiming);
+    cudaEventCreateWithFlags(&event_copy_non_sparse, cudaEventDisableTiming);
+    cudaEventCreateWithFlags(&event_copy_sparse_matrices, cudaEventDisableTiming);
 
-        // Copy constants
-        cudaMemcpyAsync(this->config.round_constants, constants,
-            sizeof(S) * round_constants_len,
-            cudaMemcpyHostToDevice, stream_copy_round_constants
-        );
-        cudaMemcpyAsync(this->config.mds_matrix, mds_offset,
-            sizeof(S) * mds_matrix_len,
-            cudaMemcpyHostToDevice, stream_copy_mds_matrix
-        );
-        cudaMemcpyAsync(this->config.non_sparse_matrix, non_sparse_offset,
-            sizeof(S) * mds_matrix_len,
-            cudaMemcpyHostToDevice, stream_copy_non_sparse
-        );
-        cudaMemcpyAsync(this->config.sparse_matrices, sparse_matrices_offset,
-            sizeof(S) * sparse_matrices_len,
-            cudaMemcpyHostToDevice, stream_copy_sparse_matrices
-        );
+    // Malloc memory for copying constants
+    cudaMallocAsync(&this->config.round_constants, sizeof(S) * round_constants_len, stream_copy_round_constants);
+    cudaMallocAsync(&this->config.mds_matrix, sizeof(S) * mds_matrix_len, stream_copy_mds_matrix);
+    cudaMallocAsync(&this->config.non_sparse_matrix, sizeof(S) * mds_matrix_len, stream_copy_non_sparse);
+    cudaMallocAsync(&this->config.sparse_matrices, sizeof(S) * sparse_matrices_len, stream_copy_sparse_matrices);
 
-        // Record finished copying event for streams
-        cudaEventRecord(event_copied_round_constants, stream_copy_round_constants);
-        cudaEventRecord(event_copy_mds_matrix, stream_copy_mds_matrix);
-        cudaEventRecord(event_copy_non_sparse, stream_copy_non_sparse);
-        cudaEventRecord(event_copy_sparse_matrices, stream_copy_sparse_matrices);
+    // Copy constants
+    cudaMemcpyAsync(
+      this->config.round_constants, constants, sizeof(S) * round_constants_len, cudaMemcpyHostToDevice,
+      stream_copy_round_constants);
+    cudaMemcpyAsync(
+      this->config.mds_matrix, mds_offset, sizeof(S) * mds_matrix_len, cudaMemcpyHostToDevice, stream_copy_mds_matrix);
+    cudaMemcpyAsync(
+      this->config.non_sparse_matrix, non_sparse_offset, sizeof(S) * mds_matrix_len, cudaMemcpyHostToDevice,
+      stream_copy_non_sparse);
+    cudaMemcpyAsync(
+      this->config.sparse_matrices, sparse_matrices_offset, sizeof(S) * sparse_matrices_len, cudaMemcpyHostToDevice,
+      stream_copy_sparse_matrices);
 
-        // Main stream waits for copying to finish
-        cudaStreamWaitEvent(stream, event_copied_round_constants);
-        cudaStreamWaitEvent(stream, event_copy_mds_matrix);
-        cudaStreamWaitEvent(stream, event_copy_non_sparse);
-        cudaStreamWaitEvent(stream, event_copy_sparse_matrices);
-    }
+    // Record finished copying event for streams
+    cudaEventRecord(event_copied_round_constants, stream_copy_round_constants);
+    cudaEventRecord(event_copy_mds_matrix, stream_copy_mds_matrix);
+    cudaEventRecord(event_copy_non_sparse, stream_copy_non_sparse);
+    cudaEventRecord(event_copy_sparse_matrices, stream_copy_sparse_matrices);
 
-    ~Poseidon() {
-        cudaFreeAsync(this->config.round_constants, this->stream);
-        cudaFreeAsync(this->config.mds_matrix, this->stream);
-        cudaFreeAsync(this->config.non_sparse_matrix, this->stream);
-        cudaFreeAsync(this->config.sparse_matrices, this->stream);
-    }
+    // Main stream waits for copying to finish
+    cudaStreamWaitEvent(stream, event_copied_round_constants);
+    cudaStreamWaitEvent(stream, event_copy_mds_matrix);
+    cudaStreamWaitEvent(stream, event_copy_non_sparse);
+    cudaStreamWaitEvent(stream, event_copy_sparse_matrices);
+  }
 
-    // Hash multiple preimages in parallel
-    void hash_blocks(const S * inp, size_t blocks, S * out, HashType hash_type, cudaStream_t stream);
+  ~Poseidon()
+  {
+    cudaFreeAsync(this->config.round_constants, this->stream);
+    cudaFreeAsync(this->config.mds_matrix, this->stream);
+    cudaFreeAsync(this->config.non_sparse_matrix, this->stream);
+    cudaFreeAsync(this->config.sparse_matrices, this->stream);
+  }
 
-  private:
-    S tree_domain_tag, const_input_no_pad_domain_tag;
-    cudaStream_t stream;
+  // Hash multiple preimages in parallel
+  void hash_blocks(const S* inp, size_t blocks, S* out, HashType hash_type, cudaStream_t stream);
+
+private:
+  S tree_domain_tag, const_input_no_pad_domain_tag;
+  cudaStream_t stream;
 };

icicle/appUtils/vector_manipulation/ve_mod_mult.cuh

@@ -1,115 +1,135 @@
 #ifndef VEC_MULT
 #define VEC_MULT
 #pragma once
-#include <stdexcept>
 #include <cuda.h>
-
+#include <stdexcept>
 
 #define MAX_THREADS_PER_BLOCK 256
 
+/**
+ * Multiply the elements of an input array by a scalar in-place.
+ * @param arr input array.
+ * @param n size of arr.
+ * @param n_inv scalar of type S (scalar).
+ */
+template <typename E, typename S>
+__global__ void template_normalize_kernel(E* arr, uint32_t n, S scalar)
+{
+  int tid = (blockIdx.x * blockDim.x) + threadIdx.x;
+  if (tid < n) { arr[tid] = scalar * arr[tid]; }
+}
+
 // TODO: headers for prototypes and .c .cpp .cu files for implementations
 template <typename E, typename S>
-__global__ void vectorModMult(S *scalar_vec, E *element_vec, E *result, size_t n_elments)
+__global__ void vectorModMult(S* scalar_vec, E* element_vec, E* result, size_t n_elments)
 {
-    int tid = blockDim.x * blockIdx.x + threadIdx.x;
-    if (tid < n_elments)
-    {
-        result[tid] = scalar_vec[tid] * element_vec[tid];
-    }
+  int tid = blockDim.x * blockIdx.x + threadIdx.x;
+  if (tid < n_elments) { result[tid] = scalar_vec[tid] * element_vec[tid]; }
 }
 
 template <typename E, typename S>
-int vector_mod_mult(S *vec_a, E *vec_b, E *result, size_t n_elments, cudaStream_t stream) // TODO: in place so no need for third result vector
+int vector_mod_mult(S* vec_a, E* vec_b, E* result, size_t n_elments, cudaStream_t stream) // TODO: in place so no need
+                                                                                          // for third result vector
 {
-    // Set the grid and block dimensions
-    int num_blocks = (int)ceil((float)n_elments / MAX_THREADS_PER_BLOCK);
-    int threads_per_block = MAX_THREADS_PER_BLOCK;
+  // Set the grid and block dimensions
+  int num_blocks = (int)ceil((float)n_elments / MAX_THREADS_PER_BLOCK);
+  int threads_per_block = MAX_THREADS_PER_BLOCK;
 
-    // Allocate memory on the device for the input vectors, the output vector, and the modulus
-    S *d_vec_a;
-    E *d_vec_b, *d_result;
-    cudaMallocAsync(&d_vec_a, n_elments * sizeof(S), stream);
-    cudaMallocAsync(&d_vec_b, n_elments * sizeof(E), stream);
-    cudaMallocAsync(&d_result, n_elments * sizeof(E), stream);
+  // Allocate memory on the device for the input vectors, the output vector, and the modulus
+  S* d_vec_a;
+  E *d_vec_b, *d_result;
+  cudaMallocAsync(&d_vec_a, n_elments * sizeof(S), stream);
+  cudaMallocAsync(&d_vec_b, n_elments * sizeof(E), stream);
+  cudaMallocAsync(&d_result, n_elments * sizeof(E), stream);
 
-    // Copy the input vectors and the modulus from the host to the device
-    cudaMemcpyAsync(d_vec_a, vec_a, n_elments * sizeof(S), cudaMemcpyHostToDevice, stream);
-    cudaMemcpyAsync(d_vec_b, vec_b, n_elments * sizeof(E), cudaMemcpyHostToDevice, stream);
+  // Copy the input vectors and the modulus from the host to the device
+  cudaMemcpyAsync(d_vec_a, vec_a, n_elments * sizeof(S), cudaMemcpyHostToDevice, stream);
+  cudaMemcpyAsync(d_vec_b, vec_b, n_elments * sizeof(E), cudaMemcpyHostToDevice, stream);
 
-    // Call the kernel to perform element-wise modular multiplication
-    vectorModMult<<<num_blocks, threads_per_block, 0, stream>>>(d_vec_a, d_vec_b, d_result, n_elments);
+  // Call the kernel to perform element-wise modular multiplication
+  vectorModMult<<<num_blocks, threads_per_block, 0, stream>>>(d_vec_a, d_vec_b, d_result, n_elments);
 
-    cudaMemcpyAsync(result, d_result, n_elments * sizeof(E), cudaMemcpyDeviceToHost, stream);
+  cudaMemcpyAsync(result, d_result, n_elments * sizeof(E), cudaMemcpyDeviceToHost, stream);
 
-    cudaFreeAsync(d_vec_a, stream);
-    cudaFreeAsync(d_vec_b, stream);
-    cudaFreeAsync(d_result, stream);
+  cudaFreeAsync(d_vec_a, stream);
+  cudaFreeAsync(d_vec_b, stream);
+  cudaFreeAsync(d_result, stream);
 
-    cudaStreamSynchronize(stream);
-    return 0;
+  cudaStreamSynchronize(stream);
+  return 0;
 }
 
 template <typename E, typename S>
-__global__ void batchVectorMult(S *scalar_vec, E *element_vec, unsigned n_scalars, unsigned batch_size)
+int vector_mod_mult_device(
+  S* d_vec_a, E* d_vec_b, E* d_result, size_t n_elments) // TODO: in place so no need for third result vector
 {
-    int tid = blockDim.x * blockIdx.x + threadIdx.x;
-    if (tid < n_scalars * batch_size)
-    {
-        int scalar_id = tid % n_scalars;
-        element_vec[tid] = scalar_vec[scalar_id] * element_vec[tid];
-    }
+  // Set the grid and block dimensions
+  int num_blocks = (int)ceil((float)n_elments / MAX_THREADS_PER_BLOCK);
+  int threads_per_block = MAX_THREADS_PER_BLOCK;
+
+  // Call the kernel to perform element-wise modular multiplication
+  vectorModMult<<<num_blocks, threads_per_block>>>(d_vec_a, d_vec_b, d_result, n_elments);
+  return 0;
 }
 
 template <typename E, typename S>
-int batch_vector_mult(S *scalar_vec, E *element_vec, unsigned n_scalars, unsigned batch_size, cudaStream_t stream)
+__global__ void batchVectorMult(S* scalar_vec, E* element_vec, unsigned n_scalars, unsigned batch_size)
 {
-    // Set the grid and block dimensions
-    int NUM_THREADS = MAX_THREADS_PER_BLOCK;
-    int NUM_BLOCKS = (n_scalars * batch_size + NUM_THREADS - 1) / NUM_THREADS;
-    batchVectorMult<<<NUM_BLOCKS, NUM_THREADS, 0, stream>>>(scalar_vec, element_vec, n_scalars, batch_size);
-    return 0;
+  int tid = blockDim.x * blockIdx.x + threadIdx.x;
+  if (tid < n_scalars * batch_size) {
+    int scalar_id = tid % n_scalars;
+    element_vec[tid] = scalar_vec[scalar_id] * element_vec[tid];
+  }
+}
+
+template <typename E, typename S>
+int batch_vector_mult(S* scalar_vec, E* element_vec, unsigned n_scalars, unsigned batch_size, cudaStream_t stream)
+{
+  // Set the grid and block dimensions
+  int NUM_THREADS = MAX_THREADS_PER_BLOCK;
+  int NUM_BLOCKS = (n_scalars * batch_size + NUM_THREADS - 1) / NUM_THREADS;
+  batchVectorMult<<<NUM_BLOCKS, NUM_THREADS, 0, stream>>>(scalar_vec, element_vec, n_scalars, batch_size);
+  return 0;
 }
 
 template <typename E>
-__global__ void matrixVectorMult(E *matrix_elements, E *vector_elements, E *result, size_t dim)
+__global__ void matrixVectorMult(E* matrix_elements, E* vector_elements, E* result, size_t dim)
 {
-
-    int tid = blockDim.x * blockIdx.x + threadIdx.x;
-    if (tid < dim)
-    {
-        result[tid] = E::zero();
-        for (int i = 0; i < dim; i++)
-            result[tid] = result[tid] + matrix_elements[tid * dim + i] * vector_elements[i];
-    }
+  int tid = blockDim.x * blockIdx.x + threadIdx.x;
+  if (tid < dim) {
+    result[tid] = E::zero();
+    for (int i = 0; i < dim; i++)
+      result[tid] = result[tid] + matrix_elements[tid * dim + i] * vector_elements[i];
+  }
 }
 
 template <typename E>
-int matrix_mod_mult(E *matrix_elements, E *vector_elements, E *result, size_t dim, cudaStream_t stream)
+int matrix_mod_mult(E* matrix_elements, E* vector_elements, E* result, size_t dim, cudaStream_t stream)
 {
-    // Set the grid and block dimensions
-    int num_blocks = (int)ceil((float)dim / MAX_THREADS_PER_BLOCK);
-    int threads_per_block = MAX_THREADS_PER_BLOCK;
+  // Set the grid and block dimensions
+  int num_blocks = (int)ceil((float)dim / MAX_THREADS_PER_BLOCK);
+  int threads_per_block = MAX_THREADS_PER_BLOCK;
 
-    // Allocate memory on the device for the input vectors, the output vector, and the modulus
-    E *d_matrix, *d_vector, *d_result;
-    cudaMallocAsync(&d_matrix, (dim * dim) * sizeof(E), stream);
-    cudaMallocAsync(&d_vector, dim * sizeof(E), stream);
-    cudaMallocAsync(&d_result, dim * sizeof(E), stream);
+  // Allocate memory on the device for the input vectors, the output vector, and the modulus
+  E *d_matrix, *d_vector, *d_result;
+  cudaMallocAsync(&d_matrix, (dim * dim) * sizeof(E), stream);
+  cudaMallocAsync(&d_vector, dim * sizeof(E), stream);
+  cudaMallocAsync(&d_result, dim * sizeof(E), stream);
 
-    // Copy the input vectors and the modulus from the host to the device
-    cudaMemcpyAsync(d_matrix, matrix_elements, (dim * dim) * sizeof(E), cudaMemcpyHostToDevice, stream);
-    cudaMemcpyAsync(d_vector, vector_elements, dim * sizeof(E), cudaMemcpyHostToDevice, stream);
+  // Copy the input vectors and the modulus from the host to the device
+  cudaMemcpyAsync(d_matrix, matrix_elements, (dim * dim) * sizeof(E), cudaMemcpyHostToDevice, stream);
+  cudaMemcpyAsync(d_vector, vector_elements, dim * sizeof(E), cudaMemcpyHostToDevice, stream);
 
-    // Call the kernel to perform element-wise modular multiplication
-    matrixVectorMult<<<num_blocks, threads_per_block, 0, stream>>>(d_matrix, d_vector, d_result, dim);
+  // Call the kernel to perform element-wise modular multiplication
+  matrixVectorMult<<<num_blocks, threads_per_block, 0, stream>>>(d_matrix, d_vector, d_result, dim);
 
-    cudaMemcpyAsync(result, d_result, dim * sizeof(E), cudaMemcpyDeviceToHost, stream);
+  cudaMemcpyAsync(result, d_result, dim * sizeof(E), cudaMemcpyDeviceToHost, stream);
 
-    cudaFreeAsync(d_matrix, stream);
-    cudaFreeAsync(d_vector, stream);
-    cudaFreeAsync(d_result, stream);
+  cudaFreeAsync(d_matrix, stream);
+  cudaFreeAsync(d_vector, stream);
+  cudaFreeAsync(d_result, stream);
 
-    cudaStreamSynchronize(stream);
-    return 0;
+  cudaStreamSynchronize(stream);
+  return 0;
 }
 #endif

icicle/curves/bls12_377/curve_config.cuh

@@ -2,21 +2,24 @@
 
 #include "../../primitives/field.cuh"
 #include "../../primitives/projective.cuh"
+#if defined(G2_DEFINED)
+#include "../../primitives/extension_field.cuh"
+#endif
 
 #include "params.cuh"
 
 namespace BLS12_377 {
-    typedef Field<PARAMS_BLS12_377::fp_config> scalar_field_t;
-    typedef scalar_field_t scalar_t;
-    typedef Field<PARAMS_BLS12_377::fq_config> point_field_t;
-    static constexpr point_field_t b = point_field_t{ PARAMS_BLS12_377::weierstrass_b };
-    typedef Projective<point_field_t, scalar_field_t, b> projective_t;
-    typedef Affine<point_field_t> affine_t;
-    #if defined(G2_DEFINED)
-    typedef ExtensionField<PARAMS_BLS12_377::fq_config> g2_point_field_t;
-    static constexpr g2_point_field_t b_g2 = g2_point_field_t{ point_field_t{ PARAMS_BLS12_377::weierstrass_b_g2_re },
-                                                               point_field_t{ PARAMS_BLS12_377::weierstrass_b_g2_im }};
-    typedef Projective<g2_point_field_t, scalar_field_t, b_g2> g2_projective_t;
-    typedef Affine<g2_point_field_t> g2_affine_t;
-    #endif
-}
+  typedef Field<PARAMS_BLS12_377::fp_config> scalar_field_t;
+  typedef scalar_field_t scalar_t;
+  typedef Field<PARAMS_BLS12_377::fq_config> point_field_t;
+  static constexpr point_field_t b = point_field_t{PARAMS_BLS12_377::weierstrass_b};
+  typedef Projective<point_field_t, scalar_field_t, b> projective_t;
+  typedef Affine<point_field_t> affine_t;
+#if defined(G2_DEFINED)
+  typedef ExtensionField<PARAMS_BLS12_377::fq_config> g2_point_field_t;
+  static constexpr g2_point_field_t b_g2 = g2_point_field_t{
+    point_field_t{PARAMS_BLS12_377::weierstrass_b_g2_re}, point_field_t{PARAMS_BLS12_377::weierstrass_b_g2_im}};
+  typedef Projective<g2_point_field_t, scalar_field_t, b_g2> g2_projective_t;
+  typedef Affine<g2_point_field_t> g2_affine_t;
+#endif
+} // namespace BLS12_377

icicle/curves/bls12_377/lde.cu

@@ -1,327 +1,610 @@
 #ifndef _BLS12_377_LDE
 #define _BLS12_377_LDE
-#include <cuda.h>
 #include "../../appUtils/ntt/lde.cu"
 #include "../../appUtils/ntt/ntt.cuh"
 #include "../../appUtils/vector_manipulation/ve_mod_mult.cuh"
+#include "../../utils/mont.cuh"
 #include "curve_config.cuh"
+#include <cuda.h>
 
-extern "C" BLS12_377::scalar_t* build_domain_cuda_bls12_377(uint32_t domain_size, uint32_t logn, bool inverse, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" BLS12_377::scalar_t* build_domain_cuda_bls12_377(
+  uint32_t domain_size, uint32_t logn, bool inverse, size_t device_id = 0, cudaStream_t stream = 0)
 {
-    try
-    {
-        cudaStreamCreate(&stream);
-        if (inverse) {
-            return fill_twiddle_factors_array(domain_size, BLS12_377::scalar_t::omega_inv(logn), stream);
-        } else {
-            return fill_twiddle_factors_array(domain_size, BLS12_377::scalar_t::omega(logn), stream);
-        }
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return nullptr;
+  try {
+    cudaStreamCreate(&stream);
+    if (inverse) {
+      return fill_twiddle_factors_array(domain_size, BLS12_377::scalar_t::omega_inv(logn), stream);
+    } else {
+      return fill_twiddle_factors_array(domain_size, BLS12_377::scalar_t::omega(logn), stream);
     }
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return nullptr;
+  }
 }
 
-extern "C" int ntt_cuda_bls12_377(BLS12_377::scalar_t *arr, uint32_t n, bool inverse, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int ntt_cuda_bls12_377(
+  BLS12_377::scalar_t* arr,
+  uint32_t n,
+  bool inverse,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        cudaStreamCreate(&stream);
-        return ntt_end2end_template<BLS12_377::scalar_t,BLS12_377::scalar_t>(arr, n, inverse, stream); // TODO: pass device_id
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        
-        return -1;        
-    }
+  try {
+    cudaStreamCreate(&stream);
+    return ntt_end2end_template<BLS12_377::scalar_t, BLS12_377::scalar_t>(
+      arr, n, inverse, stream); // TODO: pass device_id
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+
+    return -1;
+  }
 }
 
-extern "C" int ecntt_cuda_bls12_377(BLS12_377::projective_t *arr, uint32_t n, bool inverse, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int ecntt_cuda_bls12_377(
+  BLS12_377::projective_t* arr,
+  uint32_t n,
+  bool inverse,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        cudaStreamCreate(&stream);
-        return ntt_end2end_template<BLS12_377::projective_t,BLS12_377::scalar_t>(arr, n, inverse, stream); // TODO: pass device_id
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    cudaStreamCreate(&stream);
+    return ntt_end2end_template<BLS12_377::projective_t, BLS12_377::scalar_t>(
+      arr, n, inverse, stream); // TODO: pass device_id
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int ntt_batch_cuda_bls12_377(BLS12_377::scalar_t *arr, uint32_t arr_size, uint32_t batch_size, bool inverse, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int ntt_batch_cuda_bls12_377(
+  BLS12_377::scalar_t* arr,
+  uint32_t arr_size,
+  uint32_t batch_size,
+  bool inverse,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        cudaStreamCreate(&stream);
-        return ntt_end2end_batch_template<BLS12_377::scalar_t,BLS12_377::scalar_t>(arr, arr_size, batch_size, inverse, stream); // TODO: pass device_id
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    cudaStreamCreate(&stream);
+    return ntt_end2end_batch_template<BLS12_377::scalar_t, BLS12_377::scalar_t>(
+      arr, arr_size, batch_size, inverse, stream); // TODO: pass device_id
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int ecntt_batch_cuda_bls12_377(BLS12_377::projective_t *arr, uint32_t arr_size, uint32_t batch_size, bool inverse, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int ecntt_batch_cuda_bls12_377(
+  BLS12_377::projective_t* arr,
+  uint32_t arr_size,
+  uint32_t batch_size,
+  bool inverse,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        cudaStreamCreate(&stream);
-        return ntt_end2end_batch_template<BLS12_377::projective_t,BLS12_377::scalar_t>(arr, arr_size, batch_size, inverse, stream); // TODO: pass device_id
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    cudaStreamCreate(&stream);
+    return ntt_end2end_batch_template<BLS12_377::projective_t, BLS12_377::scalar_t>(
+      arr, arr_size, batch_size, inverse, stream); // TODO: pass device_id
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int interpolate_scalars_cuda_bls12_377(BLS12_377::scalar_t* d_out, BLS12_377::scalar_t *d_evaluations, BLS12_377::scalar_t *d_domain, unsigned n, unsigned device_id = 0, cudaStream_t stream = 0)
+extern "C" int interpolate_scalars_cuda_bls12_377(
+  BLS12_377::scalar_t* d_out,
+  BLS12_377::scalar_t* d_evaluations,
+  BLS12_377::scalar_t* d_domain,
+  unsigned n,
+  unsigned device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        return interpolate(d_out, d_evaluations, d_domain, n, stream);
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    BLS12_377::scalar_t* _null = nullptr;
+    return interpolate(d_out, d_evaluations, d_domain, n, false, _null, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int interpolate_scalars_batch_cuda_bls12_377(BLS12_377::scalar_t* d_out, BLS12_377::scalar_t* d_evaluations, BLS12_377::scalar_t* d_domain, unsigned n,
-                                              unsigned batch_size, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int interpolate_scalars_batch_cuda_bls12_377(
+  BLS12_377::scalar_t* d_out,
+  BLS12_377::scalar_t* d_evaluations,
+  BLS12_377::scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        cudaStreamCreate(&stream);
-        return interpolate_batch(d_out, d_evaluations, d_domain, n, batch_size, stream);
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    BLS12_377::scalar_t* _null = nullptr;
+    cudaStreamCreate(&stream);
+    return interpolate_batch(d_out, d_evaluations, d_domain, n, batch_size, false, _null, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int interpolate_points_cuda_bls12_377(BLS12_377::projective_t* d_out, BLS12_377::projective_t *d_evaluations, BLS12_377::scalar_t *d_domain, unsigned n, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int interpolate_scalars_on_coset_cuda_bls12_377(
+  BLS12_377::scalar_t* d_out,
+  BLS12_377::scalar_t* d_evaluations,
+  BLS12_377::scalar_t* d_domain,
+  unsigned n,
+  BLS12_377::scalar_t* coset_powers,
+  unsigned device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        return interpolate(d_out, d_evaluations, d_domain, n, stream);
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    return interpolate(d_out, d_evaluations, d_domain, n, true, coset_powers, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int interpolate_points_batch_cuda_bls12_377(BLS12_377::projective_t* d_out, BLS12_377::projective_t* d_evaluations, BLS12_377::scalar_t* d_domain,
-                                             unsigned n, unsigned batch_size, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int interpolate_scalars_batch_on_coset_cuda_bls12_377(
+  BLS12_377::scalar_t* d_out,
+  BLS12_377::scalar_t* d_evaluations,
+  BLS12_377::scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  BLS12_377::scalar_t* coset_powers,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        cudaStreamCreate(&stream);
-        return interpolate_batch(d_out, d_evaluations, d_domain, n, batch_size, stream);
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    cudaStreamCreate(&stream);
+    return interpolate_batch(d_out, d_evaluations, d_domain, n, batch_size, true, coset_powers, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int evaluate_scalars_cuda_bls12_377(BLS12_377::scalar_t* d_out, BLS12_377::scalar_t *d_coefficients, BLS12_377::scalar_t *d_domain, 
-                                     unsigned domain_size, unsigned n, unsigned device_id = 0, cudaStream_t stream = 0)
+extern "C" int interpolate_points_cuda_bls12_377(
+  BLS12_377::projective_t* d_out,
+  BLS12_377::projective_t* d_evaluations,
+  BLS12_377::scalar_t* d_domain,
+  unsigned n,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        BLS12_377::scalar_t* _null = nullptr;
-        cudaStreamCreate(&stream);
-        return evaluate(d_out, d_coefficients, d_domain, domain_size, n, false, _null, stream);
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    BLS12_377::scalar_t* _null = nullptr;
+    return interpolate(d_out, d_evaluations, d_domain, n, false, _null, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int evaluate_scalars_batch_cuda_bls12_377(BLS12_377::scalar_t* d_out, BLS12_377::scalar_t* d_coefficients, BLS12_377::scalar_t* d_domain, unsigned domain_size,
-                                           unsigned n, unsigned batch_size, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int interpolate_points_batch_cuda_bls12_377(
+  BLS12_377::projective_t* d_out,
+  BLS12_377::projective_t* d_evaluations,
+  BLS12_377::scalar_t* d_domain,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        BLS12_377::scalar_t* _null = nullptr;
-        cudaStreamCreate(&stream);
-        return evaluate_batch(d_out, d_coefficients, d_domain, domain_size, n, batch_size, false, _null, stream);
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    BLS12_377::scalar_t* _null = nullptr;
+    cudaStreamCreate(&stream);
+    return interpolate_batch(d_out, d_evaluations, d_domain, n, batch_size, false, _null, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int evaluate_points_cuda_bls12_377(BLS12_377::projective_t* d_out, BLS12_377::projective_t *d_coefficients, BLS12_377::scalar_t *d_domain, 
-                                    unsigned domain_size, unsigned n, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int evaluate_scalars_cuda_bls12_377(
+  BLS12_377::scalar_t* d_out,
+  BLS12_377::scalar_t* d_coefficients,
+  BLS12_377::scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        BLS12_377::scalar_t* _null = nullptr;
-        cudaStreamCreate(&stream);
-        return evaluate(d_out, d_coefficients, d_domain, domain_size, n, false, _null, stream);
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    BLS12_377::scalar_t* _null = nullptr;
+    cudaStreamCreate(&stream);
+    return evaluate(d_out, d_coefficients, d_domain, domain_size, n, false, _null, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int evaluate_points_batch_cuda_bls12_377(BLS12_377::projective_t* d_out, BLS12_377::projective_t* d_coefficients, BLS12_377::scalar_t* d_domain, unsigned domain_size,
-                                          unsigned n, unsigned batch_size, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int evaluate_scalars_batch_cuda_bls12_377(
+  BLS12_377::scalar_t* d_out,
+  BLS12_377::scalar_t* d_coefficients,
+  BLS12_377::scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        BLS12_377::scalar_t* _null = nullptr;
-        cudaStreamCreate(&stream);
-        return evaluate_batch(d_out, d_coefficients, d_domain, domain_size, n, batch_size, false, _null, stream);
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    BLS12_377::scalar_t* _null = nullptr;
+    cudaStreamCreate(&stream);
+    return evaluate_batch(d_out, d_coefficients, d_domain, domain_size, n, batch_size, false, _null, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int evaluate_scalars_on_coset_cuda_bls12_377(BLS12_377::scalar_t* d_out, BLS12_377::scalar_t *d_coefficients, BLS12_377::scalar_t *d_domain, unsigned domain_size,
-                                              unsigned n, BLS12_377::scalar_t *coset_powers, unsigned device_id = 0, cudaStream_t stream = 0)
+extern "C" int evaluate_points_cuda_bls12_377(
+  BLS12_377::projective_t* d_out,
+  BLS12_377::projective_t* d_coefficients,
+  BLS12_377::scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        cudaStreamCreate(&stream);
-        return evaluate(d_out, d_coefficients, d_domain, domain_size, n, true, coset_powers, stream);
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    BLS12_377::scalar_t* _null = nullptr;
+    cudaStreamCreate(&stream);
+    return evaluate(d_out, d_coefficients, d_domain, domain_size, n, false, _null, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int evaluate_scalars_on_coset_batch_cuda_bls12_377(BLS12_377::scalar_t* d_out, BLS12_377::scalar_t* d_coefficients, BLS12_377::scalar_t* d_domain, unsigned domain_size, 
-                                                    unsigned n, unsigned batch_size, BLS12_377::scalar_t *coset_powers, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int evaluate_points_batch_cuda_bls12_377(
+  BLS12_377::projective_t* d_out,
+  BLS12_377::projective_t* d_coefficients,
+  BLS12_377::scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        cudaStreamCreate(&stream);
-        return evaluate_batch(d_out, d_coefficients, d_domain, domain_size, n, batch_size, true, coset_powers, stream);
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    BLS12_377::scalar_t* _null = nullptr;
+    cudaStreamCreate(&stream);
+    return evaluate_batch(d_out, d_coefficients, d_domain, domain_size, n, batch_size, false, _null, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int evaluate_points_on_coset_cuda_bls12_377(BLS12_377::projective_t* d_out, BLS12_377::projective_t *d_coefficients, BLS12_377::scalar_t *d_domain, unsigned domain_size,
-                                             unsigned n, BLS12_377::scalar_t *coset_powers, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int evaluate_scalars_on_coset_cuda_bls12_377(
+  BLS12_377::scalar_t* d_out,
+  BLS12_377::scalar_t* d_coefficients,
+  BLS12_377::scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  BLS12_377::scalar_t* coset_powers,
+  unsigned device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        cudaStreamCreate(&stream);
-        return evaluate(d_out, d_coefficients, d_domain, domain_size, n, true, coset_powers, stream);
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    cudaStreamCreate(&stream);
+    return evaluate(d_out, d_coefficients, d_domain, domain_size, n, true, coset_powers, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int evaluate_points_on_coset_batch_cuda_bls12_377(BLS12_377::projective_t* d_out, BLS12_377::projective_t* d_coefficients, BLS12_377::scalar_t* d_domain, unsigned domain_size, 
-                                                   unsigned n, unsigned batch_size, BLS12_377::scalar_t *coset_powers, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int evaluate_scalars_on_coset_batch_cuda_bls12_377(
+  BLS12_377::scalar_t* d_out,
+  BLS12_377::scalar_t* d_coefficients,
+  BLS12_377::scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  BLS12_377::scalar_t* coset_powers,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        cudaStreamCreate(&stream);
-        return evaluate_batch(d_out, d_coefficients, d_domain, domain_size, n, batch_size, true, coset_powers, stream);
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    cudaStreamCreate(&stream);
+    return evaluate_batch(d_out, d_coefficients, d_domain, domain_size, n, batch_size, true, coset_powers, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int reverse_order_scalars_cuda_bls12_377(BLS12_377::scalar_t* arr, int n, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int evaluate_points_on_coset_cuda_bls12_377(
+  BLS12_377::projective_t* d_out,
+  BLS12_377::projective_t* d_coefficients,
+  BLS12_377::scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  BLS12_377::scalar_t* coset_powers,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        uint32_t logn = uint32_t(log(n) / log(2));
-        cudaStreamCreate(&stream);
-        reverse_order(arr, n, logn, stream);
-        return 0;
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    cudaStreamCreate(&stream);
+    return evaluate(d_out, d_coefficients, d_domain, domain_size, n, true, coset_powers, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int reverse_order_scalars_batch_cuda_bls12_377(BLS12_377::scalar_t* arr, int n, int batch_size, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int evaluate_points_on_coset_batch_cuda_bls12_377(
+  BLS12_377::projective_t* d_out,
+  BLS12_377::projective_t* d_coefficients,
+  BLS12_377::scalar_t* d_domain,
+  unsigned domain_size,
+  unsigned n,
+  unsigned batch_size,
+  BLS12_377::scalar_t* coset_powers,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        uint32_t logn = uint32_t(log(n) / log(2));
-        cudaStreamCreate(&stream);
-        reverse_order_batch(arr, n, logn, batch_size, stream);
-        return 0;
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    cudaStreamCreate(&stream);
+    return evaluate_batch(d_out, d_coefficients, d_domain, domain_size, n, batch_size, true, coset_powers, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int reverse_order_points_cuda_bls12_377(BLS12_377::projective_t* arr, int n, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int ntt_inplace_batch_cuda_bls12_377(
+  BLS12_377::scalar_t* d_inout,
+  BLS12_377::scalar_t* d_twiddles,
+  unsigned n,
+  unsigned batch_size,
+  bool inverse,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        uint32_t logn = uint32_t(log(n) / log(2));
-        cudaStreamCreate(&stream);
-        reverse_order(arr, n, logn, stream);
-        return 0;
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    cudaStreamCreate(&stream);
+    BLS12_377::scalar_t* _null = nullptr;
+    ntt_inplace_batch_template(d_inout, d_twiddles, n, batch_size, inverse, false, _null, stream, true);
+    return CUDA_SUCCESS; // TODO: we should implement this https://leimao.github.io/blog/Proper-CUDA-Error-Checking/
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int reverse_order_points_batch_cuda_bls12_377(BLS12_377::projective_t* arr, int n, int batch_size, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int ntt_inplace_coset_batch_cuda_bls12_377(
+  BLS12_377::scalar_t* d_inout,
+  BLS12_377::scalar_t* d_twiddles,
+  unsigned n,
+  unsigned batch_size,
+  bool inverse,
+  bool is_coset,
+  BLS12_377::scalar_t* coset,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        uint32_t logn = uint32_t(log(n) / log(2));
-        cudaStreamCreate(&stream);
-        reverse_order_batch(arr, n, logn, batch_size, stream);
-        return 0;
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    cudaStreamCreate(&stream);
+    ntt_inplace_batch_template(d_inout, d_twiddles, n, batch_size, inverse, is_coset, coset, stream, true);
+    return CUDA_SUCCESS; // TODO: we should implement this https://leimao.github.io/blog/Proper-CUDA-Error-Checking/
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
-#endif
+
+extern "C" int sub_scalars_cuda_bls12_377(
+  BLS12_377::scalar_t* d_out,
+  BLS12_377::scalar_t* d_in1,
+  BLS12_377::scalar_t* d_in2,
+  unsigned n,
+  cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    return sub_polys(d_out, d_in1, d_in2, n, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int add_scalars_cuda_bls12_377(
+  BLS12_377::scalar_t* d_out,
+  BLS12_377::scalar_t* d_in1,
+  BLS12_377::scalar_t* d_in2,
+  unsigned n,
+  cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    return add_polys(d_out, d_in1, d_in2, n, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int to_montgomery_scalars_cuda_bls12_377(BLS12_377::scalar_t* d_inout, unsigned n, cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    return to_montgomery(d_inout, n, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int from_montgomery_scalars_cuda_bls12_377(BLS12_377::scalar_t* d_inout, unsigned n, cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    return from_montgomery(d_inout, n, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int
+to_montgomery_proj_points_cuda_bls12_377(BLS12_377::projective_t* d_inout, unsigned n, cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    return to_montgomery((BLS12_377::point_field_t*)d_inout, 3 * n, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int
+from_montgomery_proj_points_cuda_bls12_377(BLS12_377::projective_t* d_inout, unsigned n, cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    return from_montgomery((BLS12_377::point_field_t*)d_inout, 3 * n, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int
+to_montgomery_aff_points_cuda_bls12_377(BLS12_377::affine_t* d_inout, unsigned n, cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    return to_montgomery((BLS12_377::point_field_t*)d_inout, 2 * n, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int
+from_montgomery_aff_points_cuda_bls12_377(BLS12_377::affine_t* d_inout, unsigned n, cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    return from_montgomery((BLS12_377::point_field_t*)d_inout, 2 * n, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+#if defined(G2_DEFINED)
+extern "C" int
+to_montgomery_proj_points_g2_cuda_bls12_377(BLS12_377::g2_projective_t* d_inout, unsigned n, cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    return to_montgomery((BLS12_377::point_field_t*)d_inout, 6 * n, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int
+from_montgomery_proj_points_g2_cuda_bls12_377(BLS12_377::g2_projective_t* d_inout, unsigned n, cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    return from_montgomery((BLS12_377::point_field_t*)d_inout, 6 * n, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int
+to_montgomery_aff_points_g2_cuda_bls12_377(BLS12_377::g2_affine_t* d_inout, unsigned n, cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    return to_montgomery((BLS12_377::point_field_t*)d_inout, 4 * n, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int
+from_montgomery_aff_points_g2_cuda_bls12_377(BLS12_377::g2_affine_t* d_inout, unsigned n, cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    return from_montgomery((BLS12_377::point_field_t*)d_inout, 4 * n, stream);
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+#endif
+
+extern "C" int
+reverse_order_scalars_cuda_bls12_377(BLS12_377::scalar_t* arr, int n, size_t device_id = 0, cudaStream_t stream = 0)
+{
+  try {
+    uint32_t logn = uint32_t(log(n) / log(2));
+    cudaStreamCreate(&stream);
+    reverse_order(arr, n, logn, stream);
+    cudaStreamSynchronize(stream);
+    return 0;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int reverse_order_scalars_batch_cuda_bls12_377(
+  BLS12_377::scalar_t* arr, int n, int batch_size, size_t device_id = 0, cudaStream_t stream = 0)
+{
+  try {
+    uint32_t logn = uint32_t(log(n) / log(2));
+    cudaStreamCreate(&stream);
+    reverse_order_batch(arr, n, logn, batch_size, stream);
+    return 0;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int
+reverse_order_points_cuda_bls12_377(BLS12_377::projective_t* arr, int n, size_t device_id = 0, cudaStream_t stream = 0)
+{
+  try {
+    uint32_t logn = uint32_t(log(n) / log(2));
+    cudaStreamCreate(&stream);
+    reverse_order(arr, n, logn, stream);
+    return 0;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int reverse_order_points_batch_cuda_bls12_377(
+  BLS12_377::projective_t* arr, int n, int batch_size, size_t device_id = 0, cudaStream_t stream = 0)
+{
+  try {
+    uint32_t logn = uint32_t(log(n) / log(2));
+    cudaStreamCreate(&stream);
+    reverse_order_batch(arr, n, logn, batch_size, stream);
+    return 0;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+#endif

icicle/curves/bls12_377/msm.cu

@@ -1,40 +1,47 @@
 #ifndef _BLS12_377_MSM
 #define _BLS12_377_MSM
 #include "../../appUtils/msm/msm.cu"
-#include <stdexcept>
-#include <cuda.h>
 #include "curve_config.cuh"
+#include <cuda.h>
+#include <stdexcept>
 
-
-extern "C"
-int msm_cuda_bls12_377(BLS12_377::projective_t *out, BLS12_377::affine_t points[],
-              BLS12_377::scalar_t scalars[], size_t count, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int msm_cuda_bls12_377(
+  BLS12_377::projective_t* out,
+  BLS12_377::affine_t points[],
+  BLS12_377::scalar_t scalars[],
+  size_t count,
+  unsigned large_bucket_factor,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-    try
-    {
-        large_msm<BLS12_377::scalar_t, BLS12_377::projective_t, BLS12_377::affine_t>(scalars, points, count, out, false, stream);
-        return CUDA_SUCCESS;
-    }
-    catch (const std::runtime_error &ex)
-    {
-        printf("error %s", ex.what());
-        return -1;
-    }
+  try {
+    cudaStreamCreate(&stream);
+    large_msm<BLS12_377::scalar_t, BLS12_377::projective_t, BLS12_377::affine_t>(
+      scalars, points, count, out, false, false, large_bucket_factor, stream);
+    cudaStreamSynchronize(stream);
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
 }
 
-extern "C" int msm_batch_cuda_bls12_377(BLS12_377::projective_t* out, BLS12_377::affine_t points[],
-                              BLS12_377::scalar_t scalars[], size_t batch_size, size_t msm_size, size_t device_id = 0, cudaStream_t stream = 0)
+extern "C" int msm_batch_cuda_bls12_377(
+  BLS12_377::projective_t* out,
+  BLS12_377::affine_t points[],
+  BLS12_377::scalar_t scalars[],
+  size_t batch_size,
+  size_t msm_size,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
 {
-  try
-  {
+  try {
     cudaStreamCreate(&stream);
-    batched_large_msm<BLS12_377::scalar_t, BLS12_377::projective_t, BLS12_377::affine_t>(scalars, points, batch_size, msm_size, out, false, stream);
+    batched_large_msm<BLS12_377::scalar_t, BLS12_377::projective_t, BLS12_377::affine_t>(
+      scalars, points, batch_size, msm_size, out, false, stream);
     cudaStreamSynchronize(stream);
-
     return CUDA_SUCCESS;
-  }
-  catch (const std::runtime_error &ex)
-  {
+  } catch (const std::runtime_error& ex) {
     printf("error %s", ex.what());
     return -1;
   }
@@ -42,52 +49,168 @@ extern "C" int msm_batch_cuda_bls12_377(BLS12_377::projective_t* out, BLS12_377:
 
 /**
  * Commit to a polynomial using the MSM.
- * Note: this function just calls the MSM, it doesn't convert between evaluation and coefficient form of scalars or points.
+ * Note: this function just calls the MSM, it doesn't convert between evaluation and coefficient form of scalars or
+ * points.
  * @param d_out Ouptut point to write the result to.
  * @param d_scalars Scalars for the MSM. Must be on device.
  * @param d_points Points for the MSM. Must be on device.
  * @param count Length of `d_scalars` and `d_points` arrays (they should have equal length).
  */
- extern "C"
- int commit_cuda_bls12_377(BLS12_377::projective_t* d_out, BLS12_377::scalar_t* d_scalars, BLS12_377::affine_t* d_points, size_t count, size_t device_id = 0, cudaStream_t stream = 0)
- {
-     try
-     {
-         large_msm<BLS12_377::scalar_t, BLS12_377::projective_t, BLS12_377::affine_t>(d_scalars, d_points, count, d_out, true, stream);
-         cudaStreamSynchronize(stream);
-         return 0;
-     }
-     catch (const std::runtime_error &ex)
-     {
-         printf("error %s", ex.what());
-         return -1;
-     }
- }
- 
- /**
-  * Commit to a batch of polynomials using the MSM.
-  * Note: this function just calls the MSM, it doesn't convert between evaluation and coefficient form of scalars or points.
-  * @param d_out Ouptut point to write the results to.
-  * @param d_scalars Scalars for the MSMs of all polynomials. Must be on device.
-  * @param d_points Points for the MSMs. Must be on device. It is assumed that this set of bases is used for each MSM.
-  * @param count Length of `d_points` array, `d_scalar` has length `count` * `batch_size`.
-  * @param batch_size Size of the batch.
-  */
- extern "C"
- int commit_batch_cuda_bls12_377(BLS12_377::projective_t* d_out, BLS12_377::scalar_t* d_scalars, BLS12_377::affine_t* d_points, size_t count, size_t batch_size, size_t device_id = 0, cudaStream_t stream = 0)
- {
-     try
-     {
-        cudaStreamCreate(&stream);
-         batched_large_msm(d_scalars, d_points, batch_size, count, d_out, true, stream);
-         cudaStreamSynchronize(stream);
-         return 0;
-     }
-     catch (const std::runtime_error &ex)
-     {
-         printf("error %s", ex.what());
-         return -1;
-     }
- }
+extern "C" int commit_cuda_bls12_377(
+  BLS12_377::projective_t* d_out,
+  BLS12_377::scalar_t* d_scalars,
+  BLS12_377::affine_t* d_points,
+  size_t count,
+  unsigned large_bucket_factor,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    large_msm(d_scalars, d_points, count, d_out, true, false, large_bucket_factor, stream);
+    cudaStreamSynchronize(stream);
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
 
+/**
+ * Commit to a batch of polynomials using the MSM.
+ * Note: this function just calls the MSM, it doesn't convert between evaluation and coefficient form of scalars or
+ * points.
+ * @param d_out Ouptut point to write the results to.
+ * @param d_scalars Scalars for the MSMs of all polynomials. Must be on device.
+ * @param d_points Points for the MSMs. Must be on device. It is assumed that this set of bases is used for each MSM.
+ * @param count Length of `d_points` array, `d_scalar` has length `count` * `batch_size`.
+ * @param batch_size Size of the batch.
+ */
+extern "C" int commit_batch_cuda_bls12_377(
+  BLS12_377::projective_t* d_out,
+  BLS12_377::scalar_t* d_scalars,
+  BLS12_377::affine_t* d_points,
+  size_t count,
+  size_t batch_size,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    batched_large_msm(d_scalars, d_points, batch_size, count, d_out, true, stream);
+    cudaStreamSynchronize(stream);
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+#if defined(G2_DEFINED)
+extern "C" int msm_g2_cuda_bls12_377(
+  BLS12_377::g2_projective_t* out,
+  BLS12_377::g2_affine_t points[],
+  BLS12_377::scalar_t scalars[],
+  size_t count,
+  unsigned large_bucket_factor,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    large_msm<BLS12_377::scalar_t, BLS12_377::g2_projective_t, BLS12_377::g2_affine_t>(
+      scalars, points, count, out, false, false, large_bucket_factor, stream);
+    cudaStreamSynchronize(stream);
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int msm_batch_g2_cuda_bls12_377(
+  BLS12_377::g2_projective_t* out,
+  BLS12_377::g2_affine_t points[],
+  BLS12_377::scalar_t scalars[],
+  size_t batch_size,
+  size_t msm_size,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
+{
+  try {
+    cudaStreamCreate(&stream);
+    batched_large_msm<BLS12_377::scalar_t, BLS12_377::g2_projective_t, BLS12_377::g2_affine_t>(
+      scalars, points, batch_size, msm_size, out, false, stream);
+    cudaStreamSynchronize(stream);
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+/**
+ * Commit to a polynomial using the MSM in G2 group.
+ * Note: this function just calls the MSM, it doesn't convert between evaluation and coefficient form of scalars or
+ * points.
+ * @param d_out Ouptut G2 point to write the result to.
+ * @param d_scalars Scalars for the MSM. Must be on device.
+ * @param d_points G2 affine points for the MSM. Must be on device.
+ * @param count Length of `d_scalars` and `d_points` arrays (they should have equal length).
+ */
+extern "C" int commit_g2_cuda_bls12_377(
+  BLS12_377::g2_projective_t* d_out,
+  BLS12_377::scalar_t* d_scalars,
+  BLS12_377::g2_affine_t* d_points,
+  size_t count,
+  unsigned large_bucket_factor,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
+{
+  // TODO: use device_id when working with multiple devices
+  (void)device_id;
+  try {
+    cudaStreamCreate(&stream);
+    large_msm(d_scalars, d_points, count, d_out, true, false, large_bucket_factor, stream);
+    cudaStreamSynchronize(stream);
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+/**
+ * Commit to a batch of polynomials using the MSM.
+ * Note: this function just calls the MSM, it doesn't convert between evaluation and coefficient form of scalars or
+ * points.
+ * @param d_out Ouptut G2 point to write the results to.
+ * @param d_scalars Scalars for the MSMs of all polynomials. Must be on device.
+ * @param d_points G2 affine points for the MSMs. Must be on device. It is assumed that this set of bases is used for
+ * each MSM.
+ * @param count Length of `d_points` array, `d_scalar` has length `count` * `batch_size`.
+ * @param batch_size Size of the batch.
+ */
+extern "C" int commit_batch_g2_cuda_bls12_377(
+  BLS12_377::g2_projective_t* d_out,
+  BLS12_377::scalar_t* d_scalars,
+  BLS12_377::g2_affine_t* d_points,
+  size_t count,
+  size_t batch_size,
+  size_t device_id = 0,
+  cudaStream_t stream = 0)
+{
+  // TODO: use device_id when working with multiple devices
+  (void)device_id;
+  try {
+    cudaStreamCreate(&stream);
+    batched_large_msm(d_scalars, d_points, batch_size, count, d_out, true, stream);
+    cudaStreamSynchronize(stream);
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+#endif
 #endif

icicle/curves/bls12_377/params.cuh

@@ -1,180 +1,329 @@
 #pragma once
 #include "../../utils/storage.cuh"
 
-namespace PARAMS_BLS12_377{
-  struct fp_config{
+namespace PARAMS_BLS12_377 {
+  struct fp_config {
     static constexpr unsigned limbs_count = 8;
     static constexpr unsigned omegas_count = 32;
 
-    static constexpr storage<limbs_count> modulus = {0x00000001, 0x0a118000, 0xd0000001, 0x59aa76fe, 0x5c37b001, 0x60b44d1e, 0x9a2ca556, 0x12ab655e};
-    static constexpr storage<limbs_count> modulus_2 = {0x00000002, 0x14230000, 0xa0000002, 0xb354edfd, 0xb86f6002, 0xc1689a3c, 0x34594aac, 0x2556cabd};
-    static constexpr storage<limbs_count> modulus_4 = {0x00000004, 0x28460000, 0x40000004, 0x66a9dbfb, 0x70dec005, 0x82d13479, 0x68b29559, 0x4aad957a};
-    static constexpr storage<2*limbs_count> modulus_wide = {0x00000001, 0x0a118000, 0xd0000001, 0x59aa76fe, 0x5c37b001, 0x60b44d1e, 0x9a2ca556, 0x12ab655e, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
-    static constexpr storage<2*limbs_count> modulus_squared = {0x00000001, 0x14230000, 0xe0000002, 0xc7dd4d2f, 0x8585d003, 0x08ee1bd4, 0xe57fc56e, 0x7e7557e3, 0x483a709d, 0x1fdebb41, 0x5678f4e6, 0x8ea77334, 0xc19c3ec5, 0xd717de29, 0xe2340781, 0x015c8d01};
-    static constexpr storage<2*limbs_count> modulus_squared_2 = {0x00000002, 0x28460000, 0xc0000004, 0x8fba9a5f, 0x0b0ba007, 0x11dc37a9, 0xcaff8adc, 0xfceaafc7, 0x9074e13a, 0x3fbd7682, 0xacf1e9cc, 0x1d4ee668, 0x83387d8b, 0xae2fbc53, 0xc4680f03, 0x02b91a03};
-    static constexpr storage<2*limbs_count> modulus_squared_4 = {0x00000004, 0x508c0000, 0x80000008, 0x1f7534bf, 0x1617400f, 0x23b86f52, 0x95ff15b8, 0xf9d55f8f, 0x20e9c275, 0x7f7aed05, 0x59e3d398, 0x3a9dccd1, 0x0670fb16, 0x5c5f78a7, 0x88d01e07, 0x05723407};
+    static constexpr storage<limbs_count> modulus = {0x00000001, 0x0a118000, 0xd0000001, 0x59aa76fe,
+                                                     0x5c37b001, 0x60b44d1e, 0x9a2ca556, 0x12ab655e};
+    static constexpr storage<limbs_count> modulus_2 = {0x00000002, 0x14230000, 0xa0000002, 0xb354edfd,
+                                                       0xb86f6002, 0xc1689a3c, 0x34594aac, 0x2556cabd};
+    static constexpr storage<limbs_count> modulus_4 = {0x00000004, 0x28460000, 0x40000004, 0x66a9dbfb,
+                                                       0x70dec005, 0x82d13479, 0x68b29559, 0x4aad957a};
+    static constexpr storage<2 * limbs_count> modulus_wide = {
+      0x00000001, 0x0a118000, 0xd0000001, 0x59aa76fe, 0x5c37b001, 0x60b44d1e, 0x9a2ca556, 0x12ab655e,
+      0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
+    static constexpr storage<2 * limbs_count> modulus_squared = {
+      0x00000001, 0x14230000, 0xe0000002, 0xc7dd4d2f, 0x8585d003, 0x08ee1bd4, 0xe57fc56e, 0x7e7557e3,
+      0x483a709d, 0x1fdebb41, 0x5678f4e6, 0x8ea77334, 0xc19c3ec5, 0xd717de29, 0xe2340781, 0x015c8d01};
+    static constexpr storage<2 * limbs_count> modulus_squared_2 = {
+      0x00000002, 0x28460000, 0xc0000004, 0x8fba9a5f, 0x0b0ba007, 0x11dc37a9, 0xcaff8adc, 0xfceaafc7,
+      0x9074e13a, 0x3fbd7682, 0xacf1e9cc, 0x1d4ee668, 0x83387d8b, 0xae2fbc53, 0xc4680f03, 0x02b91a03};
+    static constexpr storage<2 * limbs_count> modulus_squared_4 = {
+      0x00000004, 0x508c0000, 0x80000008, 0x1f7534bf, 0x1617400f, 0x23b86f52, 0x95ff15b8, 0xf9d55f8f,
+      0x20e9c275, 0x7f7aed05, 0x59e3d398, 0x3a9dccd1, 0x0670fb16, 0x5c5f78a7, 0x88d01e07, 0x05723407};
     static constexpr unsigned modulus_bit_count = 253;
-    static constexpr storage<limbs_count> m = {0x151e79ea, 0xf5204c21, 0x8d69e258, 0xfd0a180b, 0xfaa80548, 0xe4e51e49, 0xc40b2c9e, 0x36d9491e};
-    static constexpr storage<limbs_count> one = {0x00000001, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
-    static constexpr storage<limbs_count> zero = {0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
+    static constexpr storage<limbs_count> m = {0x151e79ea, 0xf5204c21, 0x8d69e258, 0xfd0a180b,
+                                               0xfaa80548, 0xe4e51e49, 0xc40b2c9e, 0x36d9491e};
+    static constexpr storage<limbs_count> one = {0x00000001, 0x00000000, 0x00000000, 0x00000000,
+                                                 0x00000000, 0x00000000, 0x00000000, 0x00000000};
+    static constexpr storage<limbs_count> zero = {0x00000000, 0x00000000, 0x00000000, 0x00000000,
+                                                  0x00000000, 0x00000000, 0x00000000, 0x00000000};
+    static constexpr storage<limbs_count> montgomery_r = {0xfffffff3, 0x7d1c7fff, 0x6ffffff2, 0x7257f50f,
+                                                          0x512c0fee, 0x16d81575, 0x2bbb9a9d, 0x0d4bda32};
+    static constexpr storage<limbs_count> montgomery_r_inv = {0x1beeec02, 0x4122dd1a, 0x74fee875, 0xbd1eae95,
+                                                              0x27b28e2f, 0x838557e2, 0x2290c02c, 0x07b30191};
 
-    static constexpr storage<limbs_count> omega1= {0x00000001, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
-    static constexpr storage<limbs_count> omega2= {0x00000001, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
-    static constexpr storage<limbs_count> omega3= {0x00000000, 0x0a118000, 0xd0000001, 0x59aa76fe, 0x5c37b001, 0x60b44d1e, 0x9a2ca556, 0x12ab655e};
-    static constexpr storage<limbs_count> omega4= {0x00000001, 0x8f1a4000, 0xb0000001, 0xcf664765, 0x970dec00, 0x23ed1347, 0x00000000, 0x00000000};
-    static constexpr storage<limbs_count> omega5= {0x0405f600, 0xfa8e7081, 0xf8a89660, 0x38b1c291, 0x6bda5fce, 0xefab9005, 0x92a3c754, 0x0b6b0756};
-    static constexpr storage<limbs_count> omega6= {0xaf0a50c8, 0xc5b2c78e, 0x4636deb3, 0x72e32a34, 0xb6f97778, 0x3d775d15, 0x2b16be6e, 0x0c4c070d};
-    static constexpr storage<limbs_count> omega7= {0x7a1ade2c, 0x3f5a4e73, 0x0120d1db, 0x71e5bca1, 0x3b2866fd, 0xbcb44162, 0x89c38db1, 0x06ed1a90};
-    static constexpr storage<limbs_count> omega8= {0xbd2cd25e, 0x61c5510e, 0x2b0d531c, 0xe2d70111, 0x94c3bd4b, 0x738f9894, 0x53182695, 0x0b1e0f1d};
-    static constexpr storage<limbs_count> omega9= {0x8cb9508c, 0xcfb2f75e, 0xf491e401, 0x4c14f244, 0x23c16afb, 0xc8f5265f, 0x70f3ff2a, 0x0cda7e27};
-    static constexpr storage<limbs_count> omega10= {0x0bdc32ee, 0xca77feb9, 0xd957f5a9, 0xf36ddfd4, 0x61ba14c4, 0x491c58f5, 0x93e8f339, 0x0618d3c9};
-    static constexpr storage<limbs_count> omega11= {0x2d89d82f, 0x68c3242e, 0x832a3729, 0xf9559645, 0xbceb62cc, 0x5c803c5e, 0x99ffa2f8, 0x1177cf5d};
-    static constexpr storage<limbs_count> omega12= {0x6932851a, 0xb6ed40f2, 0x1e0da12e, 0x79cbe7fb, 0x2a7d8f87, 0x8d408575, 0x7505d049, 0x11867341};
-    static constexpr storage<limbs_count> omega13= {0x07146cbf, 0x8cf7d87a, 0x109c4d23, 0x14ac37dc, 0x883e9660, 0x082d15f0, 0xad9ea9b8, 0x003719b1};
-    static constexpr storage<limbs_count> omega14= {0xfd0aee77, 0x2260e0dd, 0x1e33b6db, 0xc0cbbc3f, 0xfe7e1b36, 0xc8bf6747, 0x4cb802c1, 0x129e4fd5};
-    static constexpr storage<limbs_count> omega15= {0x8ac75741, 0x22f6fca2, 0xdd37b519, 0x8101b557, 0x1036226a, 0xf493bb8a, 0xfce05c2c, 0x06dbad6c};
-    static constexpr storage<limbs_count> omega16= {0x56733f8b, 0x7d246c24, 0xff70b46a, 0xbc3c4112, 0x6f13530b, 0x2c159b40, 0xc55d287b, 0x0c13137a};
-    static constexpr storage<limbs_count> omega17= {0xec8af73d, 0x8d24de3c, 0xcf722b45, 0x50f778d4, 0x15bc7dd7, 0xf4506bc3, 0xf94a16e1, 0x0e43ba91};
-    static constexpr storage<limbs_count> omega18= {0xd4405b8f, 0x0baa7b44, 0xee0f1394, 0xf8f3c7fe, 0xef0dfe6d, 0x46b153c0, 0x2dde6b95, 0x0ea2bcd9};
-    static constexpr storage<limbs_count> omega19= {0x3d1fa34e, 0x5f4dc975, 0x15af81db, 0xc28e54ee, 0x04947d99, 0x83d9a55f, 0x54a2b488, 0x08ec7ccf};
-    static constexpr storage<limbs_count> omega20= {0x0cac0ee8, 0x0d8fa7b3, 0x82ef38e4, 0x756284ed, 0xac8f90d2, 0x7014b194, 0x634e5d50, 0x092488f8};
-    static constexpr storage<limbs_count> omega21= {0x6d34ed69, 0xd85399bf, 0x09e49cef, 0x4d9012ba, 0xca00ae5d, 0x020142ee, 0x3bdfebfd, 0x12772e57};
-    static constexpr storage<limbs_count> omega22= {0x2eb41723, 0x676c8fc7, 0x5dd895bd, 0xe20380e2, 0x9bf22dde, 0x09dc8be8, 0x42638176, 0x12822f94};
-    static constexpr storage<limbs_count> omega23= {0x81a6d2de, 0x1f1df770, 0xcf29c812, 0x5d33b2da, 0x134f0e7e, 0x1bf162de, 0x1e2877a8, 0x045162c4};
-    static constexpr storage<limbs_count> omega24= {0xfecda1b6, 0x24f4503b, 0xded67d3c, 0x0e5d7ed3, 0x40cf20af, 0x2b7b7e5e, 0x4faad6af, 0x0d472650};
-    static constexpr storage<limbs_count> omega25= {0x584b9eb1, 0xcc6c474c, 0x15a8d886, 0x47670804, 0xbb8654c5, 0x07736d2f, 0xeb207a4b, 0x0d14ce7a};
-    static constexpr storage<limbs_count> omega26= {0xed25924a, 0xd1c6471c, 0x6bc312c3, 0xd98bb374, 0xfeae1a41, 0x50be0848, 0x3265c719, 0x04b07dea};
-    static constexpr storage<limbs_count> omega27= {0x618241e3, 0xab13f73e, 0x166ca902, 0x571c9267, 0x5e828a6d, 0x8586443a, 0x6daba50b, 0x093fdf2f};
-    static constexpr storage<limbs_count> omega28= {0xee11c34f, 0xe688e66b, 0xeacecf5a, 0xdc232eae, 0xb95ae685, 0x4fc35094, 0x7c1d31dc, 0x0273b5bd};
-    static constexpr storage<limbs_count> omega29= {0x1a9057bd, 0x8a8a5a77, 0x41834fbb, 0xdcbfae1d, 0xb34ede6e, 0x534f5b97, 0xb78bbd3e, 0x07313ac5};
-    static constexpr storage<limbs_count> omega30= {0x2be70731, 0x287abbb1, 0x7c35c5aa, 0x5cbcfd1e, 0x1671f4df, 0x7585b3fe, 0xb899c011, 0x08350ecf};
-    static constexpr storage<limbs_count> omega31= {0x09f7c5e2, 0x3400c14e, 0x0a649ea1, 0xc112e60c, 0x067ce95e, 0xf7510758, 0xf9daf17c, 0x040a66a5};
-    static constexpr storage<limbs_count> omega32= {0x43efecd3, 0x89d65957, 0x3bd6c318, 0x29246adc, 0xce01533c, 0xf9fb5ef6, 0x849078c3, 0x020410e4};
+    static constexpr storage<limbs_count> omega1 = {0x00000001, 0x00000000, 0x00000000, 0x00000000,
+                                                    0x00000000, 0x00000000, 0x00000000, 0x00000000};
+    static constexpr storage<limbs_count> omega2 = {0x00000001, 0x00000000, 0x00000000, 0x00000000,
+                                                    0x00000000, 0x00000000, 0x00000000, 0x00000000};
+    static constexpr storage<limbs_count> omega3 = {0x00000000, 0x0a118000, 0xd0000001, 0x59aa76fe,
+                                                    0x5c37b001, 0x60b44d1e, 0x9a2ca556, 0x12ab655e};
+    static constexpr storage<limbs_count> omega4 = {0x00000001, 0x8f1a4000, 0xb0000001, 0xcf664765,
+                                                    0x970dec00, 0x23ed1347, 0x00000000, 0x00000000};
+    static constexpr storage<limbs_count> omega5 = {0x0405f600, 0xfa8e7081, 0xf8a89660, 0x38b1c291,
+                                                    0x6bda5fce, 0xefab9005, 0x92a3c754, 0x0b6b0756};
+    static constexpr storage<limbs_count> omega6 = {0xaf0a50c8, 0xc5b2c78e, 0x4636deb3, 0x72e32a34,
+                                                    0xb6f97778, 0x3d775d15, 0x2b16be6e, 0x0c4c070d};
+    static constexpr storage<limbs_count> omega7 = {0x7a1ade2c, 0x3f5a4e73, 0x0120d1db, 0x71e5bca1,
+                                                    0x3b2866fd, 0xbcb44162, 0x89c38db1, 0x06ed1a90};
+    static constexpr storage<limbs_count> omega8 = {0xbd2cd25e, 0x61c5510e, 0x2b0d531c, 0xe2d70111,
+                                                    0x94c3bd4b, 0x738f9894, 0x53182695, 0x0b1e0f1d};
+    static constexpr storage<limbs_count> omega9 = {0x8cb9508c, 0xcfb2f75e, 0xf491e401, 0x4c14f244,
+                                                    0x23c16afb, 0xc8f5265f, 0x70f3ff2a, 0x0cda7e27};
+    static constexpr storage<limbs_count> omega10 = {0x0bdc32ee, 0xca77feb9, 0xd957f5a9, 0xf36ddfd4,
+                                                     0x61ba14c4, 0x491c58f5, 0x93e8f339, 0x0618d3c9};
+    static constexpr storage<limbs_count> omega11 = {0x2d89d82f, 0x68c3242e, 0x832a3729, 0xf9559645,
+                                                     0xbceb62cc, 0x5c803c5e, 0x99ffa2f8, 0x1177cf5d};
+    static constexpr storage<limbs_count> omega12 = {0x6932851a, 0xb6ed40f2, 0x1e0da12e, 0x79cbe7fb,
+                                                     0x2a7d8f87, 0x8d408575, 0x7505d049, 0x11867341};
+    static constexpr storage<limbs_count> omega13 = {0x07146cbf, 0x8cf7d87a, 0x109c4d23, 0x14ac37dc,
+                                                     0x883e9660, 0x082d15f0, 0xad9ea9b8, 0x003719b1};
+    static constexpr storage<limbs_count> omega14 = {0xfd0aee77, 0x2260e0dd, 0x1e33b6db, 0xc0cbbc3f,
+                                                     0xfe7e1b36, 0xc8bf6747, 0x4cb802c1, 0x129e4fd5};
+    static constexpr storage<limbs_count> omega15 = {0x8ac75741, 0x22f6fca2, 0xdd37b519, 0x8101b557,
+                                                     0x1036226a, 0xf493bb8a, 0xfce05c2c, 0x06dbad6c};
+    static constexpr storage<limbs_count> omega16 = {0x56733f8b, 0x7d246c24, 0xff70b46a, 0xbc3c4112,
+                                                     0x6f13530b, 0x2c159b40, 0xc55d287b, 0x0c13137a};
+    static constexpr storage<limbs_count> omega17 = {0xec8af73d, 0x8d24de3c, 0xcf722b45, 0x50f778d4,
+                                                     0x15bc7dd7, 0xf4506bc3, 0xf94a16e1, 0x0e43ba91};
+    static constexpr storage<limbs_count> omega18 = {0xd4405b8f, 0x0baa7b44, 0xee0f1394, 0xf8f3c7fe,
+                                                     0xef0dfe6d, 0x46b153c0, 0x2dde6b95, 0x0ea2bcd9};
+    static constexpr storage<limbs_count> omega19 = {0x3d1fa34e, 0x5f4dc975, 0x15af81db, 0xc28e54ee,
+                                                     0x04947d99, 0x83d9a55f, 0x54a2b488, 0x08ec7ccf};
+    static constexpr storage<limbs_count> omega20 = {0x0cac0ee8, 0x0d8fa7b3, 0x82ef38e4, 0x756284ed,
+                                                     0xac8f90d2, 0x7014b194, 0x634e5d50, 0x092488f8};
+    static constexpr storage<limbs_count> omega21 = {0x6d34ed69, 0xd85399bf, 0x09e49cef, 0x4d9012ba,
+                                                     0xca00ae5d, 0x020142ee, 0x3bdfebfd, 0x12772e57};
+    static constexpr storage<limbs_count> omega22 = {0x2eb41723, 0x676c8fc7, 0x5dd895bd, 0xe20380e2,
+                                                     0x9bf22dde, 0x09dc8be8, 0x42638176, 0x12822f94};
+    static constexpr storage<limbs_count> omega23 = {0x81a6d2de, 0x1f1df770, 0xcf29c812, 0x5d33b2da,
+                                                     0x134f0e7e, 0x1bf162de, 0x1e2877a8, 0x045162c4};
+    static constexpr storage<limbs_count> omega24 = {0xfecda1b6, 0x24f4503b, 0xded67d3c, 0x0e5d7ed3,
+                                                     0x40cf20af, 0x2b7b7e5e, 0x4faad6af, 0x0d472650};
+    static constexpr storage<limbs_count> omega25 = {0x584b9eb1, 0xcc6c474c, 0x15a8d886, 0x47670804,
+                                                     0xbb8654c5, 0x07736d2f, 0xeb207a4b, 0x0d14ce7a};
+    static constexpr storage<limbs_count> omega26 = {0xed25924a, 0xd1c6471c, 0x6bc312c3, 0xd98bb374,
+                                                     0xfeae1a41, 0x50be0848, 0x3265c719, 0x04b07dea};
+    static constexpr storage<limbs_count> omega27 = {0x618241e3, 0xab13f73e, 0x166ca902, 0x571c9267,
+                                                     0x5e828a6d, 0x8586443a, 0x6daba50b, 0x093fdf2f};
+    static constexpr storage<limbs_count> omega28 = {0xee11c34f, 0xe688e66b, 0xeacecf5a, 0xdc232eae,
+                                                     0xb95ae685, 0x4fc35094, 0x7c1d31dc, 0x0273b5bd};
+    static constexpr storage<limbs_count> omega29 = {0x1a9057bd, 0x8a8a5a77, 0x41834fbb, 0xdcbfae1d,
+                                                     0xb34ede6e, 0x534f5b97, 0xb78bbd3e, 0x07313ac5};
+    static constexpr storage<limbs_count> omega30 = {0x2be70731, 0x287abbb1, 0x7c35c5aa, 0x5cbcfd1e,
+                                                     0x1671f4df, 0x7585b3fe, 0xb899c011, 0x08350ecf};
+    static constexpr storage<limbs_count> omega31 = {0x09f7c5e2, 0x3400c14e, 0x0a649ea1, 0xc112e60c,
+                                                     0x067ce95e, 0xf7510758, 0xf9daf17c, 0x040a66a5};
+    static constexpr storage<limbs_count> omega32 = {0x43efecd3, 0x89d65957, 0x3bd6c318, 0x29246adc,
+                                                     0xce01533c, 0xf9fb5ef6, 0x849078c3, 0x020410e4};
 
     static constexpr storage_array<omegas_count, limbs_count> omega = {
-        omega1, omega2, omega3, omega4, omega5, omega6, omega7, omega8, 
-        omega9, omega10, omega11, omega12, omega13, omega14, omega15, omega16,
-        omega17, omega18, omega19, omega20, omega21, omega22, omega23, omega24,
-        omega25, omega26, omega27, omega28, omega29, omega30, omega31, omega32,
+      omega1,  omega2,  omega3,  omega4,  omega5,  omega6,  omega7,  omega8,  omega9,  omega10, omega11,
+      omega12, omega13, omega14, omega15, omega16, omega17, omega18, omega19, omega20, omega21, omega22,
+      omega23, omega24, omega25, omega26, omega27, omega28, omega29, omega30, omega31, omega32,
     };
 
-    static constexpr storage<limbs_count> omega_inv1= {0x00000001, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
-    static constexpr storage<limbs_count> omega_inv2= {0x00000001, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
-    static constexpr storage<limbs_count> omega_inv3= {0x00000000, 0x0a118000, 0xd0000001, 0x59aa76fe, 0x5c37b001, 0x60b44d1e, 0x9a2ca556, 0x12ab655e};
-    static constexpr storage<limbs_count> omega_inv4= {0x00000000, 0x7af74000, 0x1fffffff, 0x8a442f99, 0xc529c400, 0x3cc739d6, 0x9a2ca556, 0x12ab655e};
-    static constexpr storage<limbs_count> omega_inv5= {0x29f04fbb, 0x401766f3, 0x0a4b98b2, 0x7e4e5f63, 0x9fbc28da, 0x35887f12, 0xdabe3b97, 0x045cb225};
-    static constexpr storage<limbs_count> omega_inv6= {0xac4ce534, 0xf3883827, 0x7c4940f0, 0x9f9a114f, 0x32cc3182, 0xe48527ee, 0x2877f4c2, 0x02d4450c};
-    static constexpr storage<limbs_count> omega_inv7= {0x4afbf0bb, 0xd2533833, 0x1d646d56, 0x20987ba6, 0xb8ae7d61, 0xf2c34c11, 0xb53ae995, 0x09962e74};
-    static constexpr storage<limbs_count> omega_inv8= {0x34f5271a, 0xd6aeb755, 0x493bb125, 0xc0e24cfd, 0x35cf1879, 0xc9d2a1ad, 0x19000e58, 0x0f3570fa};
-    static constexpr storage<limbs_count> omega_inv9= {0xbec3ee61, 0x2601423e, 0xb5252af1, 0x94f5ab4b, 0x205d09ca, 0xa1184628, 0x82a1fba2, 0x0e305e1e};
-    static constexpr storage<limbs_count> omega_inv10= {0x7e3320f2, 0x3cbad3a7, 0x4269c624, 0x7866653a, 0xa2fc13a2, 0xaf6d742d, 0xfe24db2a, 0x03ed8246};
-    static constexpr storage<limbs_count> omega_inv11= {0x30cff7d3, 0xcb6ab09e, 0xd88db7e6, 0x29949e69, 0x24db3cd4, 0xb9117dc6, 0xca8d11b5, 0x01b2aadd};
-    static constexpr storage<limbs_count> omega_inv12= {0x433b851c, 0x1c8fbc5d, 0x545e622f, 0x0ccc3b8c, 0x5c624e0f, 0x0fba9df2, 0x0496ddf9, 0x02d54c5d};
-    static constexpr storage<limbs_count> omega_inv13= {0x0a176838, 0x2ddbbfdd, 0xc4c77f0f, 0xb7a1e4f3, 0x41cad032, 0x645b4383, 0xbfb123c4, 0x0f3fe2e3};
-    static constexpr storage<limbs_count> omega_inv14= {0x9ff30538, 0x1d6d50fe, 0x8576b6fa, 0xca07f2d2, 0x720da6d2, 0x587839fa, 0xe9ebd753, 0x0038d5aa};
-    static constexpr storage<limbs_count> omega_inv15= {0x8e30fb24, 0xaeac713d, 0x21906459, 0xd004e9e3, 0xa60b0a33, 0x2fc54303, 0x14e545a6, 0x039063f8};
-    static constexpr storage<limbs_count> omega_inv16= {0x74d36c47, 0x112559bd, 0x4154b77a, 0x87db7016, 0x3843df80, 0x9e779ae5, 0x297077d0, 0x024424f2};
-    static constexpr storage<limbs_count> omega_inv17= {0x65953c15, 0xd649ae5e, 0x56accc60, 0x879fe571, 0xa3ba1e39, 0xba914f52, 0xd6ea78a2, 0x01b74920};
-    static constexpr storage<limbs_count> omega_inv18= {0x3d8a82b4, 0x319dea45, 0x8fc703de, 0x49468894, 0xc6b00817, 0x703f710f, 0xe862bc53, 0x007762fd};
-    static constexpr storage<limbs_count> omega_inv19= {0x5bae083f, 0x4f433336, 0x27612fe3, 0x485e079c, 0x7f8f0a07, 0xf83b6572, 0xca91a4d4, 0x06bdcaaf};
-    static constexpr storage<limbs_count> omega_inv20= {0xb2fb63eb, 0x4a0bf5e7, 0x996004d9, 0x6f64f8ec, 0x67519c5e, 0x0fecd781, 0x1cab2760, 0x04475eb3};
-    static constexpr storage<limbs_count> omega_inv21= {0xcd83d14f, 0xadbd6ce4, 0x750b194a, 0xc664d3bc, 0x89c9f437, 0x3034dfed, 0xcc2e643b, 0x03d502b8};
-    static constexpr storage<limbs_count> omega_inv22= {0x2272320b, 0xf89478a9, 0xd2e658b7, 0x3adac024, 0x94b25831, 0xf38d840f, 0x37dc6c4c, 0x04540b1f};
-    static constexpr storage<limbs_count> omega_inv23= {0xa6d411fe, 0x19d969b1, 0xf544a648, 0x973f00f7, 0xc9ed9f93, 0xb18f166c, 0xe7f21124, 0x02fba68e};
-    static constexpr storage<limbs_count> omega_inv24= {0x94921227, 0x78b96b20, 0x23b35b65, 0x07cd90db, 0xc843f1c3, 0x111f4fd9, 0xff729f23, 0x0ec4b820};
-    static constexpr storage<limbs_count> omega_inv25= {0x4879d823, 0x53eb200b, 0x93095f4a, 0x1971fac3, 0x86989a58, 0x8467ffe6, 0x306ed29d, 0x0af20231};
-    static constexpr storage<limbs_count> omega_inv26= {0xd4793454, 0x71c907bd, 0x7700defb, 0xc11aa47e, 0xbac11769, 0xf03e0873, 0x97419136, 0x0353190d};
-    static constexpr storage<limbs_count> omega_inv27= {0xa81a701c, 0x61a3deb6, 0x91bbbecf, 0xd8a4eda1, 0x6feb65df, 0x3f5339b1, 0x8b5421f2, 0x108adc5b};
-    static constexpr storage<limbs_count> omega_inv28= {0xe7bf5a41, 0x7d6c573a, 0xfa83b1f7, 0x8038b697, 0xa6718ce9, 0x2a988bee, 0x1239b708, 0x0846f362};
-    static constexpr storage<limbs_count> omega_inv29= {0xe3373548, 0x89a068a4, 0x78a6c4e5, 0xf31284cf, 0x6e9396d6, 0x9eed5c8d, 0x7e4342f9, 0x01643c65};
-    static constexpr storage<limbs_count> omega_inv30= {0x123a81f6, 0xc03a3272, 0x115b15e8, 0x377e6d2f, 0x2d6d7206, 0xed5575e4, 0x714004f2, 0x0b1e37e4};
-    static constexpr storage<limbs_count> omega_inv31= {0xdde8ffc5, 0x62a29589, 0x618c5d62, 0xfb6716e8, 0x88d61f25, 0x787e561c, 0xd2b21c7e, 0x0e351761};
-    static constexpr storage<limbs_count> omega_inv32= {0x7aca7fbe, 0xc9fea0e9, 0xb41a8854, 0x965ff314, 0x810eea7e, 0x743415d4, 0x8275bbd1, 0x0431c01b};
-    
+    static constexpr storage<limbs_count> omega_inv1 = {0x00000001, 0x00000000, 0x00000000, 0x00000000,
+                                                        0x00000000, 0x00000000, 0x00000000, 0x00000000};
+    static constexpr storage<limbs_count> omega_inv2 = {0x00000001, 0x00000000, 0x00000000, 0x00000000,
+                                                        0x00000000, 0x00000000, 0x00000000, 0x00000000};
+    static constexpr storage<limbs_count> omega_inv3 = {0x00000000, 0x0a118000, 0xd0000001, 0x59aa76fe,
+                                                        0x5c37b001, 0x60b44d1e, 0x9a2ca556, 0x12ab655e};
+    static constexpr storage<limbs_count> omega_inv4 = {0x00000000, 0x7af74000, 0x1fffffff, 0x8a442f99,
+                                                        0xc529c400, 0x3cc739d6, 0x9a2ca556, 0x12ab655e};
+    static constexpr storage<limbs_count> omega_inv5 = {0x29f04fbb, 0x401766f3, 0x0a4b98b2, 0x7e4e5f63,
+                                                        0x9fbc28da, 0x35887f12, 0xdabe3b97, 0x045cb225};
+    static constexpr storage<limbs_count> omega_inv6 = {0xac4ce534, 0xf3883827, 0x7c4940f0, 0x9f9a114f,
+                                                        0x32cc3182, 0xe48527ee, 0x2877f4c2, 0x02d4450c};
+    static constexpr storage<limbs_count> omega_inv7 = {0x4afbf0bb, 0xd2533833, 0x1d646d56, 0x20987ba6,
+                                                        0xb8ae7d61, 0xf2c34c11, 0xb53ae995, 0x09962e74};
+    static constexpr storage<limbs_count> omega_inv8 = {0x34f5271a, 0xd6aeb755, 0x493bb125, 0xc0e24cfd,
+                                                        0x35cf1879, 0xc9d2a1ad, 0x19000e58, 0x0f3570fa};
+    static constexpr storage<limbs_count> omega_inv9 = {0xbec3ee61, 0x2601423e, 0xb5252af1, 0x94f5ab4b,
+                                                        0x205d09ca, 0xa1184628, 0x82a1fba2, 0x0e305e1e};
+    static constexpr storage<limbs_count> omega_inv10 = {0x7e3320f2, 0x3cbad3a7, 0x4269c624, 0x7866653a,
+                                                         0xa2fc13a2, 0xaf6d742d, 0xfe24db2a, 0x03ed8246};
+    static constexpr storage<limbs_count> omega_inv11 = {0x30cff7d3, 0xcb6ab09e, 0xd88db7e6, 0x29949e69,
+                                                         0x24db3cd4, 0xb9117dc6, 0xca8d11b5, 0x01b2aadd};
+    static constexpr storage<limbs_count> omega_inv12 = {0x433b851c, 0x1c8fbc5d, 0x545e622f, 0x0ccc3b8c,
+                                                         0x5c624e0f, 0x0fba9df2, 0x0496ddf9, 0x02d54c5d};
+    static constexpr storage<limbs_count> omega_inv13 = {0x0a176838, 0x2ddbbfdd, 0xc4c77f0f, 0xb7a1e4f3,
+                                                         0x41cad032, 0x645b4383, 0xbfb123c4, 0x0f3fe2e3};
+    static constexpr storage<limbs_count> omega_inv14 = {0x9ff30538, 0x1d6d50fe, 0x8576b6fa, 0xca07f2d2,
+                                                         0x720da6d2, 0x587839fa, 0xe9ebd753, 0x0038d5aa};
+    static constexpr storage<limbs_count> omega_inv15 = {0x8e30fb24, 0xaeac713d, 0x21906459, 0xd004e9e3,
+                                                         0xa60b0a33, 0x2fc54303, 0x14e545a6, 0x039063f8};
+    static constexpr storage<limbs_count> omega_inv16 = {0x74d36c47, 0x112559bd, 0x4154b77a, 0x87db7016,
+                                                         0x3843df80, 0x9e779ae5, 0x297077d0, 0x024424f2};
+    static constexpr storage<limbs_count> omega_inv17 = {0x65953c15, 0xd649ae5e, 0x56accc60, 0x879fe571,
+                                                         0xa3ba1e39, 0xba914f52, 0xd6ea78a2, 0x01b74920};
+    static constexpr storage<limbs_count> omega_inv18 = {0x3d8a82b4, 0x319dea45, 0x8fc703de, 0x49468894,
+                                                         0xc6b00817, 0x703f710f, 0xe862bc53, 0x007762fd};
+    static constexpr storage<limbs_count> omega_inv19 = {0x5bae083f, 0x4f433336, 0x27612fe3, 0x485e079c,
+                                                         0x7f8f0a07, 0xf83b6572, 0xca91a4d4, 0x06bdcaaf};
+    static constexpr storage<limbs_count> omega_inv20 = {0xb2fb63eb, 0x4a0bf5e7, 0x996004d9, 0x6f64f8ec,
+                                                         0x67519c5e, 0x0fecd781, 0x1cab2760, 0x04475eb3};
+    static constexpr storage<limbs_count> omega_inv21 = {0xcd83d14f, 0xadbd6ce4, 0x750b194a, 0xc664d3bc,
+                                                         0x89c9f437, 0x3034dfed, 0xcc2e643b, 0x03d502b8};
+    static constexpr storage<limbs_count> omega_inv22 = {0x2272320b, 0xf89478a9, 0xd2e658b7, 0x3adac024,
+                                                         0x94b25831, 0xf38d840f, 0x37dc6c4c, 0x04540b1f};
+    static constexpr storage<limbs_count> omega_inv23 = {0xa6d411fe, 0x19d969b1, 0xf544a648, 0x973f00f7,
+                                                         0xc9ed9f93, 0xb18f166c, 0xe7f21124, 0x02fba68e};
+    static constexpr storage<limbs_count> omega_inv24 = {0x94921227, 0x78b96b20, 0x23b35b65, 0x07cd90db,
+                                                         0xc843f1c3, 0x111f4fd9, 0xff729f23, 0x0ec4b820};
+    static constexpr storage<limbs_count> omega_inv25 = {0x4879d823, 0x53eb200b, 0x93095f4a, 0x1971fac3,
+                                                         0x86989a58, 0x8467ffe6, 0x306ed29d, 0x0af20231};
+    static constexpr storage<limbs_count> omega_inv26 = {0xd4793454, 0x71c907bd, 0x7700defb, 0xc11aa47e,
+                                                         0xbac11769, 0xf03e0873, 0x97419136, 0x0353190d};
+    static constexpr storage<limbs_count> omega_inv27 = {0xa81a701c, 0x61a3deb6, 0x91bbbecf, 0xd8a4eda1,
+                                                         0x6feb65df, 0x3f5339b1, 0x8b5421f2, 0x108adc5b};
+    static constexpr storage<limbs_count> omega_inv28 = {0xe7bf5a41, 0x7d6c573a, 0xfa83b1f7, 0x8038b697,
+                                                         0xa6718ce9, 0x2a988bee, 0x1239b708, 0x0846f362};
+    static constexpr storage<limbs_count> omega_inv29 = {0xe3373548, 0x89a068a4, 0x78a6c4e5, 0xf31284cf,
+                                                         0x6e9396d6, 0x9eed5c8d, 0x7e4342f9, 0x01643c65};
+    static constexpr storage<limbs_count> omega_inv30 = {0x123a81f6, 0xc03a3272, 0x115b15e8, 0x377e6d2f,
+                                                         0x2d6d7206, 0xed5575e4, 0x714004f2, 0x0b1e37e4};
+    static constexpr storage<limbs_count> omega_inv31 = {0xdde8ffc5, 0x62a29589, 0x618c5d62, 0xfb6716e8,
+                                                         0x88d61f25, 0x787e561c, 0xd2b21c7e, 0x0e351761};
+    static constexpr storage<limbs_count> omega_inv32 = {0x7aca7fbe, 0xc9fea0e9, 0xb41a8854, 0x965ff314,
+                                                         0x810eea7e, 0x743415d4, 0x8275bbd1, 0x0431c01b};
+
     static constexpr storage_array<omegas_count, limbs_count> omega_inv = {
-        omega_inv1, omega_inv2, omega_inv3, omega_inv4, omega_inv5, omega_inv6, omega_inv7, omega_inv8, 
-        omega_inv9, omega_inv10, omega_inv11, omega_inv12, omega_inv13, omega_inv14, omega_inv15, omega_inv16,
-        omega_inv17, omega_inv18, omega_inv19, omega_inv20, omega_inv21, omega_inv22, omega_inv23, omega_inv24,
-        omega_inv25, omega_inv26, omega_inv27, omega_inv28, omega_inv29, omega_inv30, omega_inv31, omega_inv32,
+      omega_inv1,  omega_inv2,  omega_inv3,  omega_inv4,  omega_inv5,  omega_inv6,  omega_inv7,  omega_inv8,
+      omega_inv9,  omega_inv10, omega_inv11, omega_inv12, omega_inv13, omega_inv14, omega_inv15, omega_inv16,
+      omega_inv17, omega_inv18, omega_inv19, omega_inv20, omega_inv21, omega_inv22, omega_inv23, omega_inv24,
+      omega_inv25, omega_inv26, omega_inv27, omega_inv28, omega_inv29, omega_inv30, omega_inv31, omega_inv32,
     };
 
-    static constexpr storage<limbs_count> inv1= {0x00000001, 0x8508c000, 0x68000000, 0xacd53b7f, 0x2e1bd800, 0x305a268f, 0x4d1652ab, 0x0955b2af};
-    static constexpr storage<limbs_count> inv2= {0x00000001, 0xc78d2000, 0x1c000000, 0x033fd93f, 0xc529c401, 0xc88739d6, 0xf3a17c00, 0x0e008c06};
-    static constexpr storage<limbs_count> inv3= {0x00000001, 0xe8cf5000, 0xf6000000, 0x2e75281e, 0x90b0ba01, 0x949dc37a, 0xc6e710ab, 0x1055f8b2};
-    static constexpr storage<limbs_count> inv4= {0x00000001, 0xf9706800, 0xe3000000, 0x440fcf8e, 0x76743501, 0xfaa9084c, 0xb089db00, 0x1180af08};
-    static constexpr storage<limbs_count> inv5= {0x00000001, 0x01c0f400, 0xd9800001, 0x4edd2346, 0x6955f281, 0xadaeaab5, 0xa55b402b, 0x12160a33};
-    static constexpr storage<limbs_count> inv6= {0x00000001, 0x05e93a00, 0xd4c00001, 0x5443cd22, 0xe2c6d141, 0x07317be9, 0x1fc3f2c1, 0x1260b7c9};
-    static constexpr storage<limbs_count> inv7= {0x00000001, 0x07fd5d00, 0xd2600001, 0x56f72210, 0x1f7f40a1, 0xb3f2e484, 0xdcf84c0b, 0x12860e93};
-    static constexpr storage<limbs_count> inv8= {0x00000001, 0x09076e80, 0xd1300001, 0x5850cc87, 0x3ddb7851, 0x0a5398d1, 0x3b9278b1, 0x1298b9f9};
-    static constexpr storage<limbs_count> inv9= {0x00000001, 0x098c7740, 0x50980001, 0x58fda1c3, 0xcd099429, 0xb583f2f7, 0xeadf8f03, 0x12a20fab};
-    static constexpr storage<limbs_count> inv10= {0x00000001, 0x09cefba0, 0x104c0001, 0x59540c61, 0x14a0a215, 0x0b1c200b, 0x42861a2d, 0x12a6ba85};
-    static constexpr storage<limbs_count> inv11= {0x00000001, 0x09f03dd0, 0xf0260001, 0x597f41af, 0xb86c290b, 0xb5e83694, 0xee595fc1, 0x12a90ff1};
-    static constexpr storage<limbs_count> inv12= {0x00000001, 0x0a00dee8, 0x60130001, 0x5994dc57, 0x8a51ec86, 0x0b4e41d9, 0x4443028c, 0x12aa3aa8};
-    static constexpr storage<limbs_count> inv13= {0x00000001, 0x0a092f74, 0x18098001, 0xd99fa9ab, 0xf344ce43, 0x3601477b, 0x6f37d3f1, 0x12aad003};
-    static constexpr storage<limbs_count> inv14= {0x00000001, 0x0a0d57ba, 0xf404c001, 0x99a51054, 0x27be3f22, 0xcb5aca4d, 0x04b23ca3, 0x12ab1ab1};
-    static constexpr storage<limbs_count> inv15= {0x00000001, 0x0a0f6bdd, 0xe2026001, 0xf9a7c3a9, 0xc1faf791, 0x16078bb5, 0xcf6f70fd, 0x12ab4007};
-    static constexpr storage<limbs_count> inv16= {0x80000001, 0x0a1075ee, 0x59013001, 0xa9a91d54, 0x0f1953c9, 0xbb5dec6a, 0x34ce0b29, 0x12ab52b3};
-    static constexpr storage<limbs_count> inv17= {0x40000001, 0x0a10faf7, 0x94809801, 0x81a9ca29, 0x35a881e5, 0x0e091cc4, 0xe77d5840, 0x12ab5c08};
-    static constexpr storage<limbs_count> inv18= {0xa0000001, 0x0a113d7b, 0x32404c01, 0x6daa2094, 0x48f018f3, 0x375eb4f1, 0xc0d4fecb, 0x12ab60b3};
-    static constexpr storage<limbs_count> inv19= {0xd0000001, 0x0a115ebd, 0x81202601, 0x63aa4bc9, 0xd293e47a, 0xcc098107, 0x2d80d210, 0x12ab6309};
-    static constexpr storage<limbs_count> inv20= {0xe8000001, 0x0a116f5e, 0x28901301, 0xdeaa6164, 0x1765ca3d, 0x965ee713, 0xe3d6bbb3, 0x12ab6433};
-    static constexpr storage<limbs_count> inv21= {0x74000001, 0x0a1177af, 0x7c480981, 0x9c2a6c31, 0xb9cebd1f, 0xfb899a18, 0x3f01b084, 0x12ab64c9};
-    static constexpr storage<limbs_count> inv22= {0xba000001, 0x0a117bd7, 0x262404c1, 0x7aea7198, 0x8b033690, 0xae1ef39b, 0xec972aed, 0x12ab6513};
-    static constexpr storage<limbs_count> inv23= {0xdd000001, 0x0a117deb, 0x7b120261, 0xea4a744b, 0xf39d7348, 0x0769a05c, 0x4361e822, 0x12ab6539};
-    static constexpr storage<limbs_count> inv24= {0xee800001, 0x0a117ef5, 0x25890131, 0x21fa75a5, 0xa7ea91a5, 0x340ef6bd, 0xeec746bc, 0x12ab654b};
-    static constexpr storage<limbs_count> inv25= {0xf7400001, 0x0a117f7a, 0xfac48099, 0x3dd27651, 0x021120d3, 0x4a61a1ee, 0x4479f609, 0x12ab6555};
-    static constexpr storage<limbs_count> inv26= {0x7ba00001, 0x0a117fbd, 0x6562404d, 0x4bbe76a8, 0x2f24686a, 0xd58af786, 0xef534daf, 0x12ab6559};
-    static constexpr storage<limbs_count> inv27= {0xbdd00001, 0x0a117fde, 0x9ab12027, 0xd2b476d3, 0x45ae0c35, 0x1b1fa252, 0x44bff983, 0x12ab655c};
-    static constexpr storage<limbs_count> inv28= {0x5ee80001, 0x0a117fef, 0x35589014, 0x962f76e9, 0x50f2de1b, 0xbde9f7b8, 0x6f764f6c, 0x12ab655d};
-    static constexpr storage<limbs_count> inv29= {0xaf740001, 0x8a117ff7, 0x02ac480a, 0x77ecf6f4, 0x5695470e, 0x8f4f226b, 0x04d17a61, 0x12ab655e};
-    static constexpr storage<limbs_count> inv30= {0xd7ba0001, 0xca117ffb, 0x69562405, 0xe8cbb6f9, 0xd9667b87, 0xf801b7c4, 0x4f7f0fdb, 0x12ab655e};
-    static constexpr storage<limbs_count> inv31= {0xebdd0001, 0x6a117ffd, 0x1cab1203, 0xa13b16fc, 0x9acf15c4, 0x2c5b0271, 0x74d5da99, 0x12ab655e};
-    static constexpr storage<limbs_count> inv32= {0xf5ee8001, 0x3a117ffe, 0x76558902, 0xfd72c6fd, 0xfb8362e2, 0xc687a7c7, 0x87813ff7, 0x12ab655e};
+    static constexpr storage<limbs_count> inv1 = {0x00000001, 0x8508c000, 0x68000000, 0xacd53b7f,
+                                                  0x2e1bd800, 0x305a268f, 0x4d1652ab, 0x0955b2af};
+    static constexpr storage<limbs_count> inv2 = {0x00000001, 0xc78d2000, 0x1c000000, 0x033fd93f,
+                                                  0xc529c401, 0xc88739d6, 0xf3a17c00, 0x0e008c06};
+    static constexpr storage<limbs_count> inv3 = {0x00000001, 0xe8cf5000, 0xf6000000, 0x2e75281e,
+                                                  0x90b0ba01, 0x949dc37a, 0xc6e710ab, 0x1055f8b2};
+    static constexpr storage<limbs_count> inv4 = {0x00000001, 0xf9706800, 0xe3000000, 0x440fcf8e,
+                                                  0x76743501, 0xfaa9084c, 0xb089db00, 0x1180af08};
+    static constexpr storage<limbs_count> inv5 = {0x00000001, 0x01c0f400, 0xd9800001, 0x4edd2346,
+                                                  0x6955f281, 0xadaeaab5, 0xa55b402b, 0x12160a33};
+    static constexpr storage<limbs_count> inv6 = {0x00000001, 0x05e93a00, 0xd4c00001, 0x5443cd22,
+                                                  0xe2c6d141, 0x07317be9, 0x1fc3f2c1, 0x1260b7c9};
+    static constexpr storage<limbs_count> inv7 = {0x00000001, 0x07fd5d00, 0xd2600001, 0x56f72210,
+                                                  0x1f7f40a1, 0xb3f2e484, 0xdcf84c0b, 0x12860e93};
+    static constexpr storage<limbs_count> inv8 = {0x00000001, 0x09076e80, 0xd1300001, 0x5850cc87,
+                                                  0x3ddb7851, 0x0a5398d1, 0x3b9278b1, 0x1298b9f9};
+    static constexpr storage<limbs_count> inv9 = {0x00000001, 0x098c7740, 0x50980001, 0x58fda1c3,
+                                                  0xcd099429, 0xb583f2f7, 0xeadf8f03, 0x12a20fab};
+    static constexpr storage<limbs_count> inv10 = {0x00000001, 0x09cefba0, 0x104c0001, 0x59540c61,
+                                                   0x14a0a215, 0x0b1c200b, 0x42861a2d, 0x12a6ba85};
+    static constexpr storage<limbs_count> inv11 = {0x00000001, 0x09f03dd0, 0xf0260001, 0x597f41af,
+                                                   0xb86c290b, 0xb5e83694, 0xee595fc1, 0x12a90ff1};
+    static constexpr storage<limbs_count> inv12 = {0x00000001, 0x0a00dee8, 0x60130001, 0x5994dc57,
+                                                   0x8a51ec86, 0x0b4e41d9, 0x4443028c, 0x12aa3aa8};
+    static constexpr storage<limbs_count> inv13 = {0x00000001, 0x0a092f74, 0x18098001, 0xd99fa9ab,
+                                                   0xf344ce43, 0x3601477b, 0x6f37d3f1, 0x12aad003};
+    static constexpr storage<limbs_count> inv14 = {0x00000001, 0x0a0d57ba, 0xf404c001, 0x99a51054,
+                                                   0x27be3f22, 0xcb5aca4d, 0x04b23ca3, 0x12ab1ab1};
+    static constexpr storage<limbs_count> inv15 = {0x00000001, 0x0a0f6bdd, 0xe2026001, 0xf9a7c3a9,
+                                                   0xc1faf791, 0x16078bb5, 0xcf6f70fd, 0x12ab4007};
+    static constexpr storage<limbs_count> inv16 = {0x80000001, 0x0a1075ee, 0x59013001, 0xa9a91d54,
+                                                   0x0f1953c9, 0xbb5dec6a, 0x34ce0b29, 0x12ab52b3};
+    static constexpr storage<limbs_count> inv17 = {0x40000001, 0x0a10faf7, 0x94809801, 0x81a9ca29,
+                                                   0x35a881e5, 0x0e091cc4, 0xe77d5840, 0x12ab5c08};
+    static constexpr storage<limbs_count> inv18 = {0xa0000001, 0x0a113d7b, 0x32404c01, 0x6daa2094,
+                                                   0x48f018f3, 0x375eb4f1, 0xc0d4fecb, 0x12ab60b3};
+    static constexpr storage<limbs_count> inv19 = {0xd0000001, 0x0a115ebd, 0x81202601, 0x63aa4bc9,
+                                                   0xd293e47a, 0xcc098107, 0x2d80d210, 0x12ab6309};
+    static constexpr storage<limbs_count> inv20 = {0xe8000001, 0x0a116f5e, 0x28901301, 0xdeaa6164,
+                                                   0x1765ca3d, 0x965ee713, 0xe3d6bbb3, 0x12ab6433};
+    static constexpr storage<limbs_count> inv21 = {0x74000001, 0x0a1177af, 0x7c480981, 0x9c2a6c31,
+                                                   0xb9cebd1f, 0xfb899a18, 0x3f01b084, 0x12ab64c9};
+    static constexpr storage<limbs_count> inv22 = {0xba000001, 0x0a117bd7, 0x262404c1, 0x7aea7198,
+                                                   0x8b033690, 0xae1ef39b, 0xec972aed, 0x12ab6513};
+    static constexpr storage<limbs_count> inv23 = {0xdd000001, 0x0a117deb, 0x7b120261, 0xea4a744b,
+                                                   0xf39d7348, 0x0769a05c, 0x4361e822, 0x12ab6539};
+    static constexpr storage<limbs_count> inv24 = {0xee800001, 0x0a117ef5, 0x25890131, 0x21fa75a5,
+                                                   0xa7ea91a5, 0x340ef6bd, 0xeec746bc, 0x12ab654b};
+    static constexpr storage<limbs_count> inv25 = {0xf7400001, 0x0a117f7a, 0xfac48099, 0x3dd27651,
+                                                   0x021120d3, 0x4a61a1ee, 0x4479f609, 0x12ab6555};
+    static constexpr storage<limbs_count> inv26 = {0x7ba00001, 0x0a117fbd, 0x6562404d, 0x4bbe76a8,
+                                                   0x2f24686a, 0xd58af786, 0xef534daf, 0x12ab6559};
+    static constexpr storage<limbs_count> inv27 = {0xbdd00001, 0x0a117fde, 0x9ab12027, 0xd2b476d3,
+                                                   0x45ae0c35, 0x1b1fa252, 0x44bff983, 0x12ab655c};
+    static constexpr storage<limbs_count> inv28 = {0x5ee80001, 0x0a117fef, 0x35589014, 0x962f76e9,
+                                                   0x50f2de1b, 0xbde9f7b8, 0x6f764f6c, 0x12ab655d};
+    static constexpr storage<limbs_count> inv29 = {0xaf740001, 0x8a117ff7, 0x02ac480a, 0x77ecf6f4,
+                                                   0x5695470e, 0x8f4f226b, 0x04d17a61, 0x12ab655e};
+    static constexpr storage<limbs_count> inv30 = {0xd7ba0001, 0xca117ffb, 0x69562405, 0xe8cbb6f9,
+                                                   0xd9667b87, 0xf801b7c4, 0x4f7f0fdb, 0x12ab655e};
+    static constexpr storage<limbs_count> inv31 = {0xebdd0001, 0x6a117ffd, 0x1cab1203, 0xa13b16fc,
+                                                   0x9acf15c4, 0x2c5b0271, 0x74d5da99, 0x12ab655e};
+    static constexpr storage<limbs_count> inv32 = {0xf5ee8001, 0x3a117ffe, 0x76558902, 0xfd72c6fd,
+                                                   0xfb8362e2, 0xc687a7c7, 0x87813ff7, 0x12ab655e};
 
     static constexpr storage_array<omegas_count, limbs_count> inv = {
-        inv1, inv2, inv3, inv4, inv5, inv6, inv7, inv8, 
-        inv9, inv10, inv11, inv12, inv13, inv14, inv15, inv16,
-        inv17, inv18, inv19, inv20, inv21, inv22, inv23, inv24,
-        inv25, inv26, inv27, inv28, inv29, inv30, inv31, inv32,
-    };    
+      inv1,  inv2,  inv3,  inv4,  inv5,  inv6,  inv7,  inv8,  inv9,  inv10, inv11, inv12, inv13, inv14, inv15, inv16,
+      inv17, inv18, inv19, inv20, inv21, inv22, inv23, inv24, inv25, inv26, inv27, inv28, inv29, inv30, inv31, inv32,
+    };
   };
 
-  struct fq_config{
+  struct fq_config {
     static constexpr unsigned limbs_count = 12;
-    static constexpr storage<limbs_count> modulus = {0x00000001, 0x8508c000, 0x30000000, 0x170b5d44, 0xba094800, 0x1ef3622f, 0x00f5138f, 0x1a22d9f3, 0x6ca1493b, 0xc63b05c0, 0x17c510ea, 0x01ae3a46};
-    static constexpr storage<limbs_count> modulus_2 = {0x00000002, 0x0a118000, 0x60000001, 0x2e16ba88, 0x74129000, 0x3de6c45f, 0x01ea271e, 0x3445b3e6, 0xd9429276, 0x8c760b80, 0x2f8a21d5, 0x035c748c};
-    static constexpr storage<limbs_count> modulus_4 = {0x00000004, 0x14230000, 0xc0000002, 0x5c2d7510, 0xe8252000, 0x7bcd88be, 0x03d44e3c, 0x688b67cc, 0xb28524ec, 0x18ec1701, 0x5f1443ab, 0x06b8e918};
-    static constexpr storage<2*limbs_count> modulus_wide = {0x00000001, 0x8508c000, 0x30000000, 0x170b5d44, 0xba094800, 0x1ef3622f, 0x00f5138f, 0x1a22d9f3, 0x6ca1493b, 0xc63b05c0, 0x17c510ea, 0x01ae3a46, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
-    static constexpr storage<2*limbs_count> modulus_squared = {0x00000001, 0x0a118000, 0xf0000001, 0x7338d254, 0x2e1bd800, 0x4ada268f, 0x35f1c09a, 0x6bcbfbd2, 0x58638c9d, 0x318324b9, 0x8bb70ae0, 0x460aaaaa, 0x502a4d6c, 0xc014e712, 0xb90660cd, 0x09d018af, 0x3dda4d5c, 0x1f5e7141, 0xa4aee93f, 0x4bb8b87d, 0xb361263c, 0x2256913b, 0xd0bbaffb, 0x0002d307};
-    static constexpr storage<2*limbs_count> modulus_squared_2 = {0x00000002, 0x14230000, 0xe0000002, 0xe671a4a9, 0x5c37b000, 0x95b44d1e, 0x6be38134, 0xd797f7a4, 0xb0c7193a, 0x63064972, 0x176e15c0, 0x8c155555, 0xa0549ad8, 0x8029ce24, 0x720cc19b, 0x13a0315f, 0x7bb49ab8, 0x3ebce282, 0x495dd27e, 0x977170fb, 0x66c24c78, 0x44ad2277, 0xa1775ff6, 0x0005a60f};
-    static constexpr storage<2*limbs_count> modulus_squared_4 = {0x00000004, 0x28460000, 0xc0000004, 0xcce34953, 0xb86f6001, 0x2b689a3c, 0xd7c70269, 0xaf2fef48, 0x618e3275, 0xc60c92e5, 0x2edc2b80, 0x182aaaaa, 0x40a935b1, 0x00539c49, 0xe4198337, 0x274062be, 0xf7693570, 0x7d79c504, 0x92bba4fc, 0x2ee2e1f6, 0xcd8498f1, 0x895a44ee, 0x42eebfec, 0x000b4c1f};
+    static constexpr storage<limbs_count> modulus = {0x00000001, 0x8508c000, 0x30000000, 0x170b5d44,
+                                                     0xba094800, 0x1ef3622f, 0x00f5138f, 0x1a22d9f3,
+                                                     0x6ca1493b, 0xc63b05c0, 0x17c510ea, 0x01ae3a46};
+    static constexpr storage<limbs_count> modulus_2 = {0x00000002, 0x0a118000, 0x60000001, 0x2e16ba88,
+                                                       0x74129000, 0x3de6c45f, 0x01ea271e, 0x3445b3e6,
+                                                       0xd9429276, 0x8c760b80, 0x2f8a21d5, 0x035c748c};
+    static constexpr storage<limbs_count> modulus_4 = {0x00000004, 0x14230000, 0xc0000002, 0x5c2d7510,
+                                                       0xe8252000, 0x7bcd88be, 0x03d44e3c, 0x688b67cc,
+                                                       0xb28524ec, 0x18ec1701, 0x5f1443ab, 0x06b8e918};
+    static constexpr storage<2 * limbs_count> modulus_wide = {
+      0x00000001, 0x8508c000, 0x30000000, 0x170b5d44, 0xba094800, 0x1ef3622f, 0x00f5138f, 0x1a22d9f3,
+      0x6ca1493b, 0xc63b05c0, 0x17c510ea, 0x01ae3a46, 0x00000000, 0x00000000, 0x00000000, 0x00000000,
+      0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
+    static constexpr storage<2 * limbs_count> modulus_squared = {
+      0x00000001, 0x0a118000, 0xf0000001, 0x7338d254, 0x2e1bd800, 0x4ada268f, 0x35f1c09a, 0x6bcbfbd2,
+      0x58638c9d, 0x318324b9, 0x8bb70ae0, 0x460aaaaa, 0x502a4d6c, 0xc014e712, 0xb90660cd, 0x09d018af,
+      0x3dda4d5c, 0x1f5e7141, 0xa4aee93f, 0x4bb8b87d, 0xb361263c, 0x2256913b, 0xd0bbaffb, 0x0002d307};
+    static constexpr storage<2 * limbs_count> modulus_squared_2 = {
+      0x00000002, 0x14230000, 0xe0000002, 0xe671a4a9, 0x5c37b000, 0x95b44d1e, 0x6be38134, 0xd797f7a4,
+      0xb0c7193a, 0x63064972, 0x176e15c0, 0x8c155555, 0xa0549ad8, 0x8029ce24, 0x720cc19b, 0x13a0315f,
+      0x7bb49ab8, 0x3ebce282, 0x495dd27e, 0x977170fb, 0x66c24c78, 0x44ad2277, 0xa1775ff6, 0x0005a60f};
+    static constexpr storage<2 * limbs_count> modulus_squared_4 = {
+      0x00000004, 0x28460000, 0xc0000004, 0xcce34953, 0xb86f6001, 0x2b689a3c, 0xd7c70269, 0xaf2fef48,
+      0x618e3275, 0xc60c92e5, 0x2edc2b80, 0x182aaaaa, 0x40a935b1, 0x00539c49, 0xe4198337, 0x274062be,
+      0xf7693570, 0x7d79c504, 0x92bba4fc, 0x2ee2e1f6, 0xcd8498f1, 0x895a44ee, 0x42eebfec, 0x000b4c1f};
     static constexpr unsigned modulus_bit_count = 377;
-    static constexpr storage<limbs_count> m = {0x5e4daffc, 0x1f9fd58c, 0x89c42a59, 0xd0ed6877, 0xd85a6d02, 0x6af2d488, 0x6776b1a0, 0x3bbad0de, 0x582ef4f7, 0x976c3ca0, 0x0cc4060e, 0x0261508d};
-    static constexpr storage<limbs_count> one = {0x00000001, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
-    static constexpr storage<limbs_count> zero = {0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
+    static constexpr storage<limbs_count> m = {0x5e4daffc, 0x1f9fd58c, 0x89c42a59, 0xd0ed6877, 0xd85a6d02, 0x6af2d488,
+                                               0x6776b1a0, 0x3bbad0de, 0x582ef4f7, 0x976c3ca0, 0x0cc4060e, 0x0261508d};
+    static constexpr storage<limbs_count> one = {0x00000001, 0x00000000, 0x00000000, 0x00000000,
+                                                 0x00000000, 0x00000000, 0x00000000, 0x00000000,
+                                                 0x00000000, 0x00000000, 0x00000000, 0x00000000};
+    static constexpr storage<limbs_count> zero = {0x00000000, 0x00000000, 0x00000000, 0x00000000,
+                                                  0x00000000, 0x00000000, 0x00000000, 0x00000000,
+                                                  0x00000000, 0x00000000, 0x00000000, 0x00000000};
+    static constexpr storage<limbs_count> montgomery_r = {0xffffff,   0xf73fffff, 0xffffff7a, 0xf4a2bbcf,
+                                                          0xf6b7ffe8, 0x0c9dd045, 0x0aec70e1, 0xdd260cff,
+                                                          0x5eb6c4e5, 0xc4fa3f93, 0x3aef1539, 0x51c5b9e8};
+    static constexpr storage<limbs_count> montgomery_r_inv = {0x934f3a1,  0xb0909a28, 0xc1cfac62, 0x3264aa55,
+                                                              0x2a491ae8, 0xaccd49ca, 0xe80e9a61, 0x28b2dce9,
+                                                              0x26f7c08a, 0x4d313ea1, 0x36254563, 0x161de1ee};
     // i^2, the square of the imaginary unit for the extension field
-    static constexpr uint32_t i_squared = 1;
+    static constexpr uint32_t i_squared = 5;
     // true if i^2 is negative
     static constexpr bool i_squared_is_negative = true;
-    // G1 and G2 generators 
-    static constexpr storage<limbs_count> g1_gen_x = {0xb21be9ef, 0xeab9b16e, 0xffcd394e, 0xd5481512, 0xbd37cb5c, 0x188282c8,
-                                                         0xaa9d41bb, 0x85951e2c, 0xbf87ff54, 0xc8fc6225, 0xfe740a67, 0x008848de};
-    static constexpr storage<limbs_count> g1_gen_y = {0x559c8ea6, 0xfd82de55, 0x34a9591a, 0xc2fe3d36, 0x4fb82305, 0x6d182ad4,
-                                                         0xca3e52d9, 0xbd7fb348, 0x30afeec4, 0x1f674f5d, 0xc5102eff, 0x01914a69};
-    static constexpr storage<limbs_count> g2_gen_x_re = {0x7c005196, 0x74e3e48f, 0xbb535402, 0x71889f52, 0x57db6b9b, 0x7ea501f5, 
-                                                            0x203e5031, 0xc565f071, 0xa3841d01, 0xc89630a2, 0x71c785fe, 0x018480be};
-    static constexpr storage<limbs_count> g2_gen_x_im = {0x6ea16afe, 0xb26bfefa, 0xbff76fe6, 0x5cf89984, 0x0799c9de, 0xe7223ece, 
-                                                            0x6651cecb, 0x532777ee, 0xb1b140d5, 0x70dc5a51, 0xe7004031, 0x00ea6040};
-    static constexpr storage<limbs_count> g2_gen_y_re = {0x09fd4ddf, 0xf0940944, 0x6d8c7c2e, 0xf2cf8888, 0xf832d204, 0xe458c282, 
-                                                            0x74b49a58, 0xde03ed72, 0xcbb2efb4, 0xd960736b, 0x5d446f7b, 0x00690d66};
-    static constexpr storage<limbs_count> g2_gen_y_im = {0x85eb8f93, 0xd9a1cdd1, 0x5e52270b, 0x4279b83f, 0xcee304c2, 0x2463b01a,
-                                                            0x3d591bf1, 0x61ef11ac, 0x151a70aa, 0x9e549da3, 0xd2835518, 0x00f8169f};
+    // G1 and G2 generators
+    static constexpr storage<limbs_count> g1_gen_x = {0xb21be9ef, 0xeab9b16e, 0xffcd394e, 0xd5481512,
+                                                      0xbd37cb5c, 0x188282c8, 0xaa9d41bb, 0x85951e2c,
+                                                      0xbf87ff54, 0xc8fc6225, 0xfe740a67, 0x008848de};
+    static constexpr storage<limbs_count> g1_gen_y = {0x559c8ea6, 0xfd82de55, 0x34a9591a, 0xc2fe3d36,
+                                                      0x4fb82305, 0x6d182ad4, 0xca3e52d9, 0xbd7fb348,
+                                                      0x30afeec4, 0x1f674f5d, 0xc5102eff, 0x01914a69};
+    static constexpr storage<limbs_count> g2_gen_x_re = {0x7c005196, 0x74e3e48f, 0xbb535402, 0x71889f52,
+                                                         0x57db6b9b, 0x7ea501f5, 0x203e5031, 0xc565f071,
+                                                         0xa3841d01, 0xc89630a2, 0x71c785fe, 0x018480be};
+    static constexpr storage<limbs_count> g2_gen_x_im = {0x6ea16afe, 0xb26bfefa, 0xbff76fe6, 0x5cf89984,
+                                                         0x0799c9de, 0xe7223ece, 0x6651cecb, 0x532777ee,
+                                                         0xb1b140d5, 0x70dc5a51, 0xe7004031, 0x00ea6040};
+    static constexpr storage<limbs_count> g2_gen_y_re = {0x09fd4ddf, 0xf0940944, 0x6d8c7c2e, 0xf2cf8888,
+                                                         0xf832d204, 0xe458c282, 0x74b49a58, 0xde03ed72,
+                                                         0xcbb2efb4, 0xd960736b, 0x5d446f7b, 0x00690d66};
+    static constexpr storage<limbs_count> g2_gen_y_im = {0x85eb8f93, 0xd9a1cdd1, 0x5e52270b, 0x4279b83f,
+                                                         0xcee304c2, 0x2463b01a, 0x3d591bf1, 0x61ef11ac,
+                                                         0x151a70aa, 0x9e549da3, 0xd2835518, 0x00f8169f};
   };
 
-  static constexpr storage<fq_config::limbs_count> weierstrass_b = {0x00000001, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000,
-                                                                    0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
-  static constexpr storage<fq_config::limbs_count> weierstrass_b_g2_re = {0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000,
-                                                                          0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
-  static constexpr storage<fq_config::limbs_count> weierstrass_b_g2_im = {0x9999999a, 0x1c9ed999, 0x1ccccccd, 0x0dd39e5c, 0x3c6bf800, 0x129207b6,
-                                                                          0xcd5fd889, 0xdc7b4f91, 0x7460c589, 0x43bd0373, 0xdb0fd6f3, 0x010222f6};
-}
+  static constexpr storage<fq_config::limbs_count> weierstrass_b = {0x00000001, 0x00000000, 0x00000000, 0x00000000,
+                                                                    0x00000000, 0x00000000, 0x00000000, 0x00000000,
+                                                                    0x00000000, 0x00000000, 0x00000000, 0x00000000};
+  static constexpr storage<fq_config::limbs_count> weierstrass_b_g2_re = {
+    0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000,
+    0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000};
+  static constexpr storage<fq_config::limbs_count> weierstrass_b_g2_im = {
+    0x9999999a, 0x1c9ed999, 0x1ccccccd, 0x0dd39e5c, 0x3c6bf800, 0x129207b6,
+    0xcd5fd889, 0xdc7b4f91, 0x7460c589, 0x43bd0373, 0xdb0fd6f3, 0x010222f6};
+} // namespace PARAMS_BLS12_377

icicle/curves/bls12_377/projective.cu

@@ -1,22 +1,111 @@
-
+#include "../../primitives/projective.cuh"
+#include "curve_config.cuh"
 #include <cuda.h>
 
-#include "curve_config.cuh"
+extern "C" int random_projective_bls12_377(BLS12_377::projective_t* out) { 
+  try {
+    out[0] = BLS12_377::projective_t::rand_host();
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
 
-#include "../../primitives/projective.cuh"
+extern "C" BLS12_377::projective_t projective_zero_bls12_377() { return BLS12_377::projective_t::zero(); }
 
-extern "C" bool eq_bls12_377(BLS12_377::projective_t *point1, BLS12_377::projective_t *point2)
+extern "C" bool projective_is_on_curve_bls12_377(BLS12_377::projective_t* point1)
 {
-    return (*point1 == *point2) && 
-    !((point1->x == BLS12_377::point_field_t::zero()) && (point1->y == BLS12_377::point_field_t::zero()) && (point1->z == BLS12_377::point_field_t::zero())) && 
-    !((point2->x == BLS12_377::point_field_t::zero()) && (point2->y == BLS12_377::point_field_t::zero()) && (point2->z == BLS12_377::point_field_t::zero()));
+  return BLS12_377::projective_t::is_on_curve(*point1);
+}
+
+extern "C" int projective_to_affine_bls12_377(BLS12_377::affine_t* out, BLS12_377::projective_t* point1)
+{
+  try {
+    out[0] = BLS12_377::projective_t::to_affine(*point1);
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int projective_from_affine_bls12_377(BLS12_377::projective_t* out, BLS12_377::affine_t* point1)
+{
+  try {
+    out[0] = BLS12_377::projective_t::from_affine(*point1);
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int random_scalar_bls12_377(BLS12_377::scalar_field_t* out) { 
+  try {
+    out[0] = BLS12_377::scalar_field_t::rand_host();
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" bool eq_bls12_377(BLS12_377::projective_t* point1, BLS12_377::projective_t* point2)
+{
+  return (*point1 == *point2) &&
+         !((point1->x == BLS12_377::point_field_t::zero()) && (point1->y == BLS12_377::point_field_t::zero()) &&
+           (point1->z == BLS12_377::point_field_t::zero())) &&
+         !((point2->x == BLS12_377::point_field_t::zero()) && (point2->y == BLS12_377::point_field_t::zero()) &&
+           (point2->z == BLS12_377::point_field_t::zero()));
 }
 
 #if defined(G2_DEFINED)
-extern "C" bool eq_g2_bls12_377(BLS12_377::g2_projective_t *point1, BLS12_377::g2_projective_t *point2)
+extern "C" bool eq_g2_bls12_377(BLS12_377::g2_projective_t* point1, BLS12_377::g2_projective_t* point2)
 {
-  return (*point1 == *point2) && 
-  !((point1->x == BLS12_377::g2_point_field_t::zero()) && (point1->y == BLS12_377::g2_point_field_t::zero()) && (point1->z == BLS12_377::g2_point_field_t::zero())) && 
-  !((point2->x == BLS12_377::g2_point_field_t::zero()) && (point2->y == BLS12_377::g2_point_field_t::zero()) && (point2->z == BLS12_377::g2_point_field_t::zero()));
+  return (*point1 == *point2) &&
+         !((point1->x == BLS12_377::g2_point_field_t::zero()) && (point1->y == BLS12_377::g2_point_field_t::zero()) &&
+           (point1->z == BLS12_377::g2_point_field_t::zero())) &&
+         !((point2->x == BLS12_377::g2_point_field_t::zero()) && (point2->y == BLS12_377::g2_point_field_t::zero()) &&
+           (point2->z == BLS12_377::g2_point_field_t::zero()));
 }
+
+extern "C" int random_g2_projective_bls12_377(BLS12_377::g2_projective_t* out) 
+{ 
+  try {
+    out[0] = BLS12_377::g2_projective_t::rand_host();
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int g2_projective_to_affine_bls12_377(BLS12_377::g2_affine_t* out, BLS12_377::g2_projective_t* point1)
+{
+  try {
+    out[0] = BLS12_377::g2_projective_t::to_affine(*point1);
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" int g2_projective_from_affine_bls12_377(BLS12_377::g2_projective_t* out, BLS12_377::g2_affine_t* point1)
+{
+  try {
+    out[0] = BLS12_377::g2_projective_t::from_affine(*point1);
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what());
+    return -1;
+  }
+}
+
+extern "C" bool g2_projective_is_on_curve_bls12_377(BLS12_377::g2_projective_t* point1)
+{
+  return BLS12_377::g2_projective_t::is_on_curve(*point1);
+}
+
 #endif

icicle/curves/bls12_377/supported_operations.cu

@@ -1,5 +1,4 @@
-
-#include "projective.cu"
 #include "lde.cu"
 #include "msm.cu"
-#include "ve_mod_mult.cu"
+#include "projective.cu"
+#include "ve_mod_mult.cu"

icicle/curves/bls12_377/ve_mod_mult.cu

@@ -1,69 +1,80 @@
 #ifndef _BLS12_377_VEC_MULT
 #define _BLS12_377_VEC_MULT
-#include <stdio.h>
-#include <iostream>
-#include "../../primitives/field.cuh"
-#include "../../utils/storage.cuh"
-#include "../../primitives/projective.cuh"
-#include "curve_config.cuh"
 #include "../../appUtils/vector_manipulation/ve_mod_mult.cuh"
+#include "../../primitives/field.cuh"
+#include "../../primitives/projective.cuh"
+#include "../../utils/storage.cuh"
+#include "curve_config.cuh"
+#include <iostream>
+#include <stdio.h>
 
-
-extern "C" int32_t vec_mod_mult_point_bls12_377(BLS12_377::projective_t *inout,
-                                      BLS12_377::scalar_t *scalar_vec,
-                                      size_t n_elments,
-                                      size_t device_id, 
-                                      cudaStream_t stream = 0)
+extern "C" int32_t vec_mod_mult_point_bls12_377(
+  BLS12_377::projective_t* inout,
+  BLS12_377::scalar_t* scalar_vec,
+  size_t n_elments,
+  size_t device_id,
+  cudaStream_t stream = 0)
 {
-  try
-  {
+  // TODO: use device_id when working with multiple devices
+  (void)device_id;
+  try {
     // TODO: device_id
     vector_mod_mult<BLS12_377::projective_t, BLS12_377::scalar_t>(scalar_vec, inout, inout, n_elments, stream);
     return CUDA_SUCCESS;
-  }
-  catch (const std::runtime_error &ex)
-  {
+  } catch (const std::runtime_error& ex) {
     printf("error %s", ex.what()); // TODO: error code and message
     return -1;
   }
 }
 
-extern "C" int32_t vec_mod_mult_scalar_bls12_377(BLS12_377::scalar_t *inout,
-                                       BLS12_377::scalar_t *scalar_vec,
-                                       size_t n_elments,
-                                       size_t device_id,
-                                       cudaStream_t stream = 0)
+extern "C" int32_t vec_mod_mult_scalar_bls12_377(
+  BLS12_377::scalar_t* inout,
+  BLS12_377::scalar_t* scalar_vec,
+  size_t n_elments,
+  size_t device_id,
+  cudaStream_t stream = 0)
 {
-  try
-  {
+  // TODO: use device_id when working with multiple devices
+  (void)device_id;
+  try {
     // TODO: device_id
     vector_mod_mult<BLS12_377::scalar_t, BLS12_377::scalar_t>(scalar_vec, inout, inout, n_elments, stream);
     return CUDA_SUCCESS;
-  }
-  catch (const std::runtime_error &ex)
-  {
+  } catch (const std::runtime_error& ex) {
     printf("error %s", ex.what()); // TODO: error code and message
     return -1;
   }
 }
 
-extern "C" int32_t matrix_vec_mod_mult_bls12_377(BLS12_377::scalar_t *matrix_flattened,
-                                       BLS12_377::scalar_t *input,
-                                       BLS12_377::scalar_t *output,
-                                       size_t n_elments,
-                                       size_t device_id,
-                                       cudaStream_t stream = 0)
+extern "C" int32_t vec_mod_mult_device_scalar_bls12_377(
+  BLS12_377::scalar_t* inout, BLS12_377::scalar_t* scalar_vec, size_t n_elements, size_t device_id)
 {
-  try
-  {
-    // TODO: device_id
-    matrix_mod_mult<BLS12_377::scalar_t>(matrix_flattened, input, output, n_elments, stream);
+  try {
+    vector_mod_mult_device<BLS12_377::scalar_t, BLS12_377::scalar_t>(scalar_vec, inout, inout, n_elements);
     return CUDA_SUCCESS;
-  }
-  catch (const std::runtime_error &ex)
-  {
+  } catch (const std::runtime_error& ex) {
     printf("error %s", ex.what()); // TODO: error code and message
     return -1;
   }
 }
-#endif
+
+extern "C" int32_t matrix_vec_mod_mult_bls12_377(
+  BLS12_377::scalar_t* matrix_flattened,
+  BLS12_377::scalar_t* input,
+  BLS12_377::scalar_t* output,
+  size_t n_elments,
+  size_t device_id,
+  cudaStream_t stream = 0)
+{
+  // TODO: use device_id when working with multiple devices
+  (void)device_id;
+  try {
+    // TODO: device_id
+    matrix_mod_mult<BLS12_377::scalar_t>(matrix_flattened, input, output, n_elments, stream);
+    return CUDA_SUCCESS;
+  } catch (const std::runtime_error& ex) {
+    printf("error %s", ex.what()); // TODO: error code and message
+    return -1;
+  }
+}
+#endif