ci: update version string in docs

.github/workflows/benchmarks.yml

@@ -8,8 +8,6 @@ on:
 jobs:
 
   bench_poseidon:
-    permissions:
-      contents: read
     runs-on: self-hosted
     steps:
       - uses: actions/checkout@v4
@@ -24,8 +22,6 @@ jobs:
         run: cargo bench --verbose --bench poseidon
 
   bench_einsum_accum_matmul:
-    permissions:
-      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
@@ -41,8 +37,6 @@ jobs:
         run: cargo bench --verbose --bench accum_einsum_matmul
 
   bench_accum_matmul_relu:
-    permissions:
-      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
@@ -58,8 +52,6 @@ jobs:
         run: cargo bench --verbose --bench accum_matmul_relu
 
   bench_accum_matmul_relu_overflow:
-    permissions:
-      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
@@ -75,8 +67,6 @@ jobs:
         run: cargo bench --verbose --bench accum_matmul_relu_overflow
 
   bench_relu:
-    permissions:
-      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
@@ -92,8 +82,6 @@ jobs:
         run: cargo bench --verbose --bench relu
 
   bench_accum_dot:
-    permissions:
-      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
@@ -109,8 +97,6 @@ jobs:
         run: cargo bench --verbose --bench accum_dot
 
   bench_accum_conv:
-    permissions:
-      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
@@ -126,8 +112,6 @@ jobs:
         run: cargo bench --verbose --bench accum_conv
 
   bench_accum_sumpool:
-    permissions:
-      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
@@ -143,8 +127,6 @@ jobs:
         run: cargo bench --verbose --bench accum_sumpool
 
   bench_pairwise_add:
-    permissions:
-      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
@@ -160,8 +142,6 @@ jobs:
         run: cargo bench --verbose --bench pairwise_add
 
   bench_accum_sum:
-    permissions:
-      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:
@@ -177,8 +157,6 @@ jobs:
         run: cargo bench --verbose --bench accum_sum
 
   bench_pairwise_pow:
-    permissions:
-      contents: read
     runs-on: self-hosted
     needs: [bench_poseidon]
     steps:

.github/workflows/engine.yml

@@ -15,9 +15,6 @@ defaults:
     working-directory: .
 jobs:
   publish-wasm-bindings:
-    permissions:
-      contents: read
-      packages: write
     name: publish-wasm-bindings
     runs-on: ubuntu-latest
     if: startsWith(github.ref, 'refs/tags/')
@@ -189,9 +186,6 @@ jobs:
 
 
   in-browser-evm-ver-publish:
-    permissions:
-      contents: read
-      packages: write
     name: publish-in-browser-evm-verifier-package
     needs: [publish-wasm-bindings]
     runs-on: ubuntu-latest

.github/workflows/large-tests.yml

@@ -6,8 +6,6 @@ on:
         description: "Test scenario tags"
 jobs:
   large-tests:
-    permissions:
-      contents: read
     runs-on: kaiju
     steps:
       - uses: actions/checkout@v4

.github/workflows/pypi-gpu.yml

@@ -18,9 +18,6 @@ defaults:
 jobs:
 
   linux:
-    permissions:
-      contents: read
-      packages: write
     runs-on: GPU
     strategy:
       matrix:

.github/workflows/pypi.yml

@@ -16,8 +16,6 @@ defaults:
 
 jobs:
   macos:
-    permissions:
-      contents: read
     runs-on: macos-latest
     if: startsWith(github.ref, 'refs/tags/')
     strategy:
@@ -49,13 +47,6 @@ jobs:
           components: rustfmt, clippy
 
       - name: Build wheels
-        if: matrix.target == 'universal2-apple-darwin'
-        uses: PyO3/maturin-action@v1
-        with:
-          target: ${{ matrix.target }}
-          args: --release --out dist --features python-bindings
-      - name: Build wheels
-        if: matrix.target == 'x86_64'
         uses: PyO3/maturin-action@v1
         with:
           target: ${{ matrix.target }}
@@ -73,8 +64,6 @@ jobs:
           path: dist
 
   windows:
-    permissions:
-      contents: read
     runs-on: windows-latest
     if: startsWith(github.ref, 'refs/tags/')
     strategy:
@@ -122,8 +111,6 @@ jobs:
           path: dist
 
   linux:
-    permissions:
-      contents: read
     runs-on: ubuntu-latest
     if: startsWith(github.ref, 'refs/tags/')
     strategy:
@@ -239,8 +226,6 @@ jobs:
   #       path: dist
 
   musllinux:
-    permissions:
-      contents: read
     runs-on: ubuntu-latest
     if: startsWith(github.ref, 'refs/tags/')
     strategy:
@@ -306,8 +291,6 @@ jobs:
           path: dist
 
   musllinux-cross:
-    permissions:
-      contents: read
     runs-on: ubuntu-latest
     if: startsWith(github.ref, 'refs/tags/')
     strategy:
@@ -395,8 +378,6 @@ jobs:
           packages-dir: ./
 
   doc-publish:
-    permissions:
-      contents: read
     name: Trigger ReadTheDocs Build
     runs-on: ubuntu-latest
     needs: pypi-publish

.github/workflows/release.yml

@@ -10,9 +10,6 @@ on:
       - "*"
 jobs:
   create-release:
-    permissions:
-      contents: read
-      packages: write
     name: create-release
     runs-on: ubuntu-22.04
     if: startsWith(github.ref, 'refs/tags/')
@@ -36,9 +33,6 @@ jobs:
           tag_name: ${{ env.EZKL_VERSION }}
 
   build-release-gpu:
-    permissions:
-      contents: read
-      packages: write
     name: build-release-gpu
     needs: ["create-release"]
     runs-on: GPU
@@ -100,10 +94,6 @@ jobs:
           asset_content_type: application/octet-stream
 
   build-release:
-    permissions:
-      contents: read
-      packages: write
-      issues: write
     name: build-release
     needs: ["create-release"]
     runs-on: ${{ matrix.os }}
@@ -196,18 +186,14 @@ jobs:
           echo "target flag is: ${{ env.TARGET_FLAGS }}"
           echo "target dir is: ${{ env.TARGET_DIR }}"
 
-      - name: Build release binary (no asm or metal) 
-        if: matrix.build != 'linux-gnu' && matrix.build != 'macos-aarch64'
+      - name: Build release binary (no asm)
+        if: matrix.build != 'linux-gnu'
         run: ${{ env.CARGO }} build --release ${{ env.TARGET_FLAGS }} -Z sparse-registry
 
       - name: Build release binary (asm)
         if: matrix.build == 'linux-gnu'
         run: ${{ env.CARGO }} build --release ${{ env.TARGET_FLAGS }} -Z sparse-registry --features asm
 
-      - name: Build release binary (metal)
-        if: matrix.build == 'macos-aarch64'
-        run: ${{ env.CARGO }} build --release ${{ env.TARGET_FLAGS }} -Z sparse-registry --features macos-metal
-
       - name: Strip release binary
         if: matrix.build != 'windows-msvc' && matrix.build != 'linux-aarch64'
         run: strip "target/${{ matrix.target }}/release/ezkl"

.github/workflows/rust.yml

@@ -21,8 +21,6 @@ env:
 jobs:
 
   fr-age-test:
-    permissions:
-      contents: read
     runs-on: large-self-hosted
     steps:
       - uses: actions/checkout@v4
@@ -37,9 +35,8 @@ jobs:
         run: cargo test --release --verbose tests::large_mock_::large_tests_6_expects -- --include-ignored
 
   build:
-    permissions:
-      contents: read
     runs-on: ubuntu-latest
+
     steps:
       - uses: actions/checkout@v4
         with:
@@ -53,8 +50,6 @@ jobs:
         run: cargo build --verbose
 
   docs:
-    permissions:
-      contents: read
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
@@ -69,8 +64,6 @@ jobs:
         run: cargo doc --verbose
 
   library-tests:
-    permissions:
-      contents: read
     runs-on: ubuntu-latest-32-cores
     steps:
       - uses: actions/checkout@v4
@@ -131,8 +124,6 @@ jobs:
   #       run: cargo nextest run conv_relu_col_ultra_overflow --no-capture --features icicle -- --include-ignored
 
   ultra-overflow-tests_og-lookup:
-    permissions:
-      contents: read
     runs-on: non-gpu
     steps:
       - uses: actions/checkout@v4
@@ -168,8 +159,6 @@ jobs:
         run: cargo nextest run  --release conv_relu_col_ultra_overflow --no-capture --no-default-features --features ezkl -- --include-ignored
 
   ultra-overflow-tests:
-    permissions:
-      contents: read
     runs-on: non-gpu
     steps:
       - uses: actions/checkout@v4
@@ -205,8 +194,6 @@ jobs:
         run: cargo nextest run  --release conv_relu_col_ultra_overflow --no-capture -- --include-ignored
 
   model-serialization:
-    permissions:
-      contents: read
     runs-on: ubuntu-latest-16-cores
     steps:
       - uses: actions/checkout@v4
@@ -225,8 +212,6 @@ jobs:
         run: cargo nextest run native_tests::tests::model_serialization_different_binaries_  --test-threads 1
 
   wasm32-tests:
-    permissions:
-      contents: read
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
@@ -254,8 +239,6 @@ jobs:
         run: wasm-pack test --chrome --headless -- -Z build-std="panic_abort,std" --features web
 
   mock-proving-tests:
-    permissions:
-      contents: read
     runs-on: non-gpu
     needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
@@ -321,8 +304,6 @@ jobs:
         run: cargo nextest run --release --verbose tests::mock_fixed_params_ --test-threads 32
 
   prove-and-verify-evm-tests:
-    permissions:
-      contents: read
     runs-on: non-gpu
     needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
@@ -405,44 +386,7 @@ jobs:
       - name: KZG prove and verify tests (EVM + hashed outputs)
         run: cargo nextest run --release --verbose tests_evm::kzg_evm_hashed_output_prove_and_verify --test-threads 1
 
-  # prove-and-verify-tests-metal:
-  #   permissions:
-  #     contents: read
-  #   runs-on: macos-13
-  #   # needs: [build, library-tests, docs]
-  #   steps:
-  #     - uses: actions/checkout@v4
-  #       with:
-  #         persist-credentials: false
-  #     - uses: actions-rs/toolchain@v1
-  #       with:
-  #         toolchain: nightly-2024-07-18
-  #         override: true
-  #         components: rustfmt, clippy
-  #     - uses: jetli/wasm-pack-action@v0.4.0
-  #       with:
-  #         # Pin to version 0.12.1
-  #         version: 'v0.12.1'
-  #     - name: Add rust-src
-  #       run: rustup component add rust-src --toolchain nightly-2024-07-18
-  #     - uses: actions/checkout@v3
-  #       with:
-  #         persist-credentials: false
-  #     - name: Use pnpm 8
-  #       uses: pnpm/action-setup@v2
-  #       with:
-  #         version: 8
-  #     - uses: baptiste0928/cargo-install@v1
-  #       with:
-  #         crate: cargo-nextest
-  #         locked: true
-  #     - name: KZG prove and verify tests (public outputs)
-  #       run: cargo nextest run --release --features macos-metal --verbose tests::kzg_prove_and_verify_::t --no-capture
-
-
   prove-and-verify-tests:
-    permissions:
-      contents: read
     runs-on: non-gpu
     needs: [build, library-tests, docs]
     steps:
@@ -560,8 +504,6 @@ jobs:
   #       run: cargo nextest run --release --verbose tests::kzg_prove_and_verify_hashed --features icicle  --test-threads 1
 
   prove-and-verify-mock-aggr-tests:
-    permissions:
-      contents: read
     runs-on: self-hosted
     needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
@@ -601,8 +543,6 @@ jobs:
   #       run: cargo nextest run --verbose tests_aggr::kzg_aggr_prove_and_verify_  --features icicle --test-threads 1 -- --include-ignored
 
   prove-and-verify-aggr-tests:
-    permissions:
-      contents: read
     runs-on: large-self-hosted
     needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
@@ -622,8 +562,6 @@ jobs:
         run: cargo nextest run --release --verbose tests_aggr::kzg_aggr_prove_and_verify_  --test-threads 4 -- --include-ignored
 
   prove-and-verify-aggr-evm-tests:
-    permissions:
-      contents: read
     runs-on: large-self-hosted
     needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
@@ -647,8 +585,6 @@ jobs:
         run: cargo nextest run  --release --verbose tests_evm::kzg_evm_aggr_prove_and_verify_::t --test-threads 4 -- --include-ignored
 
   examples:
-    permissions:
-      contents: read
     runs-on: ubuntu-latest-32-cores
     needs: [build, library-tests, docs]
     steps:
@@ -668,8 +604,6 @@ jobs:
         run: cargo nextest run --release tests_examples
 
   python-tests:
-    permissions:
-      contents: read
     runs-on: non-gpu
     needs: [build, library-tests, docs]
     steps:
@@ -698,8 +632,6 @@ jobs:
         run: source .env/bin/activate; pip install pytest-asyncio; pytest -vv
 
   accuracy-measurement-tests:
-    permissions:
-      contents: read
     runs-on: non-gpu
     needs: [build, library-tests, docs, python-tests, python-integration-tests]
     steps:
@@ -732,8 +664,6 @@ jobs:
         run: source .env/bin/activate; cargo nextest run --release --verbose tests::resources_accuracy_measurement_public_outputs_
 
   python-integration-tests:
-    permissions:
-      contents: read
     runs-on: large-self-hosted
     services:
       # Label used to access the service container
@@ -806,8 +736,6 @@ jobs:
         # run: source .env/bin/activate; cargo nextest run py_tests::tests::reusable_
 
   ios-integration-tests:
-      permissions:
-        contents: read
       runs-on: macos-latest
       steps:
         - uses: actions/checkout@v4
@@ -826,8 +754,6 @@ jobs:
           run: CARGO_BUILD_TARGET=aarch64-apple-darwin RUSTUP_TOOLCHAIN=nightly-2024-07-18-aarch64-apple-darwin cargo test --test ios_integration_tests --features ios-bindings-test --no-default-features
 
   swift-package-tests:
-      permissions:
-        contents: read
       runs-on: macos-latest
       needs: [ios-integration-tests]
 

.github/workflows/static-analysis.yml

@@ -8,9 +8,8 @@ on:
 
 jobs:
   analyze:
-    permissions:
-      contents: read
     runs-on: ubuntu-latest
+
     steps:
       - uses: actions/checkout@v4
         with:

.github/workflows/swift-pm.yml

@@ -9,9 +9,6 @@ on:
 
 jobs:
   build-and-update:
-    permissions:
-      contents: read
-      packages: write
     runs-on: macos-latest
     env:
       EZKL_SWIFT_PACKAGE_REPO: github.com/zkonduit/ezkl-swift-package.git

Cargo.lock

@@ -1835,16 +1835,6 @@ dependencies = [
  "syn 2.0.90",
 ]
 
-[[package]]
-name = "env_filter"
-version = "0.1.3"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "186e05a59d4c50738528153b83b0b0194d3a29507dfec16eccd4b342903397d0"
-dependencies = [
- "log",
- "regex",
-]
-
 [[package]]
 name = "env_logger"
 version = "0.10.2"
@@ -1858,19 +1848,6 @@ dependencies = [
  "termcolor",
 ]
 
-[[package]]
-name = "env_logger"
-version = "0.11.6"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "dcaee3d8e3cfc3fd92428d477bc97fc29ec8716d180c0d74c643bb26166660e0"
-dependencies = [
- "anstream",
- "anstyle",
- "env_filter",
- "humantime",
- "log",
-]
-
 [[package]]
 name = "equivalent"
 version = "1.0.1"
@@ -1946,7 +1923,7 @@ dependencies = [
  "console_error_panic_hook",
  "criterion 0.5.1",
  "ecc",
- "env_logger 0.10.2",
+ "env_logger",
  "ethabi",
  "foundry-compilers",
  "gag",
@@ -1954,7 +1931,7 @@ dependencies = [
  "halo2_gadgets",
  "halo2_proofs",
  "halo2_solidity_verifier",
- "halo2curves 0.7.0 (git+https://github.com/privacy-scaling-explorations/halo2curves?rev=b753a832e92d5c86c5c997327a9cf9de86a18851)",
+ "halo2curves 0.7.0",
  "hex",
  "indicatif",
  "instant",
@@ -1962,6 +1939,7 @@ dependencies = [
  "lazy_static",
  "log",
  "maybe-rayon",
+ "metal",
  "mimalloc",
  "mnist",
  "num",
@@ -2416,14 +2394,14 @@ dependencies = [
 [[package]]
 name = "halo2_proofs"
 version = "0.3.0"
-source = "git+https://github.com/zkonduit/halo2#bf9d0057a82443be48c4779bbe14961c18fb5996#bf9d0057a82443be48c4779bbe14961c18fb5996"
+source = "git+https://github.com/zkonduit/halo2#ee4e1a09ebdb1f79f797685b78951c6034c430a6#ee4e1a09ebdb1f79f797685b78951c6034c430a6"
 dependencies = [
  "bincode",
  "blake2b_simd",
- "env_logger 0.10.2",
+ "env_logger",
  "ff",
  "group",
- "halo2curves 0.7.0 (git+https://github.com/privacy-scaling-explorations/halo2curves?rev=b753a832e92d5c86c5c997327a9cf9de86a18851)",
+ "halo2curves 0.7.0",
  "icicle-bn254",
  "icicle-core",
  "icicle-cuda-runtime",
@@ -2431,7 +2409,6 @@ dependencies = [
  "lazy_static",
  "log",
  "maybe-rayon",
- "mopro-msm",
  "rand_chacha",
  "rand_core 0.6.4",
  "rustc-hash 2.0.0",
@@ -2517,36 +2494,6 @@ dependencies = [
  "subtle",
 ]
 
-[[package]]
-name = "halo2curves"
-version = "0.7.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d380afeef3f1d4d3245b76895172018cfb087d9976a7cabcd5597775b2933e07"
-dependencies = [
- "blake2",
- "digest 0.10.7",
- "ff",
- "group",
- "halo2derive 0.1.0 (registry+https://github.com/rust-lang/crates.io-index)",
- "hex",
- "lazy_static",
- "num-bigint",
- "num-integer",
- "num-traits",
- "pairing",
- "pasta_curves",
- "paste",
- "rand 0.8.5",
- "rand_core 0.6.4",
- "rayon",
- "serde",
- "serde_arrays",
- "sha2",
- "static_assertions",
- "subtle",
- "unroll",
-]
-
 [[package]]
 name = "halo2curves"
 version = "0.7.0"
@@ -2556,7 +2503,7 @@ dependencies = [
  "digest 0.10.7",
  "ff",
  "group",
- "halo2derive 0.1.0 (git+https://github.com/privacy-scaling-explorations/halo2curves?rev=b753a832e92d5c86c5c997327a9cf9de86a18851)",
+ "halo2derive",
  "hex",
  "lazy_static",
  "num-bigint",
@@ -2576,20 +2523,6 @@ dependencies = [
  "unroll",
 ]
 
-[[package]]
-name = "halo2derive"
-version = "0.1.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bdb99e7492b4f5ff469d238db464131b86c2eaac814a78715acba369f64d2c76"
-dependencies = [
- "num-bigint",
- "num-integer",
- "num-traits",
- "proc-macro2",
- "quote",
- "syn 1.0.109",
-]
-
 [[package]]
 name = "halo2derive"
 version = "0.1.0"
@@ -3350,8 +3283,7 @@ dependencies = [
 [[package]]
 name = "metal"
 version = "0.29.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7ecfd3296f8c56b7c1f6fbac3c71cefa9d78ce009850c45000015f206dc7fa21"
+source = "git+https://github.com/gfx-rs/metal-rs#0e1918b34689c4b8cd13a43372f9898680547ee9"
 dependencies = [
  "bitflags 2.5.0",
  "block",
@@ -3422,28 +3354,6 @@ dependencies = [
  "byteorder",
 ]
 
-[[package]]
-name = "mopro-msm"
-version = "0.1.0"
-source = "git+https://github.com/zkonduit/metal-msm-gpu-acceleration.git#be5f647b1a6c1a6ea9024390744a2b4d87f5d002"
-dependencies = [
- "bincode",
- "env_logger 0.11.6",
- "halo2curves 0.7.0 (registry+https://github.com/rust-lang/crates.io-index)",
- "instant",
- "itertools 0.13.0",
- "lazy_static",
- "log",
- "metal",
- "objc",
- "once_cell",
- "rand 0.8.5",
- "rayon",
- "serde",
- "thiserror",
- "walkdir",
-]
-
 [[package]]
 name = "native-tls"
 version = "0.2.11"

Cargo.toml

@@ -91,6 +91,7 @@ pyo3-async-runtimes = { git = "https://github.com/PyO3/pyo3-async-runtimes", ver
 pyo3-log = { version = "0.12.0", default-features = false, optional = true }
 tract-onnx = { git = "https://github.com/sonos/tract/", rev = "37132e0397d0a73e5bd3a8615d932dabe44f6736", default-features = false, optional = true }
 tabled = { version = "0.12.0", optional = true }
+metal = { git = "https://github.com/gfx-rs/metal-rs", optional = true }
 objc = { version = "0.2.4", optional = true }
 mimalloc = { version = "0.1", optional = true }
 pyo3-stub-gen = { version = "0.6.0", optional = true }
@@ -276,14 +277,13 @@ icicle = ["halo2_proofs/icicle_gpu"]
 empty-cmd = []
 no-banner = []
 no-update = []
-macos-metal = ["halo2_proofs/macos"]
-ios-metal = ["halo2_proofs/ios"]
+
 
 [patch.'https://github.com/zkonduit/halo2']
-halo2_proofs = { git = "https://github.com/zkonduit/halo2#bf9d0057a82443be48c4779bbe14961c18fb5996", package = "halo2_proofs" }
+halo2_proofs = { git = "https://github.com/zkonduit/halo2#ee4e1a09ebdb1f79f797685b78951c6034c430a6", package = "halo2_proofs" }
 
 [patch.'https://github.com/zkonduit/halo2#0654e92bdf725fd44d849bfef3643870a8c7d50b']
-halo2_proofs = { git = "https://github.com/zkonduit/halo2#bf9d0057a82443be48c4779bbe14961c18fb5996", package = "halo2_proofs" }
+halo2_proofs = { git = "https://github.com/zkonduit/halo2#ee4e1a09ebdb1f79f797685b78951c6034c430a6", package = "halo2_proofs" }
 
 [patch.crates-io]
 uniffi_testing = { git = "https://github.com/ElusAegis/uniffi-rs", branch = "feat/testing-feature-build-fix" }

docs/python/src/conf.py

@@ -1,7 +1,7 @@
 import ezkl
 
 project = 'ezkl'
-release = '0.0.0'
+release = '18.0.0'
 version = release
 
 

src/tensor/mod.rs

@@ -24,6 +24,9 @@ use std::path::PathBuf;
 pub use val::*;
 pub use var::*;
 
+#[cfg(feature = "metal")]
+use instant::Instant;
+
 use crate::{
     circuit::utils,
     fieldutils::{integer_rep_to_felt, IntegerRep},
@@ -37,6 +40,8 @@ use halo2_proofs::{
     poly::Rotation,
 };
 use itertools::Itertools;
+#[cfg(feature = "metal")]
+use metal::{Device, MTLResourceOptions, MTLSize};
 use std::error::Error;
 use std::fmt::Debug;
 use std::io::Read;
@@ -44,6 +49,31 @@ use std::iter::Iterator;
 use std::ops::{Add, Deref, DerefMut, Div, Mul, Neg, Range, Sub};
 use std::{cmp::max, ops::Rem};
 
+#[cfg(feature = "metal")]
+use std::collections::HashMap;
+
+#[cfg(feature = "metal")]
+const LIB_DATA: &[u8] = include_bytes!("metal/tensor_ops.metallib");
+
+#[cfg(feature = "metal")]
+lazy_static::lazy_static! {
+    static ref DEVICE: Device = Device::system_default().expect("no device found");
+
+    static ref LIB: metal::Library = DEVICE.new_library_with_data(LIB_DATA).unwrap();
+
+    static ref QUEUE: metal::CommandQueue = DEVICE.new_command_queue();
+
+    static ref PIPELINES: HashMap<String, metal::ComputePipelineState> = {
+        let mut map = HashMap::new();
+        for name in ["add", "sub", "mul"] {
+            let function = LIB.get_function(name, None).unwrap();
+            let pipeline = DEVICE.new_compute_pipeline_state_with_function(&function).unwrap();
+            map.insert(name.to_string(), pipeline);
+        }
+        map
+    };
+}
+
 /// The (inner) type of tensor elements.
 pub trait TensorType: Clone + Debug + 'static {
     /// Returns the zero value.
@@ -1374,6 +1404,10 @@ impl<T: TensorType + Add<Output = T> + std::marker::Send + std::marker::Sync> Ad
         let lhs = self.expand(&broadcasted_shape).unwrap();
         let rhs = rhs.expand(&broadcasted_shape).unwrap();
 
+        #[cfg(feature = "metal")]
+        let res = metal_tensor_op(&lhs, &rhs, "add");
+
+        #[cfg(not(feature = "metal"))]
         let res = {
             let mut res: Tensor<T> = lhs
                 .par_iter()
@@ -1471,6 +1505,10 @@ impl<T: TensorType + Sub<Output = T> + std::marker::Send + std::marker::Sync> Su
         let lhs = self.expand(&broadcasted_shape).unwrap();
         let rhs = rhs.expand(&broadcasted_shape).unwrap();
 
+        #[cfg(feature = "metal")]
+        let res = metal_tensor_op(&lhs, &rhs, "sub");
+
+        #[cfg(not(feature = "metal"))]
         let res = {
             let mut res: Tensor<T> = lhs
                 .par_iter()
@@ -1538,6 +1576,10 @@ impl<T: TensorType + Mul<Output = T> + std::marker::Send + std::marker::Sync> Mu
         let lhs = self.expand(&broadcasted_shape).unwrap();
         let rhs = rhs.expand(&broadcasted_shape).unwrap();
 
+        #[cfg(feature = "metal")]
+        let res = metal_tensor_op(&lhs, &rhs, "mul");
+
+        #[cfg(not(feature = "metal"))]
         let res = {
             let mut res: Tensor<T> = lhs
                 .par_iter()
@@ -1769,4 +1811,66 @@ mod tests {
         let b = Tensor::<IntegerRep>::new(Some(&[1, 4]), &[2, 1]).unwrap();
         assert_eq!(a.get_slice(&[0..2, 0..1]).unwrap(), b);
     }
+
+    #[test]
+    #[cfg(feature = "metal")]
+    fn tensor_metal_int() {
+        let a = Tensor::<i64>::new(Some(&[1, 2, 3, 4]), &[2, 2]).unwrap();
+        let b = Tensor::<i64>::new(Some(&[1, 2, 3, 4]), &[2, 2]).unwrap();
+        let c = metal_tensor_op(&a, &b, "add");
+        assert_eq!(c, Tensor::new(Some(&[2, 4, 6, 8]), &[2, 2]).unwrap());
+
+        let c = metal_tensor_op(&a, &b, "sub");
+        assert_eq!(c, Tensor::new(Some(&[0, 0, 0, 0]), &[2, 2]).unwrap());
+
+        let c = metal_tensor_op(&a, &b, "mul");
+        assert_eq!(c, Tensor::new(Some(&[1, 4, 9, 16]), &[2, 2]).unwrap());
+    }
+
+    #[test]
+    #[cfg(feature = "metal")]
+    fn tensor_metal_felt() {
+        use halo2curves::bn256::Fr;
+
+        let a = Tensor::<Fr>::new(
+            Some(&[Fr::from(1), Fr::from(2), Fr::from(3), Fr::from(4)]),
+            &[2, 2],
+        )
+        .unwrap();
+        let b = Tensor::<Fr>::new(
+            Some(&[Fr::from(1), Fr::from(2), Fr::from(3), Fr::from(4)]),
+            &[2, 2],
+        )
+        .unwrap();
+
+        let c = metal_tensor_op(&a, &b, "add");
+        assert_eq!(
+            c,
+            Tensor::<Fr>::new(
+                Some(&[Fr::from(2), Fr::from(4), Fr::from(6), Fr::from(8)]),
+                &[2, 2],
+            )
+            .unwrap()
+        );
+
+        let c = metal_tensor_op(&a, &b, "sub");
+        assert_eq!(
+            c,
+            Tensor::<Fr>::new(
+                Some(&[Fr::from(0), Fr::from(0), Fr::from(0), Fr::from(0)]),
+                &[2, 2],
+            )
+            .unwrap()
+        );
+
+        let c = metal_tensor_op(&a, &b, "mul");
+        assert_eq!(
+            c,
+            Tensor::<Fr>::new(
+                Some(&[Fr::from(1), Fr::from(4), Fr::from(9), Fr::from(16)]),
+                &[2, 2],
+            )
+            .unwrap()
+        );
+    }
 }

tests/integration_tests.rs

@@ -2795,10 +2795,7 @@ mod native_tests {
             "--features",
             "icicle",
         ];
-        #[cfg(feature = "macos-metal")]
-        let args = ["build", "--release", "--bin", "ezkl", "--features", "macos-metal"];
-        // not macos-metal and not icicle
-        #[cfg(all(not(feature = "icicle"), not(feature = "macos-metal")))]
+        #[cfg(not(feature = "icicle"))]
         let args = ["build", "--release", "--bin", "ezkl"];
         #[cfg(not(feature = "mv-lookup"))]
         let args = [