chore(ci): remove wasm test that is not relevant

.cargo/config.toml

@@ -1,6 +1,2 @@
 [alias]
 xtask = "run --manifest-path ./tasks/Cargo.toml --"
-
-# Accessed by wasm-bindgen when testing for the wasm target
-[target.wasm32-unknown-unknown]
-runner = 'wasm-bindgen-test-runner'

.github/actions/hyperstack_setup/action.yml

@@ -1,53 +0,0 @@
-name: Setup Cuda
-description: Setup Cuda on Hyperstack instance
-
-inputs:
-  cuda-version:
-    description: Version of Cuda to use
-    required: true
-  gcc-version:
-    description: Version of GCC to use
-    required: true
-  cmake-version:
-    description: Version of cmake to use
-    default: 3.29.6
-
-runs:
-  using: "composite"
-  steps:
-    # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-    - name: Install dependencies
-      shell: bash
-      run: |
-        sudo apt update
-        sudo apt install -y checkinstall zlib1g-dev libssl-dev libclang-dev
-        wget https://github.com/Kitware/CMake/releases/download/v${{ inputs.cmake-version }}/cmake-${{ inputs.cmake-version }}.tar.gz
-        tar -zxvf cmake-${{ inputs.cmake-version }}.tar.gz
-        cd cmake-${{ inputs.cmake-version }}
-        ./bootstrap
-        make -j"$(nproc)"
-        sudo make install
-
-    - name: Export CUDA variables
-      shell: bash
-      run: |
-        CUDA_PATH=/usr/local/cuda-${{ inputs.cuda-version }}
-        echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
-        echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
-        echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
-        echo "CUDACXX=/usr/local/cuda-${{ inputs.cuda-version }}/bin/nvcc" >> "${GITHUB_ENV}"
-
-    # Specify the correct host compilers
-    - name: Export gcc and g++ variables
-      shell: bash
-      run: |
-        {
-          echo "CC=/usr/bin/gcc-${{ inputs.gcc-version }}";
-          echo "CXX=/usr/bin/g++-${{ inputs.gcc-version }}";
-          echo "CUDAHOSTCXX=/usr/bin/g++-${{ inputs.gcc-version }}";
-          echo "HOME=/home/ubuntu";
-        } >> "${GITHUB_ENV}"
-
-    - name: Check device is detected
-      shell: bash
-      run: nvidia-smi

.github/workflows/aws_tfhe_backward_compat_tests.yml

@@ -26,7 +26,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -44,13 +44,17 @@ jobs:
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: 'false'
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
@@ -72,7 +76,7 @@ jobs:
           echo "branch=${BRANCH}" >> "${GITHUB_OUTPUT}"
 
       - name: Clone test data
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: 'false'
           repository: zama-ai/tfhe-backward-compat-data
@@ -87,7 +91,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Backward compatibility tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -100,7 +104,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -111,7 +115,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (backward-compat-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_fast_tests.yml

@@ -11,26 +11,16 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' || github.event_name == 'pull_request_target' }}
+  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
   pull_request:
-  pull_request_target:
 
 jobs:
-  check-user-permission:
-    if: github.event_name == 'pull_request' || github.event_name == 'pull_request_target'
-    uses: ./.github/workflows/check_triggering_actor.yml
-    secrets:
-      TOKEN: ${{ secrets.GITHUB_TOKEN }}
-
   should-run:
     runs-on: ubuntu-latest
-    needs: check-user-permission
-    if: github.event_name != 'pull_request_target' ||
-      needs.check-user-permission.result == 'success'
     permissions:
       pull-requests: write
     outputs:
@@ -61,27 +51,24 @@ jobs:
       any_file_changed: ${{ env.IS_PULL_REQUEST == 'false' || steps.aggregated-changes.outputs.any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-          ref: ${{ github.event.pull_request.head.sha }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           since_last_remote_commit: true
           files_yaml: |
             dependencies:
               - tfhe/Cargo.toml
-              - tfhe-csprng/**
-              - tfhe-fft/**
+              - concrete-csprng/**
               - tfhe-zk-pok/**
               - utils/tfhe-versionable/**
               - utils/tfhe-versionable-derive/**
             csprng:
-              - tfhe-csprng/**
+              - concrete-csprng/**
             zk_pok:
               - tfhe-zk-pok/**
             versionable:
@@ -144,7 +131,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -159,26 +146,29 @@ jobs:
       (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
     needs: [ should-run, setup-instance ]
     concurrency:
-      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
+      group: ${{ github.workflow }}_${{ github.ref }}
       cancel-in-progress: true
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: 'false'
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-          ref: ${{ github.event.pull_request.head.sha }}
+
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
-      - name: Run tfhe-csprng tests
+      - name: Run concrete-csprng tests
         if: needs.should-run.outputs.csprng_test == 'true'
         run: |
-          make test_tfhe_csprng
+          make test_concrete_csprng
 
       - name: Run tfhe-zk-pok tests
         if: needs.should-run.outputs.zk_pok_test == 'true'
@@ -205,37 +195,10 @@ jobs:
         run: |
           make test_user_doc
 
-      - name: Get Node version
-        run: |
-          echo "NODE_VERSION=$(make node_version)" >> "${GITHUB_ENV}"
-
-      - name: Node cache restoration
-        id: node-cache
-        uses: actions/cache/restore@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
-        with:
-          path: |
-            ~/.nvm
-            ~/.npm
-          key: node-${{ env.NODE_VERSION }}
-
-      - name: Install Node
-        if: steps.node-cache.outputs.cache-hit != 'true'
-        run: |
-          make install_node
-
-      - name: Node cache save
-        uses: actions/cache/save@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
-        if: steps.node-cache.outputs.cache-hit != 'true'
-        with:
-          path: |
-            ~/.nvm
-            ~/.npm
-          key: node-${{ env.NODE_VERSION }}
-
       - name: Run js on wasm API tests
         if: needs.should-run.outputs.wasm_test == 'true'
         run: |
-          make test_nodejs_wasm_api_ci
+          make test_nodejs_wasm_api_in_docker
 
       - name: Gen Keys if required
         if: needs.should-run.outputs.shortint_test == 'true' ||
@@ -258,18 +221,14 @@ jobs:
         run: |
           make test_high_level_api
 
-      - name: Run safe serialization tests
+      - name: Run safe deserialization tests
         run: |
-          make test_safe_serialization
-
-      - name: Run zk tests
-        run: |
-          make test_zk
+          make test_safe_deserialization
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Fast AWS tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -282,7 +241,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -293,7 +252,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (fast-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_integer_tests.yml

@@ -39,27 +39,24 @@ jobs:
         steps.changed-files.outputs.integer_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
           persist-credentials: "false"
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           since_last_remote_commit: true
           files_yaml: |
             integer:
               - tfhe/Cargo.toml
-              - tfhe-csprng/**
-              - tfhe-fft/**
+              - concrete-csprng/**
               - tfhe-zk-pok/**
               - tfhe/src/core_crypto/**
               - tfhe/src/shortint/**
               - tfhe/src/integer/**
-              - .github/workflows/aws_tfhe_integer_tests.yml
 
   setup-instance:
     name: Setup instance (unsigned-integer-tests)
@@ -75,7 +72,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -93,13 +90,17 @@ jobs:
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: "false"
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
@@ -127,7 +128,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Unsigned Integer tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -140,7 +141,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -151,7 +152,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (unsigned-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_signed_integer_tests.yml

@@ -39,27 +39,24 @@ jobs:
         steps.changed-files.outputs.integer_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
           persist-credentials: "false"
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           since_last_remote_commit: true
           files_yaml: |
             integer:
               - tfhe/Cargo.toml
-              - tfhe-csprng/**
-              - tfhe-fft/**
+              - concrete-csprng/**
               - tfhe-zk-pok/**
               - tfhe/src/core_crypto/**
               - tfhe/src/shortint/**
               - tfhe/src/integer/**
-              - .github/workflows/aws_tfhe_signed_integer_tests.yml
 
   setup-instance:
     name: Setup instance (unsigned-integer-tests)
@@ -75,7 +72,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -93,13 +90,17 @@ jobs:
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: "false"
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
@@ -131,7 +132,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Signed Integer tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -144,7 +145,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -155,7 +156,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (signed-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_tests.yml

@@ -40,9 +40,6 @@ jobs:
       shortint_test: ${{ env.IS_PULL_REQUEST == 'false' ||
         steps.changed-files.outputs.shortint_any_changed ||
         steps.changed-files.outputs.dependencies_any_changed }}
-      strings_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.strings_any_changed ||
-        steps.changed-files.outputs.dependencies_any_changed }}
       high_level_api_test: ${{ env.IS_PULL_REQUEST == 'false' ||
         steps.changed-files.outputs.high_level_api_any_changed ||
         steps.changed-files.outputs.dependencies_any_changed }}
@@ -60,24 +57,22 @@ jobs:
       any_file_changed: ${{ env.IS_PULL_REQUEST == 'false' || steps.aggregated-changes.outputs.any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           since_last_remote_commit: true
           files_yaml: |
             dependencies:
               - tfhe/Cargo.toml
-              - tfhe-csprng/**
-              - tfhe-fft/**
+              - concrete-csprng/**
               - tfhe-zk-pok/**
             csprng:
-              - tfhe-csprng/**
+              - concrete-csprng/**
             zk_pok:
               - tfhe-zk-pok/**
             core_crypto:
@@ -88,11 +83,6 @@ jobs:
             shortint:
               - tfhe/src/core_crypto/**
               - tfhe/src/shortint/**
-            strings:
-              - tfhe/src/core_crypto/**
-              - tfhe/src/shortint/**
-              - tfhe/src/integer/**
-              - tfhe/src/strings/**
             high_level_api:
               - tfhe/src/**
               - '!tfhe/src/c_api/**'
@@ -122,7 +112,6 @@ jobs:
           steps.changed-files.outputs.core_crypto_any_changed == 'true' ||
           steps.changed-files.outputs.boolean_any_changed == 'true' ||
           steps.changed-files.outputs.shortint_any_changed == 'true' ||
-          steps.changed-files.outputs.strings_any_changed == 'true' ||
           steps.changed-files.outputs.high_level_api_any_changed == 'true' ||
           steps.changed-files.outputs.c_api_any_changed == 'true' ||
           steps.changed-files.outputs.examples_any_changed == 'true' ||
@@ -142,7 +131,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -162,20 +151,24 @@ jobs:
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: 'false'
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
-      - name: Run tfhe-csprng tests
+      - name: Run concrete-csprng tests
         if: needs.should-run.outputs.csprng_test == 'true'
         run: |
-          make test_tfhe_csprng
+          make test_concrete_csprng
 
       - name: Run tfhe-zk-pok tests
         if: needs.should-run.outputs.zk_pok_test == 'true'
@@ -212,11 +205,6 @@ jobs:
         run: |
           BIG_TESTS_INSTANCE=TRUE make test_shortint_ci
 
-      - name: Run strings tests
-        if: needs.should-run.outputs.strings_test == 'true'
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_strings
-
       - name: Run high-level API tests
         if: needs.should-run.outputs.high_level_api_test == 'true'
         run: |
@@ -237,7 +225,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "CPU tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -250,7 +238,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -261,7 +249,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cpu-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_wasm_tests.yml

@@ -27,7 +27,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -45,45 +45,23 @@ jobs:
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: 'false'
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
-      - name: Get Node version
-        run: |
-          echo "NODE_VERSION=$(make node_version)" >> "${GITHUB_ENV}"
-
-      - name: Node cache restoration
-        id: node-cache
-        uses: actions/cache/restore@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
-        with:
-          path: |
-            ~/.nvm
-            ~/.npm
-          key: node-${{ env.NODE_VERSION }}
-
-      - name: Install Node
-        if: steps.node-cache.outputs.cache-hit != 'true'
-        run: |
-          make install_node
-
-      - name: Node cache save
-        uses: actions/cache/save@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
-        if: steps.node-cache.outputs.cache-hit != 'true'
-        with:
-          path: |
-            ~/.nvm
-            ~/.npm
-          key: node-${{ env.NODE_VERSION }}
-
       - name: Install web resources
         run: |
+          make install_node
           make install_chrome_browser
           make install_chrome_web_driver
 
@@ -93,20 +71,16 @@ jobs:
 
       - name: Run js on wasm API tests
         run: |
-          make test_nodejs_wasm_api_ci
+          make test_nodejs_wasm_api_in_docker
 
       - name: Run parallel wasm tests
         run: |
           make test_web_js_api_parallel_chrome_ci
 
-      - name: Run x86_64/wasm zk compatibility tests
-        run: |
-          make test_zk_wasm_x86_compat_ci
-
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "WASM tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -119,7 +93,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -130,7 +104,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (wasm-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_boolean.yml

@@ -29,7 +29,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -48,7 +48,7 @@ jobs:
     continue-on-error: true
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
@@ -61,8 +61,13 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
       - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: nightly
 
@@ -80,7 +85,8 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
-          --name-suffix avx512
+          --name-suffix avx512 \
+          --throughput
 
       - name: Measure key sizes
         run: |
@@ -89,17 +95,17 @@ jobs:
       - name: Parse key sizes results
         run: |
           python3 ./ci/benchmark_parser.py tfhe/boolean_key_sizes.csv ${{ env.RESULTS_FILENAME }} \
-          --object-sizes \
+          --key-sizes \
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
         with:
           name: ${{ github.sha }}_boolean
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           repository: zama-ai/slab
           path: slab
@@ -114,7 +120,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Boolean benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -127,7 +133,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -138,7 +144,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (boolean-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_core_crypto.yml

@@ -26,7 +26,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -44,7 +44,7 @@ jobs:
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
@@ -57,8 +57,13 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
       - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: nightly
 
@@ -78,16 +83,17 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --name-suffix avx512 \
-          --walk-subdirs
+          --walk-subdirs \
+          --throughput
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
         with:
           name: ${{ github.sha }}_core_crypto
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           repository: zama-ai/slab
           path: slab
@@ -102,7 +108,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "PBS benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -115,7 +121,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -126,7 +132,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (core-crypto-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_erc20.yml

@@ -1,141 +0,0 @@
-# Run all ERC20 benchmarks on an AWS instance and return parsed results to Slab CI bot.
-name: ERC20 benchmarks
-
-on:
-  workflow_dispatch:
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 5a.m.
-    - cron: '0 5 * * 6'
-
-env:
-  CARGO_TERM_COLOR: always
-  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-
-jobs:
-  setup-instance:
-    name: Setup instance (erc20-benchmarks)
-    runs-on: ubuntu-latest
-    if: github.event_name == 'workflow_dispatch' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: bench
-
-  erc20-benchmarks:
-    name: Execute ERC20 benchmarks
-    needs: setup-instance
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    continue-on-error: true
-    timeout-minutes: 720  # 12 hours
-    steps:
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Get benchmark details
-        run: |
-          {
-            echo "BENCH_DATE=$(date --iso-8601=seconds)";
-            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
-            echo "COMMIT_HASH=$(git describe --tags --dirty)";
-          } >> "${GITHUB_ENV}"
-
-      - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
-        with:
-          toolchain: nightly
-
-      - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          repository: zama-ai/slab
-          path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Run benchmarks
-        run: |
-          make bench_hlapi_erc20
-
-      - name: Parse results
-        run: |
-          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
-          --database tfhe_rs \
-          --hardware "hpc7a.96xlarge" \
-          --project-version "${{ env.COMMIT_HASH }}" \
-          --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
-          --bench-date "${{ env.BENCH_DATE }}" \
-          --walk-subdirs \
-          --name-suffix avx512
-
-      - name: Parse PBS counts
-        run: |
-          python3 ./ci/benchmark_parser.py tfhe/erc20_pbs_count.csv ${{ env.RESULTS_FILENAME }} \
-          --object-sizes \
-          --append-results
-
-      - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
-        with:
-          name: ${{ github.sha }}_erc20
-          path: ${{ env.RESULTS_FILENAME }}
-
-      - name: Send data to Slab
-        shell: bash
-        run: |
-          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
-          --slab-url "${{ secrets.SLAB_URL }}"
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "ERC20 benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (erc20-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, erc20-benchmarks ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (erc20-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_4090.yml

@@ -39,7 +39,7 @@ jobs:
 
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
@@ -54,12 +54,12 @@ jobs:
           echo "FAST_BENCH=TRUE" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           repository: zama-ai/slab
           path: slab
@@ -79,10 +79,11 @@ jobs:
           --branch ${{ github.ref_name }} \
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
-          --walk-subdirs
+          --walk-subdirs \
+          --throughput
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
         with:
           name: ${{ github.sha }}_integer_multi_bit_gpu_default
           path: ${{ env.RESULTS_FILENAME }}
@@ -96,7 +97,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Integer RTX 4090 full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -113,10 +114,9 @@ jobs:
 
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -127,12 +127,12 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           repository: zama-ai/slab
           path: slab
@@ -154,10 +154,10 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
-      
+          --throughput
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
         with:
           name: ${{ github.sha }}_core_crypto
           path: ${{ env.RESULTS_FILENAME }}
@@ -179,7 +179,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Core crypto RTX 4090 full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_core_crypto.yml

@@ -27,7 +27,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -48,19 +48,28 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
     steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
-
       - name: Get benchmark details
         run: |
           {
@@ -75,10 +84,31 @@ jobs:
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: nightly
 
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CUDA_PATH=$CUDA_PATH";
+            echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH";
+            echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc";
+          } >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "HOME=/home/ubuntu";
+          } >> "${GITHUB_ENV}"
+
       - name: Run benchmarks with AVX512
         run: |
           make bench_pbs_gpu
@@ -95,16 +125,17 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --name-suffix avx512 \
-          --walk-subdirs
+          --walk-subdirs \
+          --throughput
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
         with:
           name: ${{ github.sha }}_core_crypto
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           repository: zama-ai/slab
           path: slab
@@ -124,7 +155,7 @@ jobs:
     continue-on-error: true
     steps:
       - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ needs.cuda-core-crypto-benchmarks.result }}
           SLACK_MESSAGE: "PBS GPU benchmarks finished with status: ${{ needs.cuda-core-crypto-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
@@ -137,7 +168,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -148,7 +179,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-core-crypto-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_erc20.yml

@@ -1,44 +0,0 @@
-# Run CUDA ERC20 benchmarks on a Hyperstack VM and return parsed results to Slab CI bot.
-name: Cuda ERC20 benchmarks
-
-on:
-  workflow_dispatch:
-    inputs:
-      profile:
-        description: "Instance type"
-        required: true
-        type: choice
-        options:
-          - "l40 (n3-L40x1)"
-          - "single-h100 (n3-H100x1)"
-          - "2-h100 (n3-H100x2)"
-          - "4-h100 (n3-H100x4)"
-          - "multi-h100 (n3-H100x8)"
-          - "multi-h100-nvlink (n3-H100x8-NVLink)"
-          - "multi-h100-sxm5 (n3-H100x8-SXM5)"
-
-jobs:
-  parse-inputs:
-    runs-on: ubuntu-latest
-    outputs:
-      profile: ${{ steps.parse_profile.outputs.profile }}
-      hardware_name: ${{ steps.parse_hardware_name.outputs.name }}
-    steps:
-      - name: Parse profile
-        id: parse_profile
-        run: |
-          echo "profile=$(echo '${{ inputs.profile }}' | sed 's|\(.*\)[[:space:]](.*)|\1|')" >> "${GITHUB_OUTPUT}"
-
-      - name: Parse hardware name
-        id: parse_hardware_name
-        run: |
-          echo "name=$(echo '${{ inputs.profile }}' | sed 's|.*[[:space:]](\(.*\))|\1|')" >> "${GITHUB_OUTPUT}"
-
-  run-benchmarks:
-    name: Run benchmarks
-    needs: parse-inputs
-    uses: ./.github/workflows/benchmark_gpu_erc20_common.yml
-    with:
-      profile: ${{ needs.parse-inputs.outputs.profile }}
-      hardware_name: ${{ needs.parse-inputs.outputs.hardware_name }}
-    secrets: inherit

.github/workflows/benchmark_gpu_erc20_common.yml

@@ -1,180 +0,0 @@
-# Run ERC20 benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
-name: Cuda ERC20 benchmarks - common
-
-on:
-  workflow_call:
-    inputs:
-      backend:
-        type: string
-        default: hyperstack
-      profile:
-        type: string
-        required: true
-      hardware_name:
-        type: string
-        required: true
-    secrets:
-      FHE_ACTIONS_TOKEN:
-        required: true
-      SLAB_ACTION_TOKEN:
-        required: true
-      SLAB_BASE_URL:
-        required: true
-      SLAB_URL:
-        required: true
-      JOB_SECRET:
-        required: true
-      SLACK_CHANNEL:
-        required: true
-      BOT_USERNAME:
-        required: true
-      SLACK_WEBHOOK:
-        required: true
-
-env:
-  CARGO_TERM_COLOR: always
-  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
-  PARSE_INTEGER_BENCH_CSV_FILE: tfhe_rs_integer_benches_${{ github.sha }}.csv
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-
-jobs:
-  setup-instance:
-    name: Setup instance (cuda-erc20-benchmarks)
-    runs-on: ubuntu-latest
-    if:  github.event_name == 'workflow_dispatch' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: ${{ inputs.backend }}
-          profile: ${{ inputs.profile }}
-
-  cuda-erc20-benchmarks:
-    name: Cuda ERC20 benchmarks (${{ inputs.profile }})
-    needs: setup-instance
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    strategy:
-      fail-fast: false
-      # explicit include-based build matrix, of known valid options
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11
-    steps:
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
-
-      - name: Get benchmark details
-        run: |
-          {
-            echo "BENCH_DATE=$(date --iso-8601=seconds)";
-            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
-            echo "COMMIT_HASH=$(git describe --tags --dirty)";
-          } >> "${GITHUB_ENV}"
-
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
-        with:
-          toolchain: nightly
-
-      - name: Run benchmarks
-        run: |
-          make bench_hlapi_erc20_gpu
-
-      - name: Parse results
-        run: |
-          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
-          --database tfhe_rs \
-          --hardware "${{ inputs.hardware_name }}" \
-          --backend gpu \
-          --project-version "${{ env.COMMIT_HASH }}" \
-          --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
-          --bench-date "${{ env.BENCH_DATE }}" \
-          --walk-subdirs \
-          --name-suffix avx512
-
-      - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
-        with:
-          name: ${{ github.sha }}_erc20_${{ inputs.profile }}
-          path: ${{ env.RESULTS_FILENAME }}
-
-      - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          repository: zama-ai/slab
-          path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Send data to Slab
-        shell: bash
-        run: |
-          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
-          --slab-url "${{ secrets.SLAB_URL }}"
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-erc20-benchmarks ]
-    runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-erc20-benchmarks.result != 'skipped' && failure() }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ needs.cuda-erc20-benchmarks.result }}
-          SLACK_MESSAGE: "Cuda ERC20 benchmarks (${{ inputs.profile }}) finished with status: ${{ needs.cuda-erc20-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-erc20-${{ inputs.profile }}-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-erc20-benchmarks, slack-notify ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-erc20-${{ inputs.profile }}-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_erc20_weekly.yml

@@ -1,35 +0,0 @@
-# Run CUDA ERC20 benchmarks on multiple Hyperstack VMs and return parsed results to Slab CI bot.
-name: Cuda ERC20 weekly benchmarks
-
-on:
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 5a.m.
-    - cron: '0 5 * * 6'
-
-jobs:
-  run-benchmarks-1-h100:
-    name: Run benchmarks (1xH100)
-    if: github.repository == 'zama-ai/tfhe-rs'
-    uses: ./.github/workflows/benchmark_gpu_erc20_common.yml
-    with:
-      profile: single-h100
-      hardware_name: n3-H100x1
-    secrets: inherit
-
-  run-benchmarks-2-h100:
-    name: Run benchmarks (2xH100)
-    if: github.repository == 'zama-ai/tfhe-rs'
-    uses: ./.github/workflows/benchmark_gpu_erc20_common.yml
-    with:
-      profile: 2-h100
-      hardware_name: n3-H100x2
-    secrets: inherit
-
-  run-benchmarks-8-h100:
-    name: Run benchmarks (8xH100)
-    if: github.repository == 'zama-ai/tfhe-rs'
-    uses: ./.github/workflows/benchmark_gpu_erc20_common.yml
-    with:
-      profile: multi-h100
-      hardware_name: n3-H100x8
-    secrets: inherit

.github/workflows/benchmark_gpu_integer.yml

@@ -1,79 +1,201 @@
-# Run CUDA benchmarks on a Hyperstack VM and return parsed results to Slab CI bot.
-name: Cuda benchmarks
+# Run integer benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
+name: Integer GPU benchmarks
 
 on:
   workflow_dispatch:
-    inputs:
-      profile:
-        description: "Instance type"
-        required: true
-        type: choice
-        options:
-          - "l40 (n3-L40x1)"
-          - "single-h100 (n3-H100x1)"
-          - "2-h100 (n3-H100x2)"
-          - "4-h100 (n3-H100x4)"
-          - "multi-h100 (n3-H100x8)"
-          - "multi-h100-nvlink (n3-H100x8-NVLink)"
-          - "multi-h100-sxm5 (n3-H100x8-SXM5)"
-          - "multi-a100-nvlink (n3-A100x8-NVLink)"
-      command:
-        description: "Benchmark command to run"
-        type: choice
-        default: integer_multi_bit
-        options:
-          - integer
-          - integer_multi_bit
-          - integer_compression
-          - pbs
-          - ks
-      op_flavor:
-        description: "Operations set to run"
-        type: choice
-        default: default
-        options:
-          - default
-          - fast_default
-          - unchecked
-      all_precisions:
-        description: "Run all precisions"
-        type: boolean
-        default: false
-      bench_type:
-        description: "Benchmarks type"
-        type: choice
-        default: latency
-        options:
-          - latency
-          - throughput
-          - both
+  push:
+    branches:
+      - main
+
+env:
+  CARGO_TERM_COLOR: always
+  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
+  PARSE_INTEGER_BENCH_CSV_FILE: tfhe_rs_integer_benches_${{ github.sha }}.csv
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
 
 jobs:
-  parse-inputs:
+  setup-instance:
+    name: Setup instance (cuda-integer-benchmarks)
     runs-on: ubuntu-latest
+    if:  github.event_name == 'workflow_dispatch' ||
+      (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs')
     outputs:
-      profile: ${{ steps.parse_profile.outputs.profile }}
-      hardware_name: ${{ steps.parse_hardware_name.outputs.name }}
+      runner-name: ${{ steps.start-instance.outputs.label }}
     steps:
-      - name: Parse profile
-        id: parse_profile
-        run: |
-          echo "profile=$(echo '${{ inputs.profile }}' | sed 's|\(.*\)[[:space:]](.*)|\1|')" >> "${GITHUB_OUTPUT}"
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: hyperstack
+          profile: single-h100
 
-      - name: Parse hardware name
-        id: parse_hardware_name
+  cuda-integer-benchmarks:
+    name: Execute GPU integer benchmarks
+    needs: setup-instance
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    strategy:
+      fail-fast: false
+      # explicit include-based build matrix, of known valid options
+      matrix:
+        include:
+          - os: ubuntu-22.04
+            cuda: "12.2"
+            gcc: 11
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
+    steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
         run: |
-          echo "name=$(echo '${{ inputs.profile }}' | sed 's|.*[[:space:]](\(.*\))|\1|')" >> "${GITHUB_OUTPUT}"
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
 
-  run-benchmarks:
-    name: Run benchmarks
-    needs: parse-inputs
-    uses: ./.github/workflows/benchmark_gpu_integer_common.yml
-    with:
-      profile: ${{ needs.parse-inputs.outputs.profile }}
-      hardware_name: ${{ needs.parse-inputs.outputs.hardware_name }}
-      command: ${{ inputs.command }}
-      op_flavor: ${{ inputs.op_flavor }}
-      bench_type: ${{ inputs.bench_type }}
-      all_precisions: ${{ inputs.all_precisions }}
-    secrets: inherit
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(date --iso-8601=seconds)";
+            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
+            echo "COMMIT_HASH=$(git describe --tags --dirty)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        with:
+          toolchain: nightly
+
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CUDA_PATH=$CUDA_PATH";
+            echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH";
+            echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc";
+          } >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+          } >> "${GITHUB_ENV}"
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
+      - name: Run benchmarks with AVX512
+        run: |
+          make FAST_BENCH=TRUE BENCH_OP_FLAVOR=default bench_integer_gpu
+
+      - name: Parse benchmarks to csv
+        run: |
+          make PARSE_INTEGER_BENCH_CSV_FILE=${{ env.PARSE_INTEGER_BENCH_CSV_FILE }} \
+            parse_integer_benches
+
+      - name: Upload csv results artifact
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
+        with:
+          name: ${{ github.sha }}_csv_integer
+          path: ${{ env.PARSE_INTEGER_BENCH_CSV_FILE }}
+
+      - name: Parse results
+        run: |
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware "n3-H100x1" \
+          --backend gpu \
+          --project-version "${{ env.COMMIT_HASH }}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${{ env.COMMIT_DATE }}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --throughput
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
+        with:
+          name: ${{ github.sha }}_integer
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
+
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-integer-benchmarks ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-integer-benchmarks.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ needs.cuda-integer-benchmarks.result }}
+          SLACK_MESSAGE: "Integer GPU benchmarks finished with status: ${{ needs.cuda-integer-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (cuda-integer-benchmarks)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, cuda-integer-benchmarks, slack-notify ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cuda-integer-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_integer_2H100_full.yml

@@ -0,0 +1,194 @@
+# Run integer benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
+name: Integer 2xH100 benchmarks
+
+on:
+  workflow_dispatch:
+  schedule:
+    # Weekly benchmarks will be triggered each Saturday at 1a.m.
+    - cron: '0 1 * * 6'
+
+env:
+  CARGO_TERM_COLOR: always
+  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
+jobs:
+  setup-instance:
+    name: Setup instance (cuda-integer-full-2-gpu-benchmarks)
+    runs-on: ubuntu-latest
+    if: github.event_name != 'schedule' ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: hyperstack
+          profile: 2-h100
+
+  cuda-integer-full-2-gpu-benchmarks:
+    name: Execute 2xH100 integer benchmarks
+    needs: setup-instance
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    timeout-minutes: 1440 # 24 hours
+    continue-on-error: true
+    strategy:
+      fail-fast: false
+      max-parallel: 1
+      matrix:
+        command: [integer_multi_bit]
+        op_flavor: [default]
+        # explicit include-based build matrix, of known valid options
+        include:
+          - os: ubuntu-22.04
+            cuda: "12.2"
+            gcc: 11
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
+    steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(date --iso-8601=seconds)";
+            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
+            echo "COMMIT_HASH=$(git describe --tags --dirty)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        with:
+          toolchain: nightly
+
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CUDA_PATH=$CUDA_PATH";
+            echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH";
+            echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc";
+          } >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+          } >> "${GITHUB_ENV}"
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
+      - name: Run benchmarks with AVX512
+        run: |
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_${{ matrix.command }}_gpu
+
+      - name: Parse results
+        run: |
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware "n3-H100x2" \
+          --backend gpu \
+          --project-version "${{ env.COMMIT_HASH }}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${{ env.COMMIT_DATE }}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --throughput
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
+        with:
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
+
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-integer-full-2-gpu-benchmarks ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-integer-full-2-gpu-benchmarks.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ needs.cuda-integer-full-2-gpu-benchmarks.result }}
+          SLACK_MESSAGE: "Integer GPU 2xH100 benchmarks finished with status: ${{ needs.cuda-integer-full-2-gpu-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (cuda-integer-full-2-gpu-benchmarks)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, cuda-integer-full-2-gpu-benchmarks ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cuda-integer-full-2-gpu-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_integer_common.yml

@@ -1,256 +0,0 @@
-# Run integer benchmarks on CUDA instance and return parsed results to Slab CI bot.
-name: Cuda benchmarks - common
-
-on:
-  workflow_call:
-    inputs:
-      backend:
-        type: string
-        default: hyperstack
-      profile:
-        type: string
-        required: true
-      hardware_name:
-        type: string
-        required: true
-      command: # Use a comma separated values to generate an array
-        type: string
-        required: true
-      op_flavor: # Use a comma separated values to generate an array
-        type: string
-        required: true
-      bench_type:
-        type: string
-        default: latency
-      all_precisions:
-        type: boolean
-        default: false
-    secrets:
-      FHE_ACTIONS_TOKEN:
-        required: true
-      SLAB_ACTION_TOKEN:
-        required: true
-      SLAB_BASE_URL:
-        required: true
-      SLAB_URL:
-        required: true
-      JOB_SECRET:
-        required: true
-      SLACK_CHANNEL:
-        required: true
-      BOT_USERNAME:
-        required: true
-      SLACK_WEBHOOK:
-        required: true
-
-env:
-  CARGO_TERM_COLOR: always
-  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-  FAST_BENCH: TRUE
-
-jobs:
-  prepare-matrix:
-    name: Prepare operations matrix
-    runs-on: ubuntu-latest
-    outputs:
-      command: ${{ steps.set_command.outputs.command }}
-      op_flavor: ${{ steps.set_op_flavor.outputs.op_flavor }}
-      bench_type: ${{ steps.set_bench_type.outputs.bench_type }}
-    steps:
-      - name: Set single command
-        if: ${{ !contains(inputs.command, ',')}}
-        run: |
-          echo "COMMAND=[\"${{ inputs.command }}\"]" >> "${GITHUB_ENV}"
-
-      - name: Set multiple commands
-        if: ${{ contains(inputs.command, ',')}}
-        run: |
-          PARSED_COMMAND=$(echo "${{ inputs.command }}" | sed 's/[[:space:]]*,[[:space:]]*/\\", \\"/g')
-          echo "COMMAND=[\"${PARSED_COMMAND}\"]" >> "${GITHUB_ENV}"
-
-      - name: Set single operations flavor
-        if: ${{ !contains(inputs.op_flavor, ',')}}
-        run: |
-          echo "OP_FLAVOR=[\"${{ inputs.op_flavor }}\"]" >> "${GITHUB_ENV}"
-
-      - name: Set multiple operations flavors
-        if: ${{ contains(inputs.op_flavor, ',')}}
-        run: |
-          PARSED_OP_FLAVOR=$(echo "${{ inputs.op_flavor }}" | sed 's/[[:space:]]*,[[:space:]]*/", "/g')
-          echo "OP_FLAVOR=[\"${PARSED_OP_FLAVOR}\"]" >> "${GITHUB_ENV}"
-
-      - name: Set benchmark types
-        run: |
-          if [[ "${{ inputs.bench_type }}" == "both" ]]; then
-            echo "BENCH_TYPE=[\"latency\", \"throughput\"]" >> "${GITHUB_ENV}"
-          else
-            echo "BENCH_TYPE=[\"${{ inputs.bench_type }}\"]" >> "${GITHUB_ENV}"
-          fi
-
-      - name: Set command output
-        id: set_command
-        run: |
-          echo "command=${{ toJSON(env.COMMAND) }}" >> "${GITHUB_OUTPUT}"
-
-      - name: Set operation flavor output
-        id: set_op_flavor
-        run: |
-          echo "op_flavor=${{ toJSON(env.OP_FLAVOR) }}" >> "${GITHUB_OUTPUT}"
-
-      - name: Set benchmark types output
-        id: set_bench_type
-        run: |
-          echo "bench_type=${{ toJSON(env.BENCH_TYPE) }}" >> "${GITHUB_OUTPUT}"
-
-  setup-instance:
-    name: Setup instance (cuda-${{ inputs.profile }}-benchmarks)
-    needs: prepare-matrix
-    runs-on: ubuntu-latest
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: ${{ inputs.backend }}
-          profile: ${{ inputs.profile }}
-
-  cuda-benchmarks:
-    name: Cuda benchmarks (${{ inputs.profile }})
-    needs: [ prepare-matrix, setup-instance ]
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    timeout-minutes: 1440 # 24 hours
-    continue-on-error: true
-    strategy:
-      fail-fast: false
-      max-parallel: 1
-      matrix:
-        command: ${{ fromJSON(needs.prepare-matrix.outputs.command) }}
-        op_flavor: ${{ fromJSON(needs.prepare-matrix.outputs.op_flavor) }}
-        bench_type: ${{ fromJSON(needs.prepare-matrix.outputs.bench_type) }}
-        # explicit include-based build matrix, of known valid options
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11
-    steps:
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
-
-      - name: Get benchmark details
-        run: |
-          {
-            echo "BENCH_DATE=$(date --iso-8601=seconds)";
-            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
-            echo "COMMIT_HASH=$(git describe --tags --dirty)";
-          } >> "${GITHUB_ENV}"
-
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
-        with:
-          toolchain: nightly
-
-      - name: Should run benchmarks with all precisions
-        if: inputs.all_precisions
-        run: |
-          echo "FAST_BENCH=FALSE" >> "${GITHUB_ENV}"
-
-      - name: Run benchmarks
-        run: |
-          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} BENCH_TYPE=${{ matrix.bench_type }} bench_${{ matrix.command }}_gpu
-
-      - name: Parse results
-        run: |
-          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
-          --database tfhe_rs \
-          --hardware "${{ inputs.hardware_name }}" \
-          --backend gpu \
-          --project-version "${{ env.COMMIT_HASH }}" \
-          --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
-          --bench-date "${{ env.BENCH_DATE }}" \
-          --walk-subdirs \
-          --name-suffix avx512 \
-          --bench-type ${{ matrix.bench_type }}
-
-      - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
-        with:
-          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}_${{ inputs.profile }}
-          path: ${{ env.RESULTS_FILENAME }}
-
-      - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          repository: zama-ai/slab
-          path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Send data to Slab
-        shell: bash
-        run: |
-          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
-          --slab-url "${{ secrets.SLAB_URL }}"
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-benchmarks ]
-    runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-benchmarks.result != 'skipped' && failure() }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ needs.cuda-benchmarks.result }}
-          SLACK_MESSAGE: "Cuda benchmarks (${{ inputs.profile }}) finished with status: ${{ needs.cuda-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-${{ inputs.profile }}-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-benchmarks, slack-notify ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-${{ inputs.profile }}-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_integer_full.yml

@@ -0,0 +1,200 @@
+# Run all integer benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
+name: Integer GPU full benchmarks
+
+on:
+  workflow_dispatch:
+  schedule:
+    # Weekly benchmarks will be triggered each Saturday at 1a.m.
+    - cron: '0 1 * * 6'
+
+env:
+  CARGO_TERM_COLOR: always
+  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
+jobs:
+  setup-instance:
+    name: Setup instance (cuda-integer-full-benchmarks)
+    runs-on: ubuntu-latest
+    if: github.event_name != 'schedule' ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: hyperstack
+          profile: single-h100
+
+  cuda-integer-full-benchmarks:
+    name: Execute GPU integer benchmarks for all operations flavor
+    needs: setup-instance
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    timeout-minutes: 1440 # 24 hours
+    continue-on-error: true
+    strategy:
+      fail-fast: false
+      max-parallel: 1
+      matrix:
+        command: [integer, integer_multi_bit]
+        op_flavor: [default]
+        # explicit include-based build matrix, of known valid options
+        include:
+          - os: ubuntu-22.04
+            cuda: "12.2"
+            gcc: 11
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
+    steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(date --iso-8601=seconds)";
+            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
+            echo "COMMIT_HASH=$(git describe --tags --dirty)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        with:
+          toolchain: nightly
+
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CUDA_PATH=$CUDA_PATH";
+            echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH";
+            echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc";
+          } >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+          } >> "${GITHUB_ENV}"
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
+      - name: Run benchmarks with AVX512
+        run: |
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_${{ matrix.command }}_gpu
+
+      # Run these benchmarks only once
+      - name: Run compression benchmarks with AVX512
+        if: matrix.op_flavor == 'default' && matrix.command == 'integer'
+        run: |
+          make bench_integer_compression_gpu
+
+      - name: Parse results
+        run: |
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware "n3-H100x1" \
+          --backend gpu \
+          --project-version "${{ env.COMMIT_HASH }}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${{ env.COMMIT_DATE }}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --throughput
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
+        with:
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
+
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-integer-full-benchmarks ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-integer-full-benchmarks.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ needs.cuda-integer-full-benchmarks.result }}
+          SLACK_MESSAGE: "Integer GPU full benchmarks finished with status: ${{ needs.cuda-integer-full-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (cuda-integer-full-benchmarks)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, cuda-integer-full-benchmarks, slack-notify ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cuda-integer-full-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_integer_multi_bit.yml

@@ -0,0 +1,224 @@
+# Run integer benchmarks with multi-bit cryptographic parameters on an instance and return parsed results to Slab CI bot.
+name: Integer GPU Multi-bit benchmarks
+
+on:
+  workflow_dispatch:
+    inputs:
+      all_precisions:
+        description: "Run all precisions"
+        type: boolean
+        default: false
+      fast_default:
+        description: "Run only deduplicated default operations without scalar variants"
+        type: boolean
+        default: false
+
+  schedule:
+    # Weekly benchmarks will be triggered each Saturday at 1a.m.
+    - cron: '0 1 * * 6'
+
+env:
+  CARGO_TERM_COLOR: always
+  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
+  PARSE_INTEGER_BENCH_CSV_FILE: tfhe_rs_integer_benches_${{ github.sha }}.csv
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  FAST_BENCH: TRUE
+  BENCH_OP_FLAVOR: default
+
+jobs:
+  setup-instance:
+    name: Setup instance (cuda-integer-multi-bit-benchmarks)
+    runs-on: ubuntu-latest
+    if: github.event_name != 'schedule' ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: hyperstack
+          profile: single-h100
+
+  cuda-integer-multi-bit-benchmarks:
+    name: Execute GPU integer multi-bit benchmarks
+    needs: setup-instance
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    timeout-minutes: 1440 # 24 hours
+    strategy:
+      fail-fast: false
+      # explicit include-based build matrix, of known valid options
+      matrix:
+        include:
+          - os: ubuntu-22.04
+            cuda: "12.2"
+            gcc: 11
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
+    steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(date --iso-8601=seconds)";
+            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
+            echo "COMMIT_HASH=$(git describe --tags --dirty)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        with:
+          toolchain: nightly
+
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CUDA_PATH=$CUDA_PATH";
+            echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH";
+            echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc";
+          } >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+          } >> "${GITHUB_ENV}"
+
+      - name: Should run benchmarks with all precisions
+        if: inputs.all_precisions
+        run: |
+          echo "FAST_BENCH=FALSE" >> "${GITHUB_ENV}"
+
+      - name: Should run fast subset benchmarks
+        if: inputs.fast_default
+        run: |
+          echo "BENCH_OP_FLAVOR=fast_default" >> "${GITHUB_ENV}"
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
+      - name: Run multi-bit benchmarks with AVX512
+        run: |
+          make bench_unsigned_integer_multi_bit_gpu
+
+      - name: Parse benchmarks to csv
+        run: |
+          make PARSE_INTEGER_BENCH_CSV_FILE=${{ env.PARSE_INTEGER_BENCH_CSV_FILE }} \
+            parse_integer_benches
+
+      - name: Upload csv results artifact
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
+        with:
+          name: ${{ github.sha }}_csv_integer
+          path: ${{ env.PARSE_INTEGER_BENCH_CSV_FILE }}
+
+      - name: Parse results
+        run: |
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware "n3-H100x1" \
+          --backend gpu \
+          --project-version "${{ env.COMMIT_HASH }}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${{ env.COMMIT_DATE }}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --throughput
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
+        with:
+          name: ${{ github.sha }}_integer
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
+
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-integer-multi-bit-benchmarks ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-integer-multi-bit-benchmarks.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ needs.cuda-integer-multi-bit-benchmarks.result }}
+          SLACK_MESSAGE: "Integer GPU multi-bit benchmarks finished with status: ${{ needs.cuda-integer-multi-bit-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (cuda-integer-full-benchmarks)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, cuda-integer-multi-bit-benchmarks, slack-notify ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cuda-integer-multi-bit-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_integer_multi_bit_multi_gpu.yml

@@ -0,0 +1,214 @@
+# Run 64-bit multi-bit integer benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
+name: Integer multi GPU Multi-bit benchmarks
+
+on:
+  workflow_dispatch:
+    inputs:
+      all_precisions:
+        description: "Run all precisions"
+        type: boolean
+        default: false
+      fast_default:
+        description: "Run only deduplicated default operations without scalar variants"
+        type: boolean
+        default: false
+
+  schedule:
+    # Weekly benchmarks will be triggered each Saturday at 1a.m.
+    - cron: '0 1 * * 6'
+
+env:
+  CARGO_TERM_COLOR: always
+  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  FAST_BENCH: TRUE
+  BENCH_OP_FLAVOR: default
+
+jobs:
+  setup-instance:
+    name: Setup instance (cuda-integer-multi-bit-multi-gpu-benchmarks)
+    runs-on: ubuntu-latest
+    if: ${{ (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs') ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
+      github.event_name == 'workflow_dispatch' }}
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: hyperstack
+          profile: multi-h100
+
+  cuda-integer-multi-bit-multi-gpu-benchmarks:
+    name: Execute multi GPU integer multi-bit benchmarks
+    needs: setup-instance
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    timeout-minutes: 1440 # 24 hours
+    continue-on-error: true
+    strategy:
+      fail-fast: false
+      max-parallel: 1
+      matrix:
+        include:
+          - os: ubuntu-22.04
+            cuda: "12.2"
+            gcc: 11
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
+    steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(date --iso-8601=seconds)";
+            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
+            echo "COMMIT_HASH=$(git describe --tags --dirty)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        with:
+          toolchain: nightly
+
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CUDA_PATH=$CUDA_PATH";
+            echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH";
+            echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc";
+          } >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+          } >> "${GITHUB_ENV}"
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Should run benchmarks with all precisions
+        if: inputs.all_precisions
+        run: |
+          echo "FAST_BENCH=FALSE" >> "${GITHUB_ENV}"
+
+      - name: Should run fast subset benchmarks
+        if: inputs.fast_default
+        run: |
+          echo "BENCH_OP_FLAVOR=fast_default" >> "${GITHUB_ENV}"
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
+      - name: Run multi-bit benchmarks with AVX512
+        run: |
+          make bench_unsigned_integer_multi_bit_gpu
+
+      - name: Parse results
+        run: |
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware "n3-H100x8" \
+          --backend gpu \
+          --project-version "${{ env.COMMIT_HASH }}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${{ env.COMMIT_DATE }}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --throughput
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
+        with:
+          name: ${{ github.sha }}_integer
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
+
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-integer-multi-bit-multi-gpu-benchmarks ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-integer-multi-bit-multi-gpu-benchmarks.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ needs.cuda-integer-multi-bit-multi-gpu-benchmarks.result }}
+          SLACK_MESSAGE: "Integer multi GPU multi-bit benchmarks finished with status: ${{ needs.cuda-integer-multi-bit-multi-gpu-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (cuda-integer-multi-bit-multi-gpu-benchmarks)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, cuda-integer-multi-bit-multi-gpu-benchmarks ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cuda-integer-multi-bit-multi-gpu-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_integer_multi_gpu_full.yml

@@ -0,0 +1,194 @@
+# Run all integer benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
+name: Integer multi GPU full benchmarks
+
+on:
+  workflow_dispatch:
+  schedule:
+    # Weekly benchmarks will be triggered each Saturday at 1a.m.
+    - cron: '0 1 * * 6'
+
+env:
+  CARGO_TERM_COLOR: always
+  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
+jobs:
+  setup-instance:
+    name: Setup instance (cuda-integer-full-multi-gpu-benchmarks)
+    runs-on: ubuntu-latest
+    if: github.event_name != 'schedule' ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: hyperstack
+          profile: multi-h100
+
+  cuda-integer-full-multi-gpu-benchmarks:
+    name: Execute multi GPU integer benchmarks
+    needs: setup-instance
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    timeout-minutes: 1440 # 24 hours
+    continue-on-error: true
+    strategy:
+      fail-fast: false
+      max-parallel: 1
+      matrix:
+        command: [integer_multi_bit]
+        op_flavor: [default]
+        # explicit include-based build matrix, of known valid options
+        include:
+          - os: ubuntu-22.04
+            cuda: "12.2"
+            gcc: 11
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
+    steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(date --iso-8601=seconds)";
+            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
+            echo "COMMIT_HASH=$(git describe --tags --dirty)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        with:
+          toolchain: nightly
+
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CUDA_PATH=$CUDA_PATH";
+            echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH";
+            echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc";
+          } >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+          } >> "${GITHUB_ENV}"
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
+      - name: Run benchmarks with AVX512
+        run: |
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_${{ matrix.command }}_gpu
+
+      - name: Parse results
+        run: |
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware "n3-H100x8" \
+          --backend gpu \
+          --project-version "${{ env.COMMIT_HASH }}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${{ env.COMMIT_DATE }}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --throughput
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
+        with:
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
+
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-integer-full-multi-gpu-benchmarks ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-integer-full-multi-gpu-benchmarks.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ needs.cuda-integer-full-multi-gpu-benchmarks.result }}
+          SLACK_MESSAGE: "Integer GPU full benchmarks finished with status: ${{ needs.cuda-integer-full-multi-gpu-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (cuda-integer-full-multi-gpu-benchmarks)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, cuda-integer-full-multi-gpu-benchmarks ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cuda-integer-full-multi-gpu-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_integer_weekly.yml

@@ -1,60 +0,0 @@
-# Run CUDA benchmarks on multiple Hyperstack VMs and return parsed results to Slab CI bot.
-name: Cuda weekly benchmarks
-
-on:
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 1a.m.
-    - cron: '0 1 * * 6'
-
-jobs:
-  run-benchmarks-1-h100:
-    name: Run benchmarks (1xH100)
-    if: github.repository == 'zama-ai/tfhe-rs'
-    uses: ./.github/workflows/benchmark_gpu_integer_common.yml
-    with:
-      profile: single-h100
-      hardware_name: n3-H100x1
-      command: integer,integer_multi_bit
-      op_flavor: default
-      bench_type: latency
-      all_precisions: true
-    secrets: inherit
-
-  run-benchmarks-2-h100:
-    name: Run benchmarks (2xH100)
-    if: github.repository == 'zama-ai/tfhe-rs'
-    uses: ./.github/workflows/benchmark_gpu_integer_common.yml
-    with:
-      profile: 2-h100
-      hardware_name: n3-H100x2
-      command: integer_multi_bit
-      op_flavor: default
-      bench_type: latency
-      all_precisions: true
-    secrets: inherit
-
-  run-benchmarks-8-h100:
-    name: Run benchmarks (8xH100)
-    if: github.repository == 'zama-ai/tfhe-rs'
-    uses: ./.github/workflows/benchmark_gpu_integer_common.yml
-    with:
-      profile: multi-h100
-      hardware_name: n3-H100x8
-      command: integer_multi_bit
-      op_flavor: default
-      bench_type: latency
-      all_precisions: true
-    secrets: inherit
-
-  run-benchmarks-l40:
-    name: Run benchmarks (L40)
-    if: github.repository == 'zama-ai/tfhe-rs'
-    uses: ./.github/workflows/benchmark_gpu_integer_common.yml
-    with:
-      profile: l40
-      hardware_name: n3-L40x1
-      command: integer_multi_bit,integer_compression,pbs,ks
-      op_flavor: default
-      bench_type: latency
-      all_precisions: true
-    secrets: inherit

.github/workflows/benchmark_gpu_l40.yml

@@ -0,0 +1,206 @@
+# Run benchmarks on an L40 VM and return parsed results to Slab CI bot.
+name: Cuda benchmarks (L40)
+
+on:
+  workflow_dispatch:
+  schedule:
+    # Weekly benchmarks will be triggered each Saturday at 1a.m.
+    - cron: '0 1 * * 6'
+
+env:
+  CARGO_TERM_COLOR: always
+  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
+jobs:
+  setup-instance:
+    name: Setup instance (cuda-l40-benchmarks)
+    runs-on: ubuntu-latest
+    if: github.event_name != 'schedule' ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: hyperstack
+          profile: l40 
+
+  cuda-l40-benchmarks:
+    name: Cuda benchmarks (L40)
+    needs: setup-instance
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    timeout-minutes: 1440 # 24 hours
+    continue-on-error: true
+    strategy:
+      fail-fast: false
+      max-parallel: 1
+      matrix:
+        command: [integer_multi_bit]
+        op_flavor: [default]
+        # explicit include-based build matrix, of known valid options
+        include:
+          - os: ubuntu-22.04
+            cuda: "12.2"
+            gcc: 11
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
+    steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(date --iso-8601=seconds)";
+            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
+            echo "COMMIT_HASH=$(git describe --tags --dirty)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        with:
+          toolchain: nightly
+
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CUDA_PATH=$CUDA_PATH";
+            echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH";
+            echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc";
+          } >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+          } >> "${GITHUB_ENV}"
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
+      - name: Run benchmarks with AVX512
+        run: |
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_${{ matrix.command }}_gpu
+
+      - name: Run compression benchmarks with AVX512
+        run: |
+          make bench_integer_compression_gpu
+
+      - name: Run PBS benchmarks 
+        run: |
+          make bench_pbs_gpu
+
+      - name: Run KS benchmarks 
+        run: |
+          make bench_ks_gpu
+
+      - name: Parse results
+        run: |
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware "n3-L40x1" \
+          --backend gpu \
+          --project-version "${{ env.COMMIT_HASH }}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${{ env.COMMIT_DATE }}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --throughput
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
+        with:
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
+
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-l40-benchmarks ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-l40-benchmarks.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ needs.cuda-l40-benchmarks.result }}
+          SLACK_MESSAGE: "Cuda benchmarks (L40) finished with status: ${{ needs.cuda-l40-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (cuda-l40-benchmarks)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, cuda-l40-benchmarks, slack-notify ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cuda-l40-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_integer.yml

@@ -8,14 +8,6 @@ on:
         description: "Run all precisions"
         type: boolean
         default: false
-      bench_type:
-        description: "Benchmarks type"
-        type: choice
-        default: latency
-        options:
-          - latency
-          - throughput
-          - both
 
   schedule:
     # Weekly benchmarks will be triggered each Saturday at 1a.m.
@@ -44,10 +36,10 @@ jobs:
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
     outputs:
       op_flavor: ${{ steps.set_op_flavor.outputs.op_flavor }}
-      bench_type: ${{ steps.set_bench_type.outputs.bench_type }}
     steps:
       - name: Weekly benchmarks
-        if: github.event.schedule == '0 1 * * 6'
+        if: github.event_name == 'workflow_dispatch' ||
+          github.event.schedule == '0 1 * * 6'
         run: |
           echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
 
@@ -56,31 +48,11 @@ jobs:
         run: |
           echo "OP_FLAVOR=[\"default\", \"smart\", \"unchecked\", \"misc\"]" >> "${GITHUB_ENV}"
 
-      - name: Set benchmark types
-        if: github.event_name == 'workflow_dispatch'
-        run: |
-          echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
-          if [[ "${{ inputs.bench_type }}" == "both" ]]; then
-            echo "BENCH_TYPE=[\"latency\", \"throughput\"]" >> "${GITHUB_ENV}"
-          else
-            echo "BENCH_TYPE=[\"${{ inputs.bench_type }}\"]" >> "${GITHUB_ENV}"
-          fi
-
-      - name: Default benchmark type
-        if: github.event_name != 'workflow_dispatch'
-        run: |
-          echo "BENCH_TYPE=[\"latency\"]" >> "${GITHUB_ENV}"
-
-      - name: Set operation flavor output
-        id: set_op_flavor
-        run: |
+      -  name: Set operation flavor output
+         id: set_op_flavor
+         run: |
           echo "op_flavor=${{ toJSON(env.OP_FLAVOR) }}" >> "${GITHUB_OUTPUT}"
 
-      - name: Set benchmark types output
-        id: set_bench_type
-        run: |
-          echo "bench_type=${{ toJSON(env.BENCH_TYPE) }}" >> "${GITHUB_OUTPUT}"
-
   setup-instance:
     name: Setup instance (integer-benchmarks)
     needs: prepare-matrix
@@ -90,7 +62,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -113,10 +85,9 @@ jobs:
       matrix:
         command: [ integer, integer_multi_bit]
         op_flavor: ${{ fromJson(needs.prepare-matrix.outputs.op_flavor) }}
-        bench_type: ${{ fromJSON(needs.prepare-matrix.outputs.bench_type) }}
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
@@ -129,13 +100,18 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
       - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           repository: zama-ai/slab
           path: slab
@@ -148,13 +124,13 @@ jobs:
 
       - name: Run benchmarks with AVX512
         run: |
-          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} BENCH_TYPE=${{ matrix.bench_type }} bench_${{ matrix.command }}
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_${{ matrix.command }}
 
-      # Run these benchmarks only once per benchmark type
+      # Run these benchmarks only once
       - name: Run compression benchmarks with AVX512
         if: matrix.op_flavor == 'default' && matrix.command == 'integer'
         run: |
-          make BENCH_TYPE=${{ matrix.bench_type }} bench_integer_compression
+          make bench_integer_compression
 
       - name: Parse results
         run: |
@@ -167,12 +143,12 @@ jobs:
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
           --name-suffix avx512 \
-          --bench-type ${{ matrix.bench_type }}
+          --throughput
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
         with:
-          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}_${{ matrix.bench_type }}
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Send data to Slab
@@ -184,7 +160,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Integer full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -197,7 +173,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -208,7 +184,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (integer-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_shortint.yml

@@ -56,7 +56,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -79,7 +79,7 @@ jobs:
         op_flavor: ${{ fromJson(needs.prepare-matrix.outputs.op_flavor) }}
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
@@ -92,13 +92,18 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
       - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           repository: zama-ai/slab
           path: slab
@@ -120,7 +125,8 @@ jobs:
           --commit-date "${COMMIT_DATE}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
-          --name-suffix avx512
+          --name-suffix avx512 \
+          --throughput
 
       # This small benchmark needs to be executed only once.
       - name: Measure key sizes
@@ -132,11 +138,11 @@ jobs:
         if: matrix.op_flavor == 'default'
         run: |
           python3 ./ci/benchmark_parser.py tfhe/shortint_key_sizes.csv ${{ env.RESULTS_FILENAME }} \
-          --object-sizes \
+          --key-sizes \
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
         with:
           name: ${{ github.sha }}_shortint_${{ matrix.op_flavor }}
           path: ${{ env.RESULTS_FILENAME }}
@@ -150,7 +156,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Shortint full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -163,7 +169,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -174,7 +180,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (shortint-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_signed_integer.yml

@@ -8,14 +8,6 @@ on:
         description: "Run all precisions"
         type: boolean
         default: false
-      bench_type:
-        description: "Benchmarks type"
-        type: choice
-        default: latency
-        options:
-          - latency
-          - throughput
-          - both
 
   schedule:
     # Weekly benchmarks will be triggered each Saturday at 1a.m.
@@ -44,10 +36,10 @@ jobs:
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
     outputs:
       op_flavor: ${{ steps.set_op_flavor.outputs.op_flavor }}
-      bench_type: ${{ steps.set_bench_type.outputs.bench_type }}
     steps:
       - name: Weekly benchmarks
-        if: github.event.schedule == '0 1 * * 6'
+        if: github.event_name == 'workflow_dispatch' ||
+          github.event.schedule == '0 1 * * 6'
         run: |
           echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
 
@@ -56,31 +48,11 @@ jobs:
         run: |
           echo "OP_FLAVOR=[\"default\", \"unchecked\"]" >> "${GITHUB_ENV}"
 
-      - name: Set benchmark types
-        if: github.event_name == 'workflow_dispatch'
-        run: |
-          echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
-          if [[ "${{ inputs.bench_type }}" == "both" ]]; then
-            echo "BENCH_TYPE=[\"latency\", \"throughput\"]" >> "${GITHUB_ENV}"
-          else
-            echo "BENCH_TYPE=[\"${{ inputs.bench_type }}\"]" >> "${GITHUB_ENV}"
-          fi
-
-      - name: Default benchmark type
-        if: github.event_name != 'workflow_dispatch'
-        run: |
-          echo "BENCH_TYPE=[\"latency\"]" >> "${GITHUB_ENV}"
-
       - name: Set operation flavor output
         id: set_op_flavor
         run: |
           echo "op_flavor=${{ toJSON(env.OP_FLAVOR) }}" >> "${GITHUB_OUTPUT}"
 
-      - name: Set benchmark types output
-        id: set_bench_type
-        run: |
-          echo "bench_type=${{ toJSON(env.BENCH_TYPE) }}" >> "${GITHUB_OUTPUT}"
-
   setup-instance:
     name: Setup instance (signed-integer-benchmarks)
     needs: prepare-matrix
@@ -90,7 +62,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -112,11 +84,10 @@ jobs:
       max-parallel: 1
       matrix:
         command: [ integer, integer_multi_bit ]
-        op_flavor: ${{ fromJSON(needs.prepare-matrix.outputs.op_flavor) }}
-        bench_type: ${{ fromJSON(needs.prepare-matrix.outputs.bench_type) }}
+        op_flavor: [ default, unchecked ]
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
@@ -129,13 +100,18 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
       - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           repository: zama-ai/slab
           path: slab
@@ -148,7 +124,7 @@ jobs:
 
       - name: Run benchmarks with AVX512
         run: |
-          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} BENCH_TYPE=${{ matrix.bench_type }} bench_signed_${{ matrix.command }}
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_signed_${{ matrix.command }}
 
       - name: Parse results
         run: |
@@ -161,12 +137,12 @@ jobs:
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
           --name-suffix avx512 \
-          --bench-type ${{ matrix.bench_type }}
+          --throughput
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
         with:
-          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}_${{ matrix.bench_type }}
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Send data to Slab
@@ -178,7 +154,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Signed integer full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -191,7 +167,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -202,7 +178,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (signed-integer-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_tfhe_fft.yml

@@ -1,143 +0,0 @@
-# Run FFT benchmarks on an AWS instance and return parsed results to Slab CI bot.
-name: FFT benchmarks
-
-env:
-  CARGO_TERM_COLOR: always
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
-  RUST_BACKTRACE: "full"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-
-on:
-  workflow_dispatch:
-  push:
-    branches:
-      - "main"
-    paths:
-      - tfhe-fft/**
-      - .github/workflows/benchmark_tfhe_fft.yml
-  schedule:
-    # Job will be triggered each Thursday at 11p.m.
-    - cron: '0 23 * * 4'
-
-jobs:
-  setup-ec2:
-    name: Setup EC2 instance (fft-benchmarks)
-    runs-on: ubuntu-latest
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: bench
-
-  fft-benchmarks:
-    name: Execute FFT benchmarks in EC2
-    needs: setup-ec2
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: true
-    runs-on: ${{ needs.setup-ec2.outputs.runner-name }}
-    steps:
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          fetch-depth: 0
-
-      - name: Get benchmark details
-        run: |
-          {
-            echo "BENCH_DATE=$(date --iso-8601=seconds)";
-            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
-            echo "COMMIT_HASH=$(git describe --tags --dirty)";
-          } >> "${GITHUB_ENV}"
-
-      - name: Install rust
-        uses: actions-rs/toolchain@16499b5e05bf2e26879000db0c1d13f7e13fa3af
-        with:
-          toolchain: nightly
-          override: true
-
-      - name: Run benchmarks with AVX512
-        run: |
-          make bench_fft
-
-      - name: Parse AVX512 results
-        run: |
-          python3 ./ci/fft_benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
-          --database concrete_fft \
-          --hardware "hpc7a.96xlarge" \
-          --project-version "${{ env.COMMIT_HASH }}" \
-          --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
-          --bench-date "${{ env.BENCH_DATE }}" \
-          --name-suffix avx512
-
-      - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
-        with:
-          name: ${{ github.sha }}_fft
-          path: ${{ env.RESULTS_FILENAME }}
-
-      - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          repository: zama-ai/slab
-          path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Send data to Slab
-        shell: bash
-        run: |
-          echo "Computing HMac on downloaded artifact"
-          SIGNATURE="$(slab/scripts/hmac_calculator.sh ${{ env.RESULTS_FILENAME }} '${{ secrets.JOB_SECRET }}')"
-          echo "Sending results to Slab..."
-          curl -v -k \
-          -H "Content-Type: application/json" \
-          -H "X-Slab-Repository: ${{ github.repository }}" \
-          -H "X-Slab-Command: store_data_v2" \
-          -H "X-Hub-Signature-256: sha256=${SIGNATURE}" \
-          -d @${{ env.RESULTS_FILENAME }} \
-          ${{ secrets.SLAB_URL }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "tfhe-fft benchmarks failed. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-ec2:
-    name: Teardown EC2 instance (fft-benchmarks)
-    if: ${{ always() && needs.setup-ec2.result != 'skipped' }}
-    needs: [ setup-ec2, fft-benchmarks ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-ec2.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "EC2 teardown (fft-benchmarks) failed. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_tfhe_ntt.yml

@@ -1,143 +0,0 @@
-# Run NTT benchmarks on an AWS instance and return parsed results to Slab CI bot.
-name: NTT benchmarks
-
-env:
-  CARGO_TERM_COLOR: always
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
-  RUST_BACKTRACE: "full"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-
-on:
-  workflow_dispatch:
-  push:
-    branches:
-      - "main"
-    paths:
-      - tfhe-ntt/**
-      - .github/workflows/benchmark_tfhe_ntt.yml
-  schedule:
-    # Job will be triggered each Friday at 11p.m.
-    - cron: "0 23 * * 5"
-
-jobs:
-  setup-ec2:
-    name: Setup EC2 instance (ntt-benchmarks)
-    runs-on: ubuntu-latest
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: bench
-
-  ntt-benchmarks:
-    name: Execute NTT benchmarks in EC2
-    needs: setup-ec2
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: true
-    runs-on: ${{ needs.setup-ec2.outputs.runner-name }}
-    steps:
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          fetch-depth: 0
-
-      - name: Get benchmark details
-        run: |
-          {
-            echo "BENCH_DATE=$(date --iso-8601=seconds)";
-            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
-            echo "COMMIT_HASH=$(git describe --tags --dirty)";
-          } >> "${GITHUB_ENV}"
-
-      - name: Install rust
-        uses: actions-rs/toolchain@16499b5e05bf2e26879000db0c1d13f7e13fa3af
-        with:
-          toolchain: nightly
-          override: true
-
-      - name: Run benchmarks
-        run: |
-          make bench_ntt
-
-      - name: Parse results
-        run: |
-          python3 ./ci/ntt_benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
-          --database concrete_ntt \
-          --hardware "hpc7a.96xlarge" \
-          --project-version "${{ env.COMMIT_HASH }}" \
-          --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
-          --bench-date "${{ env.BENCH_DATE }}" \
-          --name-suffix avx512
-
-      - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
-        with:
-          name: ${{ github.sha }}_ntt
-          path: ${{ env.RESULTS_FILENAME }}
-
-      - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          repository: zama-ai/slab
-          path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Send data to Slab
-        shell: bash
-        run: |
-          echo "Computing HMac on downloaded artifact"
-          SIGNATURE="$(slab/scripts/hmac_calculator.sh ${{ env.RESULTS_FILENAME }} '${{ secrets.JOB_SECRET }}')"
-          echo "Sending results to Slab..."
-          curl -v -k \
-          -H "Content-Type: application/json" \
-          -H "X-Slab-Repository: ${{ github.repository }}" \
-          -H "X-Slab-Command: store_data_v2" \
-          -H "X-Hub-Signature-256: sha256=${SIGNATURE}" \
-          -d @${{ env.RESULTS_FILENAME }} \
-          ${{ secrets.SLAB_URL }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "tfhe-ntt benchmarks failed. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-ec2:
-    name: Teardown EC2 instance (ntt-benchmarks)
-    if: ${{ always() && needs.setup-ec2.result != 'skipped' }}
-    needs: [setup-ec2, ntt-benchmarks]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-ec2.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "EC2 teardown (ntt-benchmarks) failed. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_tfhe_zk_pok.yml

@@ -1,172 +0,0 @@
-# Run benchmarks of the tfhe-zk-pok crate on an instance and return parsed results to Slab CI bot.
-name: tfhe-zk-pok benchmarks
-
-on:
-  workflow_dispatch:
-  push:
-    branches:
-      - main
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 3a.m.
-    - cron: '0 3 * * 6'
-env:
-  CARGO_TERM_COLOR: always
-  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
-  PARSE_INTEGER_BENCH_CSV_FILE: tfhe_rs_integer_benches_${{ github.sha }}.csv
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-
-jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    if: github.event_name == 'workflow_dispatch' ||
-      ((github.event_name == 'push' || github.event_name == 'schedule') && github.repository == 'zama-ai/tfhe-rs')
-    outputs:
-      zk_pok_changed: ${{ steps.changed-files.outputs.zk_pok_any_changed }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          fetch-depth: 0
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            zk_pok:
-              - tfhe-zk-pok/**
-              - .github/workflows/benchmark_tfhe_zk_pok.yml
-
-  setup-instance:
-    name: Setup instance (tfhe-zk-pok-benchmarks)
-    runs-on: ubuntu-latest
-    needs: should-run
-    if: github.event_name == 'workflow_dispatch' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
-      (github.event_name == 'push' &&
-      github.repository == 'zama-ai/tfhe-rs' &&
-      needs.should-run.outputs.zk_pok_changed == 'true')
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: bench
-
-  tfhe-zk-pok-benchmarks:
-    name: Execute tfhe-zk-pok benchmarks
-    if: needs.setup-instance.result != 'skipped'
-    needs: setup-instance
-    concurrency:
-      group: ${{ github.workflow }}_${{github.event_name}}_${{ github.ref }}${{ github.ref == 'refs/heads/main' && github.sha || '' }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    steps:
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Get benchmark details
-        run: |
-          {
-            echo "BENCH_DATE=$(date --iso-8601=seconds)";
-            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
-            echo "COMMIT_HASH=$(git describe --tags --dirty)";
-          } >> "${GITHUB_ENV}"
-
-      - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
-        with:
-          toolchain: nightly
-
-      - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          repository: zama-ai/slab
-          path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Run benchmarks
-        run: |
-          make bench_tfhe_zk_pok
-
-      - name: Parse results
-        run: |
-          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
-          --database tfhe_rs \
-          --crate tfhe-zk-pok \
-          --hardware "hpc7a.96xlarge" \
-          --backend cpu \
-          --project-version "${{ env.COMMIT_HASH }}" \
-          --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
-          --bench-date "${{ env.BENCH_DATE }}" \
-          --walk-subdirs \
-          --name-suffix avx512
-
-      - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
-        with:
-          name: ${{ github.sha }}_tfhe_zk_pok
-          path: ${{ env.RESULTS_FILENAME }}
-
-      - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          repository: zama-ai/slab
-          path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Send data to Slab
-        shell: bash
-        run: |
-          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
-          --slab-url "${{ secrets.SLAB_URL }}"
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "tfhe-zk-pok benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (tfhe-zk-pok-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, tfhe-zk-pok-benchmarks ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (tfhe-zk-pok-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_wasm_client.yml

@@ -33,20 +33,19 @@ jobs:
       wasm_bench: ${{ steps.changed-files.outputs.wasm_bench_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           since_last_remote_commit: true
           files_yaml: |
             wasm_bench:
               - tfhe/Cargo.toml
-              - tfhe-csprng/**
+              - concrete-csprng/**
               - tfhe-zk-pok/**
               - tfhe/src/**
               - '!tfhe/src/c_api/**'
@@ -65,7 +64,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -85,7 +84,7 @@ jobs:
         browser: [ chrome, firefox ]
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
@@ -98,42 +97,21 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
       - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: nightly
 
-      - name: Get Node version
-        run: |
-          echo "NODE_VERSION=$(make node_version)" >> "${GITHUB_ENV}"
-
-      - name: Node cache restoration
-        id: node-cache
-        uses: actions/cache/restore@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
-        with:
-          path: |
-            ~/.nvm
-            ~/.npm
-          key: node-${{ env.NODE_VERSION }}
-
-      - name: Install Node
-        if: steps.node-cache.outputs.cache-hit != 'true'
-        run: |
-          make install_node
-
-      - name: Node cache save
-        uses: actions/cache/save@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
-        if: steps.node-cache.outputs.cache-hit != 'true'
-        with:
-          path: |
-            ~/.nvm
-            ~/.npm
-          key: node-${{ env.NODE_VERSION }}
-
       - name: Install web resources
         run: |
+          make install_node
           make install_${{ matrix.browser }}_browser
-          make install_${{ matrix.browser }}_web_driver
+          make install_${{ matrix.browser }}_driver
 
       - name: Run benchmarks
         run: |
@@ -166,13 +144,13 @@ jobs:
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
         with:
           name: ${{ github.sha }}_wasm_${{ matrix.browser }}
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           repository: zama-ai/slab
           path: slab
@@ -187,7 +165,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "WASM benchmarks (${{ matrix.browser }}) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -200,7 +178,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -211,7 +189,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (wasm-client-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_zk_pke.yml

@@ -3,16 +3,6 @@ name: PKE ZK benchmarks
 
 on:
   workflow_dispatch:
-    inputs:
-      bench_type:
-        description: "Benchmarks type"
-        type: choice
-        default: latency
-        options:
-          - latency
-          - throughput
-          - both
-
   push:
     branches:
       - main
@@ -40,21 +30,19 @@ jobs:
       zk_pok_changed: ${{ steps.changed-files.outputs.zk_pok_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           since_last_remote_commit: true
           files_yaml: |
             zk_pok:
               - tfhe/Cargo.toml
-              - tfhe-csprng/**
-              - tfhe-fft/**
+              - concrete-csprng/**
               - tfhe-zk-pok/**
               - tfhe/src/core_crypto/**
               - tfhe/src/shortint/**
@@ -63,37 +51,10 @@ jobs:
               - tfhe/benches/integer/zk_pke.rs
               - .github/workflows/zk_pke_benchmark.yml
 
-  prepare-matrix:
-    name: Prepare operations matrix
-    runs-on: ubuntu-latest
-    if: github.event_name != 'schedule' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
-    outputs:
-      bench_type: ${{ steps.set_bench_type.outputs.bench_type }}
-    steps:
-      - name: Set benchmark types
-        if: github.event_name == 'workflow_dispatch'
-        run: |
-          if [[ "${{ inputs.bench_type }}" == "both" ]]; then
-            echo "BENCH_TYPE=[\"latency\", \"throughput\"]" >> "${GITHUB_ENV}"
-          else
-            echo "BENCH_TYPE=[\"${{ inputs.bench_type }}\"]" >> "${GITHUB_ENV}"
-          fi
-
-      - name: Default benchmark type
-        if: github.event_name != 'workflow_dispatch'
-        run: |
-          echo "BENCH_TYPE=[\"latency\"]" >> "${GITHUB_ENV}"
-
-      - name: Set benchmark types output
-        id: set_bench_type
-        run: |
-          echo "bench_type=${{ toJSON(env.BENCH_TYPE) }}" >> "${GITHUB_OUTPUT}"
-
   setup-instance:
     name: Setup instance (pke-zk-benchmarks)
     runs-on: ubuntu-latest
-    needs: [ should-run, prepare-matrix ]
+    needs: should-run
     if: github.event_name == 'workflow_dispatch' ||
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
       (github.event_name == 'push' &&
@@ -104,7 +65,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -116,18 +77,14 @@ jobs:
   pke-zk-benchmarks:
     name: Execute PKE ZK benchmarks
     if: needs.setup-instance.result != 'skipped'
-    needs: [ prepare-matrix, setup-instance ]
+    needs: setup-instance
     concurrency:
       group: ${{ github.workflow }}_${{github.event_name}}_${{ github.ref }}${{ github.ref == 'refs/heads/main' && github.sha || '' }}
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    strategy:
-      max-parallel: 1
-      matrix:
-        bench_type: ${{ fromJSON(needs.prepare-matrix.outputs.bench_type) }}
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
@@ -140,13 +97,18 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
       - name: Install rust
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           repository: zama-ai/slab
           path: slab
@@ -154,7 +116,7 @@ jobs:
 
       - name: Run benchmarks with AVX512
         run: |
-          make BENCH_TYPE=${{ matrix.bench_type }} bench_integer_zk
+          make bench_integer_zk
 
       - name: Parse results
         run: |
@@ -168,22 +130,22 @@ jobs:
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
           --name-suffix avx512 \
-          --bench-type ${{ matrix.bench_type }}
+          --throughput
 
       - name: Parse CRS sizes results
         run: |
           python3 ./ci/benchmark_parser.py tfhe/pke_zk_crs_sizes.csv ${{ env.RESULTS_FILENAME }} \
-          --object-sizes \
+          --key-sizes \
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b
+        uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874
         with:
           name: ${{ github.sha }}_integer_zk
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           repository: zama-ai/slab
           path: slab
@@ -198,7 +160,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "PKE ZK benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -211,7 +173,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -222,7 +184,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (pke-zk-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/cargo_build.yml

@@ -1,4 +1,4 @@
-name: Cargo Build TFHE-rs
+name: Cargo Build
 
 on:
   pull_request:
@@ -25,10 +25,10 @@ jobs:
       fail-fast: false
 
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+      - uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
@@ -47,10 +47,10 @@ jobs:
         run: |
           make pcc
 
-      - name: Build tfhe-csprng
+      - name: Build concrete-csprng
         if: ${{ contains(matrix.os, 'ubuntu') }}
         run: |
-          make build_tfhe_csprng
+          make build_concrete_csprng
 
       - name: Build Release core
         if: ${{ contains(matrix.os, 'ubuntu') }}

.github/workflows/cargo_build_tfhe_fft.yml

@@ -1,44 +0,0 @@
-# Build tfhe-fft
-name: Cargo Build tfhe-fft
-
-on:
-  pull_request:
-
-env:
-  CARGO_TERM_COLOR: always
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.head_ref }}
-  cancel-in-progress: true
-
-jobs:
-  cargo-builds-fft:
-    runs-on: ${{ matrix.runner_type }}
-
-    strategy:
-      matrix:
-        runner_type: [ubuntu-latest, macos-latest, windows-latest]
-      fail-fast: false
-
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-
-      - name: Install Rust
-        uses: actions-rs/toolchain@16499b5e05bf2e26879000db0c1d13f7e13fa3af
-        with:
-          toolchain: stable
-          override: true
-
-      - name: Run pcc checks
-        if: matrix.runner_type == 'ubuntu-latest'
-        run: |
-          sudo apt install -y libfftw3-dev
-          make pcc_fft
-
-      - name: Build release
-        run: |
-          make build_fft
-
-      - name: Build release no-std
-        run: |
-          make build_fft_no_std

.github/workflows/cargo_build_tfhe_ntt.yml

@@ -1,40 +0,0 @@
-# Build tfhe-ntt
-name: Cargo Build tfhe-ntt
-
-on:
-  pull_request:
-
-env:
-  CARGO_TERM_COLOR: always
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.head_ref }}
-  cancel-in-progress: true
-
-jobs:
-  cargo-builds:
-    runs-on: ${{ matrix.os }}
-    strategy:
-      matrix:
-        os: [ubuntu-latest, macos-latest, windows-latest]
-      fail-fast: false
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-
-      - name: Install Rust
-        uses: actions-rs/toolchain@16499b5e05bf2e26879000db0c1d13f7e13fa3af
-        with:
-          toolchain: stable
-          override: true
-
-      - name: Run pcc checks
-        run: |
-          make pcc_ntt
-
-      - name: Build release
-        run: |
-          make build_ntt
-
-      - name: Build release no-std
-        run: |
-          make build_ntt_no_std

.github/workflows/cargo_test_fft.yml

@@ -1,71 +0,0 @@
-# Test tfhe-fft
-name: Cargo Test tfhe-fft
-
-on:
-  pull_request:
-
-env:
-  CARGO_TERM_COLOR: always
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.head_ref }}
-  cancel-in-progress: true
-
-jobs:
-  cargo-tests:
-    runs-on: ${{ matrix.runner_type }}
-    strategy:
-      matrix:
-        runner_type: [ubuntu-latest, macos-latest, windows-latest]
-      fail-fast: false
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-
-      - name: Install Rust
-        uses: actions-rs/toolchain@16499b5e05bf2e26879000db0c1d13f7e13fa3af
-        with:
-          toolchain: stable
-          override: true
-
-      - name: Test debug
-        run: |
-          make test_fft
-
-      - name: Test serialization
-        run: make test_fft_serde
-
-      - name: Test no-std
-        run: |
-          make test_fft_no_std
-
-  cargo-tests-nightly:
-    runs-on: ${{ matrix.runner_type }}
-    strategy:
-      matrix:
-        runner_type: [ubuntu-latest, macos-latest, windows-latest]
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-
-      - name: Install Rust
-        uses: actions-rs/toolchain@16499b5e05bf2e26879000db0c1d13f7e13fa3af
-        with:
-          toolchain: nightly
-          override: true
-
-      - name: Test nightly
-        run: |
-          make test_fft_nightly
-
-      - name: Test no-std nightly
-        run: |
-          make test_fft_no_std_nightly
-
-  cargo-tests-node-js:
-    runs-on: "ubuntu-latest"
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-
-      - name: Test node js
-        run: |
-          make install_node
-          make test_fft_node_js_ci

.github/workflows/cargo_test_ntt.yml

@@ -1,54 +0,0 @@
-# Test tfhe-ntt
-name: Cargo Test tfhe-ntt
-
-on:
-  pull_request:
-
-env:
-  CARGO_TERM_COLOR: always
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.head_ref }}
-  cancel-in-progress: true
-
-jobs:
-  cargo-tests:
-    runs-on: ${{ matrix.os }}
-    strategy:
-      matrix:
-        os: [ubuntu-latest, macos-latest, windows-latest]
-      fail-fast: false
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-
-      - name: Install Rust
-        uses: actions-rs/toolchain@16499b5e05bf2e26879000db0c1d13f7e13fa3af
-        with:
-          toolchain: stable
-          override: true
-
-      - name: Test debug
-        run: make test_ntt
-
-      - name: Test no-std
-        run: make test_ntt_no_std
-
-  cargo-tests-nightly:
-    runs-on: ${{ matrix.os }}
-    strategy:
-      matrix:
-        os: [ubuntu-latest, macos-latest, windows-latest]
-    steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-
-      - name: Install Rust
-        uses: actions-rs/toolchain@16499b5e05bf2e26879000db0c1d13f7e13fa3af
-        with:
-          toolchain: nightly
-          override: true
-
-      - name: Test nightly
-        run: make test_ntt_nightly
-
-      - name: Test no-std nightly
-        run: make test_ntt_no_std_nightly

.github/workflows/check_commit.yml

@@ -28,6 +28,3 @@ jobs:
           excludeTitle: "true" # optional: this excludes the title of a pull request
           checkAllCommitMessages: "true" # optional: this checks all commits associated with a pull request
           accessToken: ${{ secrets.GITHUB_TOKEN }} # github access token is only required if checkAllCommitMessages is true
-
-      - name: Check commit signatures
-        uses: 1Password/check-signed-commits-action@ed2885f3ed2577a4f5d3c3fe895432a557d23d52

.github/workflows/check_triggering_actor.yml

@@ -1,29 +0,0 @@
-# Check if triggering actor is a collaborator and has write access
-name: Check Triggering Actor
-
-on:
-  workflow_call:
-    secrets:
-      TOKEN:
-        required: true
-
-jobs:
-  check-actor-permission:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Get User Permission
-        id: check-access
-        uses: actions-cool/check-user-permission@956b2e73cdfe3bcb819bb7225e490cb3b18fd76e # v2.2.1
-        with:
-          require: write
-          username: ${{ github.triggering_actor }}
-        env:
-          GITHUB_TOKEN: ${{ secrets.TOKEN }}
-
-      - name: Check User Permission
-        if: steps.check-access.outputs.require-result == 'false'
-        run: |
-          echo "${{ github.triggering_actor }} does not have permissions on this repo."
-          echo "Current permission level is ${{ steps.check-access.outputs.user-permission }}"
-          echo "Job originally triggered by ${{ github.actor }}"
-          exit 1

.github/workflows/ci_lint.yml

@@ -13,7 +13,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
 
       - name: Get actionlint
         run: |
@@ -27,7 +27,7 @@ jobs:
           make lint_workflow
 
       - name: Ensure SHA pinned actions
-        uses: zgosalvez/github-actions-ensure-sha-pinned-actions@6ae615f6475d2ede5ad88bea6baa7a1d5e93ffaa # v3.0.19
+        uses: zgosalvez/github-actions-ensure-sha-pinned-actions@b8f9a25a51fe633d9215ac7734854dc11cd299cb # v3.0.13
         with:
           allowlist: |
             slsa-framework/slsa-github-generator

.github/workflows/code_coverage.yml

@@ -25,7 +25,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -44,22 +44,26 @@ jobs:
     timeout-minutes: 5760 # 4 days
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
+
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           files_yaml: |
             tfhe:
               - tfhe/src/**
-            tfhe_csprng:
-              - tfhe-csprng/src/**
+            concrete_csprng:
+              - concrete-csprng/src/**
 
       - name: Generate Keys
         if: steps.changed-files.outputs.tfhe_any_changed == 'true'
@@ -83,7 +87,7 @@ jobs:
           make test_shortint_cov
 
       - name: Upload tfhe coverage to Codecov
-        uses: codecov/codecov-action@1e68e06f1dbfde0e4cefc87efeba9e4643565303
+        uses: codecov/codecov-action@b9fd7d16f6d7d1b5d2bec1a2887e65ceed900238
         if: steps.changed-files.outputs.tfhe_any_changed == 'true'
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
@@ -97,7 +101,7 @@ jobs:
           make test_integer_cov
 
       - name: Upload tfhe coverage to Codecov
-        uses: codecov/codecov-action@1e68e06f1dbfde0e4cefc87efeba9e4643565303
+        uses: codecov/codecov-action@b9fd7d16f6d7d1b5d2bec1a2887e65ceed900238
         if: steps.changed-files.outputs.tfhe_any_changed == 'true'
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
@@ -108,7 +112,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Code coverage finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -121,7 +125,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -132,7 +136,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (code-coverage) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/csprng_randomness_tests.yml

@@ -27,7 +27,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -45,13 +45,17 @@ jobs:
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: 'false'
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
@@ -62,10 +66,10 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "tfhe-csprng randomness check finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "concrete-csprng randomness check finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (csprng-randomness-tests)
@@ -75,7 +79,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -86,7 +90,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (csprng-randomness-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/data_pr_close.yml

@@ -117,7 +117,7 @@ jobs:
     - name: Slack Notification
       if: ${{ always() && job.status == 'failure' }}
       continue-on-error: true
-      uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+      uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
       env:
         SLACK_COLOR: ${{ job.status }}
         SLACK_MESSAGE: "Failed to auto-${{ env.CLOSE_TYPE }} PR on data repo: ${{ fromJson(env.GH_API_RES || env.TARGET_REPO_PR).message }}"

.github/workflows/gpu_4090_tests.yml

@@ -34,13 +34,13 @@ jobs:
 
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: 'false'
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
@@ -77,7 +77,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "CUDA RTX 4090 tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_fast_h100_tests.yml

@@ -28,14 +28,13 @@ jobs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           since_last_remote_commit: true
           files_yaml: |
@@ -45,16 +44,14 @@ jobs:
               - backends/tfhe-cuda-backend/**
               - tfhe/src/core_crypto/gpu/**
               - tfhe/src/integer/gpu/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
               - 'tfhe/docs/**.md'
+              - Makefile
               - '.github/workflows/gpu_fast_h100_tests.yml'
-              - scripts/integer-tests.sh
-              - ci/slab.toml
+              - scripts/**
+              - ci/**
 
   setup-instance:
     name: Setup instance (cuda-h100-tests)
@@ -68,7 +65,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -94,28 +91,60 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
     steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: 'false'
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
-
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
+          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "HOME=/home/ubuntu";
+          } >> "${GITHUB_ENV}"
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
       - name: Run core crypto and internal CUDA backend tests
         run: |
           BIG_TESTS_INSTANCE=TRUE make test_core_crypto_gpu
@@ -142,7 +171,7 @@ jobs:
     continue-on-error: true
     steps:
       - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
           SLACK_MESSAGE: "Fast H100 tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
@@ -155,7 +184,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -166,7 +195,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_fast_tests.yml

@@ -27,14 +27,13 @@ jobs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           since_last_remote_commit: true
           files_yaml: |
@@ -43,17 +42,15 @@ jobs:
               - tfhe/build.rs
               - backends/tfhe-cuda-backend/**
               - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
               - tfhe/src/integer/gpu/**
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
               - 'tfhe/docs/**.md'
               - '.github/workflows/gpu_fast_tests.yml'
-              - scripts/integer-tests.sh
-              - ci/slab.toml
+              - Makefile
+              - scripts/**
+              - ci/**
 
   setup-instance:
     name: Setup instance (cuda-tests)
@@ -66,7 +63,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -92,28 +89,60 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
     steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: 'false'
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
-
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
+          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "HOME=/home/ubuntu";
+          } >> "${GITHUB_ENV}"
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
       - name: Run core crypto and internal CUDA backend tests
         run: |
           make test_core_crypto_gpu
@@ -140,7 +169,7 @@ jobs:
     continue-on-error: true
     steps:
       - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
           SLACK_MESSAGE: "Base GPU tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
@@ -153,7 +182,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -164,7 +193,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_full_h100_tests.yml

@@ -25,7 +25,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -49,12 +49,15 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
     steps:
       # Mandatory on hyperstack since a bootable volume is not re-usable yet.
       - name: Install dependencies
         run: |
           sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev libclang-dev
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
           wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
           tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
           cd cmake-${{ env.CMAKE_VERSION }}
@@ -63,26 +66,43 @@ jobs:
           sudo make install
 
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: 'false'
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
-
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
+          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "HOME=/home/ubuntu";
+          } >> "${GITHUB_ENV}"
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
       - name: Run core crypto, integer and internal CUDA backend tests
         run: |
           make test_gpu
@@ -107,7 +127,7 @@ jobs:
     continue-on-error: true
     steps:
       - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
           SLACK_MESSAGE: "Full H100 tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
@@ -119,7 +139,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -130,7 +150,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_full_multi_gpu_tests.yml

@@ -28,14 +28,13 @@ jobs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           since_last_remote_commit: true
           files_yaml: |
@@ -44,17 +43,15 @@ jobs:
               - tfhe/build.rs
               - backends/tfhe-cuda-backend/**
               - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
               - tfhe/src/integer/gpu/**
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
               - 'tfhe/docs/**.md'
+              - Makefile
               - '.github/workflows/**_multi_gpu_tests.yml'
-              - scripts/integer-tests.sh
-              - ci/slab.toml
+              - scripts/**
+              - ci/**
 
   setup-instance:
     name: Setup instance (cuda-tests-multi-gpu)
@@ -68,7 +65,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -94,28 +91,60 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
     steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: 'false'
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
-
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
+          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "HOME=/home/ubuntu";
+          } >> "${GITHUB_ENV}"
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
       - name: Run multi-bit CUDA integer compression tests
         run: |
           BIG_TESTS_INSTANCE=TRUE make test_integer_compression_gpu
@@ -145,7 +174,7 @@ jobs:
     continue-on-error: true
     steps:
       - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
           SLACK_MESSAGE: "Multi-GPU tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
@@ -158,7 +187,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -169,7 +198,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-tests-multi-gpu) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_integer_long_run_tests.yml

@@ -1,115 +0,0 @@
-name: Long Run Tests on GPU
-
-env:
-  CARGO_TERM_COLOR: always
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUSTFLAGS: "-C target-cpu=native"
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-
-on:
-  # Allows you to run this workflow manually from the Actions tab as an alternative.
-  workflow_dispatch:
-  schedule:
-    # Weekly tests will be triggered each Friday at 9p.m.
-    - cron: "0 21 * * 5"
-
-jobs:
-  setup-instance:
-    name: Setup instance (gpu-tests)
-    if: github.event_name != 'schedule' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
-    runs-on: ubuntu-latest
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: hyperstack
-          profile: multi-gpu-test
-
-  cuda-tests:
-    name: Long run GPU tests
-    needs: [ setup-instance ]
-    concurrency:
-      group: ${{ github.workflow }}_${{github.event_name}}_${{ github.ref }}
-      cancel-in-progress: true
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    strategy:
-      fail-fast: false
-      # explicit include-based build matrix, of known valid options
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11 
-    timeout-minutes: 4320 # 72 hours
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
-        with:
-          toolchain: stable
-
-      - name: Run tests
-        run: |
-          make test_integer_long_run_gpu
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-tests ]
-    runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-tests.result != 'skipped' && failure() }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ needs.cuda-tests.result }}
-          SLACK_MESSAGE: "Integer GPU long run tests finished with status: ${{ needs.cuda-tests.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (gpu-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-tests ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (gpu-long-run-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_pcc.yml

@@ -24,7 +24,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -53,7 +53,7 @@ jobs:
 
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: 'false'
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
@@ -63,7 +63,7 @@ jobs:
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
@@ -97,7 +97,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "CUDA AWS post-commit checks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -110,7 +110,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -121,7 +121,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-pcc) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_signed_integer_classic_tests.yml

@@ -1,155 +0,0 @@
-# Signed integer GPU tests on an RTXA6000 VM on hyperstack with classical PBS
-name: TFHE Cuda Backend - Signed integer tests with classical PBS
-
-env:
-  CARGO_TERM_COLOR: always
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUSTFLAGS: "-C target-cpu=native"
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' }}
-
-on:
-  # Allows you to run this workflow manually from the Actions tab as an alternative.
-  workflow_dispatch:
-  pull_request:
-      types: [ labeled ]
-
-jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    permissions:
-      pull-requests: write
-    outputs:
-      gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            gpu:
-              - tfhe/Cargo.toml
-              - tfhe/build.rs
-              - backends/tfhe-cuda-backend/**
-              - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
-              - tfhe/src/integer/gpu/**
-              - tfhe/src/shortint/parameters/**
-              - tfhe/src/high_level_api/**
-              - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
-              - '.github/workflows/gpu_signed_integer_classic_tests.yml'
-              - scripts/integer-tests.sh
-              - ci/slab.toml
-
-  setup-instance:
-    name: Setup instance (cuda-signed-classic-tests)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
-      (github.event.action != 'labeled' && needs.should-run.outputs.gpu_test == 'true') ||
-      (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.gpu_test == 'true')
-    runs-on: ubuntu-latest
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: hyperstack
-          profile: gpu-test
-
-  cuda-tests-linux:
-    name: CUDA signed integer tests with classical PBS
-    needs: [ should-run, setup-instance ]
-    if: github.event_name != 'pull_request' ||
-      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    strategy:
-      fail-fast: false
-      # explicit include-based build matrix, of known valid options
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11 
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
-        with:
-          toolchain: stable
-
-      - name: Run signed integer tests
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_signed_integer_gpu_ci
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-tests-linux.result != 'skipped' && failure() }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
-          SLACK_MESSAGE: "Integer GPU signed integer tests with classical PBS finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-signed-classic-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-signed-classic-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_signed_integer_h100_tests.yml

@@ -28,14 +28,14 @@ jobs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           since_last_remote_commit: true
           files_yaml: |
@@ -44,17 +44,15 @@ jobs:
               - tfhe/build.rs
               - backends/tfhe-cuda-backend/**
               - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
               - tfhe/src/integer/gpu/**
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
               - 'tfhe/docs/**.md'
+              - Makefile
               - '.github/workflows/gpu_signed_integer_h100_tests.yml'
-              - scripts/integer-tests.sh
-              - ci/slab.toml
+              - scripts/**
+              - ci/**
 
   setup-instance:
     name: Setup instance (cuda-h100-tests)
@@ -68,7 +66,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -94,25 +92,61 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
     steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
 
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
+          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "HOME=/home/ubuntu";
+          } >> "${GITHUB_ENV}"
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
+      - name: Run signed integer tests
+        run: |
+          BIG_TESTS_INSTANCE=TRUE make test_signed_integer_gpu_ci
+
       - name: Run signed integer multi-bit tests
         run: |
           BIG_TESTS_INSTANCE=TRUE make test_signed_integer_multi_bit_gpu_ci
@@ -125,7 +159,7 @@ jobs:
     continue-on-error: true
     steps:
       - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
           SLACK_MESSAGE: "Integer GPU H100 tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
@@ -138,7 +172,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -149,7 +183,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_signed_integer_tests.yml

@@ -22,6 +22,7 @@ on:
     types:
       - opened
       - synchronize
+      - labeled
   schedule:
     # Nightly tests @ 1AM after each work day
     - cron: "0 1 * * MON-FRI"
@@ -35,14 +36,13 @@ jobs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           since_last_remote_commit: true
           files_yaml: |
@@ -51,17 +51,15 @@ jobs:
               - tfhe/build.rs
               - backends/tfhe-cuda-backend/**
               - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
               - tfhe/src/integer/gpu/**
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
               - 'tfhe/docs/**.md'
               - '.github/workflows/gpu_signed_integer_tests.yml'
-              - scripts/integer-tests.sh
-              - ci/slab.toml
+              - Makefile
+              - scripts/**
+              - ci/**
 
   setup-instance:
     name: Setup instance (cuda-signed-integer-tests)
@@ -69,13 +67,13 @@ jobs:
     needs: should-run
     if: (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
       github.event_name == 'workflow_dispatch' ||
-      needs.should-run.outputs.gpu_test == 'true'
+      (github.event.action != 'labeled' && needs.should-run.outputs.gpu_test == 'true')
     outputs:
       runner-name: ${{ steps.start-instance.outputs.label }}
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -101,28 +99,56 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
     steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           persist-credentials: 'false'
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
-
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
+          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "HOME=/home/ubuntu";
+          } >> "${GITHUB_ENV}"
+
       - name: Should run nightly tests
         if: github.event_name == 'schedule'
         run: |
@@ -131,6 +157,10 @@ jobs:
             echo "NIGHTLY_TESTS=TRUE";
           } >> "${GITHUB_ENV}"
 
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
       - name: Run signed integer multi-bit tests
         run: |
           make test_signed_integer_multi_bit_gpu_ci
@@ -143,7 +173,7 @@ jobs:
     continue-on-error: true
     steps:
       - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ needs.cuda-signed-integer-tests.result }}
           SLACK_MESSAGE: "Base GPU tests finished with status: ${{ needs.cuda-signed-integer-tests.result }}. (${{ env.ACTION_RUN_URL }})"
@@ -156,7 +186,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -167,7 +197,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-signed-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_unsigned_integer_classic_tests.yml

@@ -1,155 +0,0 @@
-# Test unsigned integers on an RTXA6000 VM on hyperstack with the classical PBS
-name: TFHE Cuda Backend - Unsigned integer tests with classical PBS
-
-env:
-  CARGO_TERM_COLOR: always
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUSTFLAGS: "-C target-cpu=native"
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' }}
-
-on:
-  # Allows you to run this workflow manually from the Actions tab as an alternative.
-  workflow_dispatch:
-  pull_request:
-      types: [ labeled ]
-
-jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    permissions:
-      pull-requests: write
-    outputs:
-      gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            gpu:
-              - tfhe/Cargo.toml
-              - tfhe/build.rs
-              - backends/tfhe-cuda-backend/**
-              - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
-              - tfhe/src/integer/gpu/**
-              - tfhe/src/shortint/parameters/**
-              - tfhe/src/high_level_api/**
-              - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
-              - '.github/workflows/gpu_unsigned_integer_classic_tests.yml'
-              - scripts/integer-tests.sh
-              - ci/slab.toml
-
-  setup-instance:
-    name: Setup instance (cuda-unsigned-classic-tests)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
-      (github.event.action != 'labeled' && needs.should-run.outputs.gpu_test == 'true') ||
-      (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.gpu_test == 'true')
-    runs-on: ubuntu-latest
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: hyperstack
-          profile: gpu-test
-
-  cuda-tests-linux:
-    name: CUDA unsigned integer tests with classical PBS
-    needs: [ should-run, setup-instance ]
-    if: github.event_name != 'pull_request' ||
-      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    strategy:
-      fail-fast: false
-      # explicit include-based build matrix, of known valid options
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11 
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
-        with:
-          toolchain: stable
-
-      - name: Run unsigned integer tests
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_unsigned_integer_gpu_ci
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-tests-linux.result != 'skipped' && failure() }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
-          SLACK_MESSAGE: "Unsigned integer GPU classic tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-unsigned-classic-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-unsigned-classic-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_unsigned_integer_h100_tests.yml

@@ -28,14 +28,14 @@ jobs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           since_last_remote_commit: true
           files_yaml: |
@@ -44,17 +44,15 @@ jobs:
               - tfhe/build.rs
               - backends/tfhe-cuda-backend/**
               - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
               - tfhe/src/integer/gpu/**
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
               - 'tfhe/docs/**.md'
-              - '.github/workflows/gpu_unsigned_integer_h100_tests.yml'
-              - scripts/integer-tests.sh
-              - ci/slab.toml
+              - Makefile
+              - '.github/workflows/gpu_unsigned_integer_tests.yml'
+              - scripts/**
+              - ci/**
 
   setup-instance:
     name: Setup instance (cuda-h100-tests)
@@ -68,7 +66,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -94,25 +92,61 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
     steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
 
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
+          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "HOME=/home/ubuntu";
+          } >> "${GITHUB_ENV}"
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
+      - name: Run unsigned integer tests
+        run: |
+          BIG_TESTS_INSTANCE=TRUE make test_unsigned_integer_gpu_ci
+
       - name: Run unsigned integer multi-bit tests
         run: |
           BIG_TESTS_INSTANCE=TRUE make test_unsigned_integer_multi_bit_gpu_ci
@@ -125,7 +159,7 @@ jobs:
     continue-on-error: true
     steps:
       - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
           SLACK_MESSAGE: "Unsigned integer GPU H100 tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
@@ -138,7 +172,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -149,7 +183,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_unsigned_integer_tests.yml

@@ -21,6 +21,7 @@ on:
     types:
       - opened
       - synchronize
+      - labeled
   schedule:
     # Nightly tests @ 1AM after each work day
     - cron: "0 1 * * MON-FRI"
@@ -34,14 +35,14 @@ jobs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
           token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@d6e91a2266cdb9d62096cebf1e8546899c6aa18f
+        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
         with:
           since_last_remote_commit: true
           files_yaml: |
@@ -50,31 +51,29 @@ jobs:
               - tfhe/build.rs
               - backends/tfhe-cuda-backend/**
               - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
-              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
               - tfhe/src/integer/gpu/**
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
               - 'tfhe/docs/**.md'
               - '.github/workflows/gpu_unsigned_integer_tests.yml'
-              - scripts/integer-tests.sh
-              - ci/slab.toml
+              - Makefile
+              - scripts/**
+              - ci/**
 
   setup-instance:
     name: Setup instance (cuda-unsigned-integer-tests)
     needs: should-run
     if: (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
       github.event_name == 'workflow_dispatch' ||
-      needs.should-run.outputs.gpu_test == 'true'
+      (github.event.action != 'labeled' && needs.should-run.outputs.gpu_test == 'true')
     runs-on: ubuntu-latest
     outputs:
       runner-name: ${{ steps.start-instance.outputs.label }}
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -100,25 +99,53 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
     steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
 
-      - name: Setup Hyperstack dependencies
-        uses: ./.github/actions/hyperstack_setup
-        with:
-          cuda-version: ${{ matrix.cuda }}
-          gcc-version: ${{ matrix.gcc }}
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
+          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "HOME=/home/ubuntu";
+          } >> "${GITHUB_ENV}"
+
       - name: Should run nightly tests
         if: github.event_name == 'schedule'
         run: |
@@ -127,6 +154,10 @@ jobs:
             echo "NIGHTLY_TESTS=TRUE";
           } >> "${GITHUB_ENV}"
 
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
       - name: Run unsigned integer multi-bit tests
         run: |
           make test_unsigned_integer_multi_bit_gpu_ci
@@ -139,7 +170,7 @@ jobs:
     continue-on-error: true
     steps:
       - name: Send message
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ needs.cuda-unsigned-integer-tests.result }}
           SLACK_MESSAGE: "Unsigned integer GPU tests finished with status: ${{ needs.cuda-unsigned-integer-tests.result }}. (${{ env.ACTION_RUN_URL }})"
@@ -152,7 +183,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -163,7 +194,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-unsigned-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/integer_long_run_tests.yml

@@ -1,95 +0,0 @@
-name: AWS Long Run Tests on CPU
-
-env:
-  CARGO_TERM_COLOR: always
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUSTFLAGS: "-C target-cpu=native"
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-
-on:
-  # Allows you to run this workflow manually from the Actions tab as an alternative.
-  workflow_dispatch:
-  schedule:
-    # Weekly tests will be triggered each Friday at 9p.m.
-    - cron: "0 21 * * 5"
-
-jobs:
-  setup-instance:
-    name: Setup instance (cpu-tests)
-    if: github.event_name != 'schedule' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
-    runs-on: ubuntu-latest
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: cpu-big
-
-  cpu-tests:
-    name: Long run CPU tests
-    needs: [ setup-instance ]
-    concurrency:
-      group: ${{ github.workflow }}_${{github.event_name}}_${{ github.ref }}
-      cancel-in-progress: true
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    timeout-minutes: 4320 # 72 hours
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          persist-credentials: 'false'
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
-        with:
-          toolchain: stable
-
-      - name: Run tests
-        run: |
-          make test_integer_long_run
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "CPU long run tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cpu-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cpu-tests ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cpu-long-run-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/m1_tests.yml

@@ -3,7 +3,7 @@ name: Tests on M1 CPU
 on:
   workflow_dispatch:
   pull_request:
-    types: [labeled]
+    types: [ labeled ]
   # Have a nightly build for M1 tests
   schedule:
     # * is a special character in YAML so you have to quote this string
@@ -27,66 +27,29 @@ concurrency:
   cancel-in-progress: true
 
 jobs:
-  cargo-builds-m1:
+  cargo-builds:
     if: ${{ (github.event_name == 'schedule' &&  github.repository == 'zama-ai/tfhe-rs') || github.event_name == 'workflow_dispatch' || contains(github.event.label.name, 'm1_test') }}
     runs-on: ["self-hosted", "m1mac"]
     # 12 hours, default is 6 hours, hopefully this is more than enough
     timeout-minutes: 720
 
     steps:
-      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+      - uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
-          persist-credentials: "false"
+          persist-credentials: 'false'
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
-      - name: Run pcc FFT checks
-        run: |
-          make pcc_fft
-
-      - name: Build FFT release
-        run: |
-          make build_fft
-
-      - name: Build FFT release no-std
-        run: |
-          make build_fft_no_std
-
-      - name: Run FFT tests
-        run: |
-          make test_fft
-          make test_fft_serde
-          make test_fft_nightly
-          make test_fft_no_std
-          make test_fft_no_std_nightly
-          # we don't run the js stuff here as it's causing issues with the M1 config
-
-      - name: Run pcc NTT checks
-        run: |
-          make pcc_ntt
-
-      - name: Build NTT release
-        run: |
-          make build_ntt
-
-      - name: Build NTT release no-std
-        run: |
-          make build_ntt_no_std
-
-      - name: Run NTT tests
-        run: |
-          make test_ntt_all
-
       - name: Run pcc checks
         run: |
           make pcc
 
-      - name: Build tfhe-csprng
+      - name: Build concrete-csprng
         run: |
-          make build_tfhe_csprng
+          make build_concrete_csprng
 
       - name: Build Release core
         run: |
@@ -112,9 +75,9 @@ jobs:
         run: |
           make build_c_api
 
-      - name: Run tfhe-csprng tests
+      - name: Run concrete-csprng tests
         run: |
-          make test_tfhe_csprng
+          make test_concrete_csprng
 
       - name: Run tfhe-zk-pok tests
         run: |
@@ -174,7 +137,7 @@ jobs:
     name: Remove m1_test label
     runs-on: ubuntu-latest
     needs:
-      - cargo-builds-m1
+      - cargo-builds
     if: ${{ always() }}
     steps:
       - uses: actions-ecosystem/action-remove-labels@2ce5d41b4b6aa8503e285553f75ed56e0a40bae0
@@ -184,13 +147,13 @@ jobs:
           github_token: ${{ secrets.GITHUB_TOKEN }}
 
       - name: Slack Notification
-        if: ${{ needs.cargo-builds-m1.result != 'skipped' }}
+        if: ${{ needs.cargo-builds.result != 'skipped' }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
-          SLACK_COLOR: ${{ needs.cargo-builds-m1.result }}
+          SLACK_COLOR: ${{ needs.cargo-builds.result }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
           SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-          SLACK_MESSAGE: "M1 tests finished with status: ${{ needs.cargo-builds-m1.result }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "M1 tests finished with status: ${{ needs.cargo-builds.result }}. (${{ env.ACTION_RUN_URL }})"
           SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
           SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/make_release.yml

@@ -30,27 +30,19 @@ env:
   NPM_TAG: ""
 
 jobs:
-  verify_tag:
-    uses: ./.github/workflows/verify_tagged_commit.yml
-    secrets:
-      RELEASE_TEAM: ${{ secrets.RELEASE_TEAM }}
-      READ_ORG_TOKEN: ${{ secrets.READ_ORG_TOKEN }}
-
   package:
     runs-on: ubuntu-latest
-    needs: verify_tag
     outputs:
       hash: ${{ steps.hash.outputs.hash }}
     steps:
       - name: Checkout
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
       - name: Prepare package
         run: |
           cargo package -p tfhe
-      - uses: actions/upload-artifact@6f51ac03b9356f520e9adb1b1b7802705f340c2b # v4.5.0
+      - uses: actions/upload-artifact@50769540e7f4bd5e21e526ee35c689e35e0d6874 # v4.4.0
         with:
           name: crate
           path: target/package/*.crate
@@ -82,10 +74,9 @@ jobs:
       id-token: write
     steps:
       - name: Checkout
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
       - name: Create NPM version tag
         if: ${{ inputs.npm_latest_tag }}
         run: |
@@ -110,7 +101,7 @@ jobs:
       - name: Slack notification (hashes comparison)
         if: ${{ needs.package.outputs.hash != steps.published_hash.outputs.pub_hash }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: failure
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
@@ -155,7 +146,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/make_release_concrete_csprng.yml

@@ -1,4 +1,4 @@
-name: Publish tfhe-csprng release
+name: Publish concrete-csprng release
 
 on:
   workflow_dispatch:
@@ -12,38 +12,30 @@ env:
   ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
 
 jobs:
-  verify_tag:
-    uses: ./.github/workflows/verify_tagged_commit.yml
-    secrets:
-      RELEASE_TEAM: ${{ secrets.RELEASE_TEAM }}
-      READ_ORG_TOKEN: ${{ secrets.READ_ORG_TOKEN }}
-
   publish_release:
-    name: Publish tfhe-csprng Release
-    needs: verify_tag
+    name: Publish concrete-csprng Release
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Publish crate.io package
         env:
           CRATES_TOKEN: ${{ secrets.CARGO_REGISTRY_TOKEN }}
           DRY_RUN: ${{ inputs.dry_run && '--dry-run' || '' }}
         run: |
-          cargo publish -p tfhe-csprng --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
+          cargo publish -p concrete-csprng --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
           SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-          SLACK_MESSAGE: "tfhe-csprng release finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "concrete-csprng release finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
           SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
           SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/make_release_cuda.yml

@@ -21,22 +21,15 @@ env:
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
 
 jobs:
-  verify_tag:
-    uses: ./.github/workflows/verify_tagged_commit.yml
-    secrets:
-      RELEASE_TEAM: ${{ secrets.RELEASE_TEAM }}
-      READ_ORG_TOKEN: ${{ secrets.READ_ORG_TOKEN }}
-
   setup-instance:
     name: Setup instance (publish-cuda-release)
-    needs: verify_tag
     runs-on: ubuntu-latest
     outputs:
       runner-name: ${{ steps.start-instance.outputs.label }}
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -61,17 +54,16 @@ jobs:
       CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
     steps:
       - name: Checkout
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
+        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
         with:
           toolchain: stable
 
@@ -107,7 +99,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "tfhe-cuda-backend release finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -120,7 +112,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
+        uses: zama-ai/slab-github-runner@c0e7168795bd78f61f61146951ed9d0c73c9b701
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -131,7 +123,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (publish-cuda-release) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/make_release_tfhe_fft.yml

@@ -1,49 +0,0 @@
-# Publish new release of tfhe-fft
-name: Publish tfhe-fft release
-
-on:
-  workflow_dispatch:
-    inputs:
-      dry_run:
-        description: "Dry-run"
-        type: boolean
-        default: true
-
-env:
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-
-jobs:
-  verify_tag:
-    uses: ./.github/workflows/verify_tagged_commit.yml
-    secrets:
-      RELEASE_TEAM: ${{ secrets.RELEASE_TEAM }}
-      READ_ORG_TOKEN: ${{ secrets.READ_ORG_TOKEN }}
-
-  publish_release:
-    name: Publish tfhe-fft Release
-    runs-on: ubuntu-latest
-    needs: verify_tag
-    steps:
-      - name: Checkout
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          fetch-depth: 0
-
-      - name: Publish crate.io package
-        env:
-          CRATES_TOKEN: ${{ secrets.CARGO_REGISTRY_TOKEN }}
-          DRY_RUN: ${{ inputs.dry_run && '--dry-run' || '' }}
-        run: |
-          cargo publish -p tfhe-fft --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-          SLACK_MESSAGE: "tfhe-fft release failed: (${{ env.ACTION_RUN_URL }})"
-          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/make_release_tfhe_ntt.yml

@@ -1,49 +0,0 @@
-# Publish new release of tfhe-ntt
-name: Publish tfhe-ntt release
-
-on:
-  workflow_dispatch:
-    inputs:
-      dry_run:
-        description: "Dry-run"
-        type: boolean
-        default: true
-
-env:
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-
-jobs:
-  verify_tag:
-    uses: ./.github/workflows/verify_tagged_commit.yml
-    secrets:
-      RELEASE_TEAM: ${{ secrets.RELEASE_TEAM }}
-      READ_ORG_TOKEN: ${{ secrets.READ_ORG_TOKEN }}
-
-  publish_release:
-    name: Publish tfhe-ntt Release
-    runs-on: ubuntu-latest
-    needs: verify_tag
-    steps:
-      - name: Checkout
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          fetch-depth: 0
-
-      - name: Publish crate.io package
-        env:
-          CRATES_TOKEN: ${{ secrets.CARGO_REGISTRY_TOKEN }}
-          DRY_RUN: ${{ inputs.dry_run && '--dry-run' || '' }}
-        run: |
-          cargo publish -p tfhe-ntt --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-          SLACK_MESSAGE: "tfhe-ntt release failed: (${{ env.ACTION_RUN_URL }})"
-          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/make_release_tfhe_versionable.yml

@@ -12,22 +12,14 @@ env:
   ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
 
 jobs:
-  verify_tag:
-    uses: ./.github/workflows/verify_tagged_commit.yml
-    secrets:
-      RELEASE_TEAM: ${{ secrets.RELEASE_TEAM }}
-      READ_ORG_TOKEN: ${{ secrets.READ_ORG_TOKEN }}
-
   publish_release:
     name: Publish tfhe-versionable Release
-    needs: verify_tag
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Publish proc-macro crate
         env:
@@ -46,7 +38,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/make_release_zk_pok.yml

@@ -13,22 +13,14 @@ env:
   ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
 
 jobs:
-  verify_tag:
-    uses: ./.github/workflows/verify_tagged_commit.yml
-    secrets:
-      RELEASE_TEAM: ${{ secrets.RELEASE_TEAM }}
-      READ_ORG_TOKEN: ${{ secrets.READ_ORG_TOKEN }}
-
   publish_release:
     name: Publish tfhe-zk-pok Release
-    needs: verify_tag
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Publish crate.io package
         env:
@@ -40,7 +32,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/parameters_check.yml

@@ -17,10 +17,10 @@ jobs:
     runs-on: large_ubuntu_16
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
 
       - name: Checkout lattice-estimator
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           repository: malb/lattice-estimator
           path: lattice_estimator
@@ -42,7 +42,7 @@ jobs:
       - name: Slack Notification
         if: ${{ always() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/sync_on_push.yml

@@ -13,10 +13,9 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout repo
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@d632683dd7b4114ad314bca15554477dd762a938
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
       - name: git-sync
         uses: wei/git-sync@55c6b63b4f21607da0e9877ca9b4d11a29fc6d83
         with:

.github/workflows/verify_tagged_commit.yml

@@ -1,32 +0,0 @@
-# Verify a tagged commit
-name: Verify tagged commit
-
-on:
-  workflow_call:
-    secrets:
-      RELEASE_TEAM:
-        required: true
-      READ_ORG_TOKEN:
-        required: true
-
-jobs:
-  checks:
-    runs-on: ubuntu-latest
-    if: startsWith(github.ref, 'refs/tags/')
-    steps:
-      # Check triggering actor membership
-      - name: Actor verification
-        id: actor_check
-        uses: morfien101/actions-authorized-user@4a3cfbf0bcb3cafe4a71710a278920c5d94bb38b
-        with:
-          username: ${{ github.triggering_actor }}
-          org: ${{ github.repository_owner }}
-          team: ${{ secrets.RELEASE_TEAM }}
-          github_token: ${{ secrets.READ_ORG_TOKEN }}
-
-      - name: Actor authorized
-        run: |
-          if [ "${{ steps.actor_check.outputs.authorized }}" == "false" ]; then
-            echo "Actor '${{ github.triggering_actor }}' is not authorized to perform release"
-            exit 1
-          fi

.gitignore

@@ -13,7 +13,6 @@ target/
 
 # Some of our bench outputs
 /tfhe/benchmarks_parameters
-/tfhe-zk-pok/benchmarks_parameters
 **/*.csv
 
 # dieharder run log
@@ -29,8 +28,6 @@ backends/tfhe-cuda-backend/cuda/cmake-build-debug/
 tfhe/web_wasm_parallel_tests/server.PID
 venv/
 web-test-runner/
-node_modules/
-package-lock.json
 
 # Dir used for backward compatibility test data
 tfhe/tfhe-backward-compat-data/

Cargo.toml

@@ -2,12 +2,10 @@
 resolver = "2"
 members = [
     "tfhe",
-    "tfhe-fft",
-    "tfhe-ntt",
     "tfhe-zk-pok",
     "tasks",
     "apps/trivium",
-    "tfhe-csprng",
+    "concrete-csprng",
     "backends/tfhe-cuda-backend",
     "utils/tfhe-versionable",
     "utils/tfhe-versionable-derive",
@@ -18,17 +16,6 @@ exclude = [
     "utils/cargo-tfhe-lints-inner",
     "utils/cargo-tfhe-lints"
 ]
-[workspace.dependencies]
-aligned-vec = { version = "0.6", default-features = false }
-bytemuck = "1.14.3"
-dyn-stack = { version = "0.11", default-features = false }
-itertools = "0.13"
-num-complex = "0.4"
-pulp = { version = "0.20.0", default-features = false }
-rand = "0.8"
-rayon = "1"
-serde = { version = "1.0", default-features = false }
-wasm-bindgen = ">=0.2.86,<0.2.94"
 
 [profile.bench]
 lto = "fat"

README.md

@@ -70,12 +70,26 @@ production-ready library for all the advanced features of TFHE.
 ### Cargo.toml configuration
 To use the latest version of `TFHE-rs` in your project, you first need to add it as a dependency in your `Cargo.toml`:
 
++ For x86_64-based machines running Unix-like OSes:
+
 ```toml
-tfhe = { version = "*", features = ["boolean", "shortint", "integer"] }
+tfhe = { version = "*", features = ["boolean", "shortint", "integer", "x86_64-unix"] }
+```
+
++ For Apple Silicon or aarch64-based machines running Unix-like OSes:
+
+```toml
+tfhe = { version = "*", features = ["boolean", "shortint", "integer", "aarch64-unix"] }
+```
+
++ For x86_64-based machines with the [`rdseed instruction`](https://en.wikipedia.org/wiki/RDRAND) running Windows:
+
+```toml
+tfhe = { version = "*", features = ["boolean", "shortint", "integer", "x86_64"] }
 ```
 
 > [!Note]
-> Note: You need to use a Rust version >= 1.81 to compile TFHE-rs.
+> Note: You need to use a Rust version >= 1.73 to compile TFHE-rs.
 
 > [!Note]
 > Note: aarch64-based machines are not yet supported for Windows as it's currently missing an entropy source to be able to seed the [CSPRNGs](https://en.wikipedia.org/wiki/Cryptographically_secure_pseudorandom_number_generator) used in TFHE-rs.

apps/trivium/Cargo.toml

@@ -6,8 +6,15 @@ edition = "2021"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
-rayon = { workspace = true }
-tfhe = { path = "../../tfhe", features = [ "boolean", "shortint", "integer" ] }
+rayon = { version = "1.7.0"}
+
+[target.'cfg(target_arch = "x86_64")'.dependencies.tfhe]
+path = "../../tfhe"
+features = [ "boolean", "shortint", "integer", "x86_64" ]
+
+[target.'cfg(target_arch = "aarch64")'.dependencies.tfhe]
+path = "../../tfhe"
+features = [ "boolean", "shortint", "integer", "aarch64-unix" ]
 
 [dev-dependencies]
 criterion = { version = "0.5.1", features = [ "html_reports" ]}

apps/trivium/README.md

@@ -7,7 +7,7 @@ using multithreading to accelerate the computation.
 
 
 Quite similarly, the function `TriviumStream::<FheBool>::new` will return a very similar object running in FHE space. Its arguments are
-2 arrays of 80 FheBool representing the encrypted Trivium key, and the encrypted IV. It also requires a reference to the server key of the 
+2 arrays of 80 FheBool representing the encrypted Trivium key, and the encrypted IV. It also requires a reference to the the server key of the 
 current scheme. This means that any user of this feature must also have the `tfhe-rs` crate as a dependency.
 
 

apps/trivium/benches/kreyvium_shortint.rs

@@ -1,6 +1,5 @@
 use criterion::Criterion;
 use tfhe::prelude::*;
-use tfhe::shortint::parameters::V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64;
 use tfhe::shortint::prelude::*;
 use tfhe::{generate_keys, ConfigBuilder, FheUint64};
 use tfhe_trivium::{KreyviumStreamShortint, TransCiphering};
@@ -11,13 +10,12 @@ pub fn kreyvium_shortint_warmup(c: &mut Criterion) {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) =
-        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB000000000000".to_string();
@@ -62,13 +60,12 @@ pub fn kreyvium_shortint_gen(c: &mut Criterion) {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) =
-        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB000000000000".to_string();
@@ -108,13 +105,12 @@ pub fn kreyvium_shortint_trans(c: &mut Criterion) {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) =
-        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB000000000000".to_string();

apps/trivium/benches/trivium_shortint.rs

@@ -1,6 +1,5 @@
 use criterion::Criterion;
 use tfhe::prelude::*;
-use tfhe::shortint::parameters::V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64;
 use tfhe::shortint::prelude::*;
 use tfhe::{generate_keys, ConfigBuilder, FheUint64};
 use tfhe_trivium::{TransCiphering, TriviumStreamShortint};
@@ -11,13 +10,12 @@ pub fn trivium_shortint_warmup(c: &mut Criterion) {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) =
-        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB".to_string();
@@ -62,13 +60,12 @@ pub fn trivium_shortint_gen(c: &mut Criterion) {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) =
-        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB".to_string();
@@ -108,13 +105,12 @@ pub fn trivium_shortint_trans(c: &mut Criterion) {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) =
-        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB".to_string();

apps/trivium/src/kreyvium/test.rs

@@ -1,7 +1,7 @@
 use crate::{KreyviumStream, KreyviumStreamByte, KreyviumStreamShortint, TransCiphering};
 use tfhe::prelude::*;
-use tfhe::shortint::parameters::V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64;
 use tfhe::{generate_keys, ConfigBuilder, FheBool, FheUint64, FheUint8};
+
 // Values for these tests come from the github repo renaud1239/Kreyvium,
 // commit fd6828f68711276c25f55e605935028f5e843f43
 
@@ -221,13 +221,12 @@ fn kreyvium_test_shortint_long() {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) =
-        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB000000000000".to_string();

apps/trivium/src/trivium/test.rs

@@ -1,7 +1,7 @@
 use crate::{TransCiphering, TriviumStream, TriviumStreamByte, TriviumStreamShortint};
 use tfhe::prelude::*;
-use tfhe::shortint::parameters::V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64;
 use tfhe::{generate_keys, ConfigBuilder, FheBool, FheUint64, FheUint8};
+
 // Values for these tests come from the github repo cantora/avr-crypto-lib, commit 2a5b018,
 // file testvectors/trivium-80.80.test-vectors
 
@@ -357,13 +357,12 @@ fn trivium_test_shortint_long() {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) =
-        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB".to_string();

backends/tfhe-cuda-backend/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "tfhe-cuda-backend"
-version = "0.7.0"
+version = "0.4.0"
 edition = "2021"
 authors = ["Zama team"]
 license = "BSD-3-Clause-Clear"
@@ -14,4 +14,3 @@ keywords = ["fully", "homomorphic", "encryption", "fhe", "cryptography"]
 [build-dependencies]
 cmake = { version = "0.1" }
 pkg-config = { version = "0.3" }
-bindgen = "0.70.1"

backends/tfhe-cuda-backend/build.rs

@@ -1,8 +1,8 @@
-use std::path::PathBuf;
+use std::env;
 use std::process::Command;
 
 fn main() {
-    if let Ok(val) = std::env::var("DOCS_RS") {
+    if let Ok(val) = env::var("DOCS_RS") {
         if val.parse::<u32>() == Ok(1) {
             return;
         }
@@ -26,8 +26,7 @@ fn main() {
     println!("cargo::rerun-if-changed=cuda/tests_and_benchmarks");
     println!("cargo::rerun-if-changed=cuda/CMakeLists.txt");
     println!("cargo::rerun-if-changed=src");
-
-    if std::env::consts::OS == "linux" {
+    if env::consts::OS == "linux" {
         let output = Command::new("./get_os_name.sh").output().unwrap();
         let distribution = String::from_utf8(output.stdout).unwrap();
         if distribution != "Ubuntu\n" {
@@ -36,7 +35,6 @@ fn main() {
                 Only Ubuntu is supported by tfhe-cuda-backend at this time. Build may fail\n"
             );
         }
-
         let dest = cmake::build("cuda");
         println!("cargo:rustc-link-search=native={}", dest.display());
         println!("cargo:rustc-link-lib=static=tfhe_cuda_backend");
@@ -53,57 +51,6 @@ fn main() {
         println!("cargo:rustc-link-lib=cudart");
         println!("cargo:rustc-link-search=native=/usr/lib/x86_64-linux-gnu/");
         println!("cargo:rustc-link-lib=stdc++");
-
-        let header_path = "wrapper.h";
-        let headers = vec![
-            "wrapper.h",
-            "cuda/include/ciphertext.h",
-            "cuda/include/integer/compression/compression.h",
-            "cuda/include/integer/integer.h",
-            "cuda/include/keyswitch.h",
-            "cuda/include/linear_algebra.h",
-            "cuda/include/pbs/programmable_bootstrap.h",
-            "cuda/include/pbs/programmable_bootstrap_multibit.h",
-        ];
-        let out_path = PathBuf::from("src").join("bindings.rs");
-        let bindings_modified = if out_path.exists() {
-            std::fs::metadata(&out_path).unwrap().modified().unwrap()
-        } else {
-            std::time::SystemTime::UNIX_EPOCH // If bindings file doesn't exist, consider it older
-        };
-        let mut headers_modified = bindings_modified;
-        for header in headers {
-            println!("cargo:rerun-if-changed={}", header);
-            // Check modification times
-            let header_modified = std::fs::metadata(header).unwrap().modified().unwrap();
-            if header_modified > headers_modified {
-                headers_modified = header_modified;
-            }
-        }
-
-        // Regenerate bindings only if header has been modified
-        if headers_modified > bindings_modified {
-            let bindings = bindgen::Builder::default()
-                .header(header_path)
-                // allow only what we are interested in, the custom types appearing in the interface
-                .allowlist_type("PBS_TYPE")
-                .allowlist_type("SHIFT_OR_ROTATE_TYPE")
-                // and the functions reachable from the headers included in wrapper.h
-                .allowlist_function(".*")
-                .clang_arg("-x")
-                .clang_arg("c++")
-                .clang_arg("-std=c++17")
-                .clang_arg("-I/usr/include")
-                .clang_arg("-I/usr/local/include")
-                .ctypes_prefix("ffi")
-                .raw_line("use crate::ffi;")
-                .generate()
-                .expect("Unable to generate bindings");
-
-            bindings
-                .write_to_file(&out_path)
-                .expect("Couldn't write bindings!");
-        }
     } else {
         panic!(
             "Error: platform not supported, tfhe-cuda-backend not built (only Linux is supported)"

backends/tfhe-cuda-backend/cuda/include/ciphertext.h

@@ -1,24 +1,25 @@
 #ifndef CUDA_CIPHERTEXT_H
 #define CUDA_CIPHERTEXT_H
 
-#include "stdint.h"
+#include "device.h"
+#include <cstdint>
 
 extern "C" {
 void cuda_convert_lwe_ciphertext_vector_to_gpu_64(void *stream,
                                                   uint32_t gpu_index,
-                                                  void *dest, void const *src,
+                                                  void *dest, void *src,
                                                   uint32_t number_of_cts,
                                                   uint32_t lwe_dimension);
 void cuda_convert_lwe_ciphertext_vector_to_cpu_64(void *stream,
                                                   uint32_t gpu_index,
-                                                  void *dest, void const *src,
+                                                  void *dest, void *src,
                                                   uint32_t number_of_cts,
                                                   uint32_t lwe_dimension);
 
 void cuda_glwe_sample_extract_64(void *stream, uint32_t gpu_index,
-                                 void *lwe_array_out, void const *glwe_array_in,
-                                 uint32_t const *nth_array, uint32_t num_nths,
+                                 void *lwe_array_out, void *glwe_array_in,
+                                 uint32_t *nth_array, uint32_t num_nths,
                                  uint32_t glwe_dimension,
                                  uint32_t polynomial_size);
-}
+};
 #endif

backends/tfhe-cuda-backend/cuda/include/compression.h

@@ -1,7 +1,46 @@
-#ifndef CUDA_INTEGER_COMPRESSION_UTILITIES_H
-#define CUDA_INTEGER_COMPRESSION_UTILITIES_H
+#ifndef CUDA_INTEGER_COMPRESSION_H
+#define CUDA_INTEGER_COMPRESSION_H
 
-#include "../integer_utilities.h"
+#include "integer.h"
+
+extern "C" {
+void scratch_cuda_integer_compress_radix_ciphertext_64(
+    void **streams, uint32_t *gpu_indexes, uint32_t gpu_count, int8_t **mem_ptr,
+    uint32_t compression_glwe_dimension, uint32_t compression_polynomial_size,
+    uint32_t lwe_dimension, uint32_t ks_level, uint32_t ks_base_log,
+    uint32_t num_radix_blocks, uint32_t message_modulus, uint32_t carry_modulus,
+    PBS_TYPE pbs_type, uint32_t lwe_per_glwe, uint32_t storage_log_modulus,
+    bool allocate_gpu_memory);
+
+void scratch_cuda_integer_decompress_radix_ciphertext_64(
+    void **streams, uint32_t *gpu_indexes, uint32_t gpu_count, int8_t **mem_ptr,
+    uint32_t encryption_glwe_dimension, uint32_t encryption_polynomial_size,
+    uint32_t compression_glwe_dimension, uint32_t compression_polynomial_size,
+    uint32_t lwe_dimension, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t num_radix_blocks, uint32_t message_modulus, uint32_t carry_modulus,
+    PBS_TYPE pbs_type, uint32_t storage_log_modulus, uint32_t body_count,
+    bool allocate_gpu_memory);
+
+void cuda_integer_compress_radix_ciphertext_64(
+    void **streams, uint32_t *gpu_indexes, uint32_t gpu_count,
+    void *glwe_array_out, void *lwe_array_in, void **fp_ksk, uint32_t num_nths,
+    int8_t *mem_ptr);
+
+void cuda_integer_decompress_radix_ciphertext_64(
+    void **streams, uint32_t *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array_out, void *glwe_in, uint32_t *indexes_array,
+    uint32_t indexes_array_size, void **bsks, int8_t *mem_ptr);
+
+void cleanup_cuda_integer_compress_radix_ciphertext_64(void **streams,
+                                                       uint32_t *gpu_indexes,
+                                                       uint32_t gpu_count,
+                                                       int8_t **mem_ptr_void);
+
+void cleanup_cuda_integer_decompress_radix_ciphertext_64(void **streams,
+                                                         uint32_t *gpu_indexes,
+                                                         uint32_t gpu_count,
+                                                         int8_t **mem_ptr_void);
+}
 
 template <typename Torus> struct int_compression {
   int_radix_params compression_params;
@@ -15,7 +54,7 @@ template <typename Torus> struct int_compression {
   Torus *tmp_lwe;
   Torus *tmp_glwe_array_out;
 
-  int_compression(cudaStream_t const *streams, uint32_t const *gpu_indexes,
+  int_compression(cudaStream_t *streams, uint32_t *gpu_indexes,
                   uint32_t gpu_count, int_radix_params compression_params,
                   uint32_t num_radix_blocks, uint32_t lwe_per_glwe,
                   uint32_t storage_log_modulus, bool allocate_gpu_memory) {
@@ -38,12 +77,11 @@ template <typename Torus> struct int_compression {
 
       scratch_packing_keyswitch_lwe_list_to_glwe_64(
           streams[0], gpu_indexes[0], &fp_ks_buffer,
-          compression_params.small_lwe_dimension,
           compression_params.glwe_dimension, compression_params.polynomial_size,
           num_radix_blocks, true);
     }
   }
-  void release(cudaStream_t const *streams, uint32_t const *gpu_indexes,
+  void release(cudaStream_t *streams, uint32_t *gpu_indexes,
                uint32_t gpu_count) {
     cuda_drop_async(tmp_lwe, streams[0], gpu_indexes[0]);
     cuda_drop_async(tmp_glwe_array_out, streams[0], gpu_indexes[0]);
@@ -65,9 +103,9 @@ template <typename Torus> struct int_decompression {
   Torus *tmp_extracted_lwe;
   uint32_t *tmp_indexes_array;
 
-  int_radix_lut<Torus> *decompression_rescale_lut;
+  int_radix_lut<Torus> *carry_extract_lut;
 
-  int_decompression(cudaStream_t const *streams, uint32_t const *gpu_indexes,
+  int_decompression(cudaStream_t *streams, uint32_t *gpu_indexes,
                     uint32_t gpu_count, int_radix_params encryption_params,
                     int_radix_params compression_params,
                     uint32_t num_radix_blocks, uint32_t body_count,
@@ -84,7 +122,7 @@ template <typename Torus> struct int_decompression {
       Torus lwe_accumulator_size = (compression_params.glwe_dimension *
                                         compression_params.polynomial_size +
                                     1);
-      decompression_rescale_lut = new int_radix_lut<Torus>(
+      carry_extract_lut = new int_radix_lut<Torus>(
           streams, gpu_indexes, gpu_count, encryption_params, 1,
           num_radix_blocks, allocate_gpu_memory);
 
@@ -97,38 +135,29 @@ template <typename Torus> struct int_decompression {
           num_radix_blocks * lwe_accumulator_size * sizeof(Torus), streams[0],
           gpu_indexes[0]);
 
-      // Rescale is done using an identity LUT
-      // Here we do not divide by message_modulus
-      // Example: in the 2_2 case we are mapping a 2 bits message onto a 4 bits
-      // space, we want to keep the original 2 bits value in the 4 bits space,
-      // so we apply the identity and the encoding will rescale it for us.
-      auto decompression_rescale_f = [encryption_params](Torus x) -> Torus {
-        return x;
+      // Carry extract LUT
+      auto carry_extract_f = [encryption_params](Torus x) -> Torus {
+        return x / encryption_params.message_modulus;
       };
 
-      auto effective_compression_message_modulus =
-          encryption_params.carry_modulus;
-      auto effective_compression_carry_modulus = 1;
-
-      generate_device_accumulator_with_encoding<Torus>(
-          streams[0], gpu_indexes[0], decompression_rescale_lut->get_lut(0, 0),
+      generate_device_accumulator<Torus>(
+          streams[0], gpu_indexes[0],
+          carry_extract_lut->get_lut(gpu_indexes[0], 0),
           encryption_params.glwe_dimension, encryption_params.polynomial_size,
-          effective_compression_message_modulus,
-          effective_compression_carry_modulus,
           encryption_params.message_modulus, encryption_params.carry_modulus,
-          decompression_rescale_f);
+          carry_extract_f);
 
-      decompression_rescale_lut->broadcast_lut(streams, gpu_indexes, 0);
+      carry_extract_lut->broadcast_lut(streams, gpu_indexes, gpu_indexes[0]);
     }
   }
-  void release(cudaStream_t const *streams, uint32_t const *gpu_indexes,
+  void release(cudaStream_t *streams, uint32_t *gpu_indexes,
                uint32_t gpu_count) {
     cuda_drop_async(tmp_extracted_glwe, streams[0], gpu_indexes[0]);
     cuda_drop_async(tmp_extracted_lwe, streams[0], gpu_indexes[0]);
     cuda_drop_async(tmp_indexes_array, streams[0], gpu_indexes[0]);
 
-    decompression_rescale_lut->release(streams, gpu_indexes, gpu_count);
-    delete decompression_rescale_lut;
+    carry_extract_lut->release(streams, gpu_indexes, gpu_count);
+    delete carry_extract_lut;
   }
 };
 #endif

backends/tfhe-cuda-backend/cuda/include/device.h

@@ -27,23 +27,12 @@ inline void cuda_error(cudaError_t code, const char *file, int line) {
     std::abort();                                                              \
   }
 
-cudaEvent_t cuda_create_event(uint32_t gpu_index);
-
-void cuda_event_record(cudaEvent_t event, cudaStream_t stream,
-                       uint32_t gpu_index);
-void cuda_stream_wait_event(cudaStream_t stream, cudaEvent_t event,
-                            uint32_t gpu_index);
-
-void cuda_event_destroy(cudaEvent_t event, uint32_t gpu_index);
-
 cudaStream_t cuda_create_stream(uint32_t gpu_index);
 
 void cuda_destroy_stream(cudaStream_t stream, uint32_t gpu_index);
 
 void cuda_synchronize_stream(cudaStream_t stream, uint32_t gpu_index);
 
-uint32_t cuda_is_available();
-
 void *cuda_malloc(uint64_t size, uint32_t gpu_index);
 
 void *cuda_malloc_async(uint64_t size, cudaStream_t stream, uint32_t gpu_index);
@@ -53,7 +42,7 @@ void cuda_check_valid_malloc(uint64_t size, uint32_t gpu_index);
 void cuda_memcpy_async_to_gpu(void *dest, void *src, uint64_t size,
                               cudaStream_t stream, uint32_t gpu_index);
 
-void cuda_memcpy_async_gpu_to_gpu(void *dest, void const *src, uint64_t size,
+void cuda_memcpy_async_gpu_to_gpu(void *dest, void *src, uint64_t size,
                                   cudaStream_t stream, uint32_t gpu_index);
 
 void cuda_memcpy_gpu_to_gpu(void *dest, void *src, uint64_t size,

backends/tfhe-cuda-backend/cuda/include/integer/compression/compression.h

@@ -1,45 +0,0 @@
-#ifndef CUDA_INTEGER_COMPRESSION_H
-#define CUDA_INTEGER_COMPRESSION_H
-
-#include "../../pbs/pbs_enums.h"
-
-extern "C" {
-void scratch_cuda_integer_compress_radix_ciphertext_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t compression_glwe_dimension,
-    uint32_t compression_polynomial_size, uint32_t lwe_dimension,
-    uint32_t ks_level, uint32_t ks_base_log, uint32_t num_radix_blocks,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    uint32_t lwe_per_glwe, uint32_t storage_log_modulus,
-    bool allocate_gpu_memory);
-
-void scratch_cuda_integer_decompress_radix_ciphertext_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t encryption_glwe_dimension,
-    uint32_t encryption_polynomial_size, uint32_t compression_glwe_dimension,
-    uint32_t compression_polynomial_size, uint32_t lwe_dimension,
-    uint32_t pbs_level, uint32_t pbs_base_log, uint32_t num_radix_blocks,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    uint32_t storage_log_modulus, uint32_t body_count,
-    bool allocate_gpu_memory);
-
-void cuda_integer_compress_radix_ciphertext_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *glwe_array_out, void const *lwe_array_in, void *const *fp_ksk,
-    uint32_t num_nths, int8_t *mem_ptr);
-
-void cuda_integer_decompress_radix_ciphertext_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array_out, void const *glwe_in, uint32_t const *indexes_array,
-    uint32_t indexes_array_size, void *const *bsks, int8_t *mem_ptr);
-
-void cleanup_cuda_integer_compress_radix_ciphertext_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr_void);
-
-void cleanup_cuda_integer_decompress_radix_ciphertext_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr_void);
-}
-
-#endif

backends/tfhe-cuda-backend/cuda/include/integer/integer.h

@@ -1,487 +0,0 @@
-#ifndef CUDA_INTEGER_H
-#define CUDA_INTEGER_H
-
-#include "../pbs/pbs_enums.h"
-#include <stdint.h>
-
-enum OUTPUT_CARRY { NONE = 0, GENERATED = 1, PROPAGATED = 2 };
-enum SHIFT_OR_ROTATE_TYPE {
-  LEFT_SHIFT = 0,
-  RIGHT_SHIFT = 1,
-  LEFT_ROTATE = 2,
-  RIGHT_ROTATE = 3
-};
-enum BITOP_TYPE {
-  BITAND = 0,
-  BITOR = 1,
-  BITXOR = 2,
-  SCALAR_BITAND = 3,
-  SCALAR_BITOR = 4,
-  SCALAR_BITXOR = 5,
-};
-
-enum COMPARISON_TYPE {
-  EQ = 0,
-  NE = 1,
-  GT = 2,
-  GE = 3,
-  LT = 4,
-  LE = 5,
-  MAX = 6,
-  MIN = 7,
-};
-
-enum CMP_ORDERING { IS_INFERIOR = 0, IS_EQUAL = 1, IS_SUPERIOR = 2 };
-
-enum SIGNED_OPERATION { ADDITION = 1, SUBTRACTION = -1 };
-
-enum outputFlag { FLAG_NONE = 0, FLAG_OVERFLOW = 1, FLAG_CARRY = 2 };
-
-extern "C" {
-void scratch_cuda_apply_univariate_lut_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, void const *input_lut, uint32_t lwe_dimension,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t input_lwe_ciphertext_count,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    bool allocate_gpu_memory);
-void scratch_cuda_apply_many_univariate_lut_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, void const *input_lut, uint32_t lwe_dimension,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t num_radix_blocks,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    uint32_t num_many_lut, bool allocate_gpu_memory);
-void cuda_apply_univariate_lut_kb_64(void *const *streams,
-                                     uint32_t const *gpu_indexes,
-                                     uint32_t gpu_count, void *output_radix_lwe,
-                                     void const *input_radix_lwe,
-                                     int8_t *mem_ptr, void *const *ksks,
-                                     void *const *bsks, uint32_t num_blocks);
-
-void cleanup_cuda_apply_univariate_lut_kb_64(void *const *streams,
-                                             uint32_t const *gpu_indexes,
-                                             uint32_t gpu_count,
-                                             int8_t **mem_ptr_void);
-
-void scratch_cuda_apply_bivariate_lut_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, void const *input_lut, uint32_t lwe_dimension,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t input_lwe_ciphertext_count,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    bool allocate_gpu_memory);
-
-void cuda_apply_bivariate_lut_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *output_radix_lwe, void const *input_radix_lwe_1,
-    void const *input_radix_lwe_2, int8_t *mem_ptr, void *const *ksks,
-    void *const *bsks, uint32_t num_blocks, uint32_t shift);
-
-void cleanup_cuda_apply_bivariate_lut_kb_64(void *const *streams,
-                                            uint32_t const *gpu_indexes,
-                                            uint32_t gpu_count,
-                                            int8_t **mem_ptr_void);
-
-void cuda_apply_many_univariate_lut_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *output_radix_lwe, void const *input_radix_lwe, int8_t *mem_ptr,
-    void *const *ksks, void *const *bsks, uint32_t num_blocks,
-    uint32_t num_luts, uint32_t lut_stride);
-
-void scratch_cuda_full_propagation_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t ks_level, uint32_t ks_base_log,
-    uint32_t pbs_level, uint32_t pbs_base_log, uint32_t grouping_factor,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    bool allocate_gpu_memory);
-
-void cuda_full_propagation_64_inplace(void *const *streams,
-                                      uint32_t const *gpu_indexes,
-                                      uint32_t gpu_count, void *input_blocks,
-                                      int8_t *mem_ptr, void *const *ksks,
-                                      void *const *bsks, uint32_t num_blocks);
-
-void cleanup_cuda_full_propagation(void *const *streams,
-                                   uint32_t const *gpu_indexes,
-                                   uint32_t gpu_count, int8_t **mem_ptr_void);
-
-void scratch_cuda_integer_mult_radix_ciphertext_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, bool const is_boolean_left, bool const is_boolean_right,
-    uint32_t message_modulus, uint32_t carry_modulus, uint32_t glwe_dimension,
-    uint32_t lwe_dimension, uint32_t polynomial_size, uint32_t pbs_base_log,
-    uint32_t pbs_level, uint32_t ks_base_log, uint32_t ks_level,
-    uint32_t grouping_factor, uint32_t num_blocks, PBS_TYPE pbs_type,
-    bool allocate_gpu_memory);
-
-void cuda_integer_mult_radix_ciphertext_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *radix_lwe_out, void const *radix_lwe_left, bool const is_bool_left,
-    void const *radix_lwe_right, bool const is_bool_right, void *const *bsks,
-    void *const *ksks, int8_t *mem_ptr, uint32_t polynomial_size,
-    uint32_t num_blocks);
-
-void cleanup_cuda_integer_mult(void *const *streams,
-                               uint32_t const *gpu_indexes, uint32_t gpu_count,
-                               int8_t **mem_ptr_void);
-
-void cuda_negate_integer_radix_ciphertext_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array_out, void const *lwe_array_in, uint32_t lwe_dimension,
-    uint32_t lwe_ciphertext_count, uint32_t message_modulus,
-    uint32_t carry_modulus);
-
-void cuda_scalar_addition_integer_radix_ciphertext_64_inplace(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array, void const *scalar_input, uint32_t lwe_dimension,
-    uint32_t lwe_ciphertext_count, uint32_t message_modulus,
-    uint32_t carry_modulus);
-
-void scratch_cuda_integer_radix_logical_scalar_shift_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
-    uint32_t carry_modulus, PBS_TYPE pbs_type, SHIFT_OR_ROTATE_TYPE shift_type,
-    bool allocate_gpu_memory);
-
-void cuda_integer_radix_logical_scalar_shift_kb_64_inplace(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array, uint32_t shift, int8_t *mem_ptr, void *const *bsks,
-    void *const *ksks, uint32_t num_blocks);
-
-void scratch_cuda_integer_radix_arithmetic_scalar_shift_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
-    uint32_t carry_modulus, PBS_TYPE pbs_type, SHIFT_OR_ROTATE_TYPE shift_type,
-    bool allocate_gpu_memory);
-
-void cuda_integer_radix_arithmetic_scalar_shift_kb_64_inplace(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array, uint32_t shift, int8_t *mem_ptr, void *const *bsks,
-    void *const *ksks, uint32_t num_blocks);
-
-void cleanup_cuda_integer_radix_logical_scalar_shift(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr_void);
-
-void cleanup_cuda_integer_radix_arithmetic_scalar_shift(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr_void);
-
-void scratch_cuda_integer_radix_shift_and_rotate_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
-    uint32_t carry_modulus, PBS_TYPE pbs_type, SHIFT_OR_ROTATE_TYPE shift_type,
-    bool is_signed, bool allocate_gpu_memory);
-
-void cuda_integer_radix_shift_and_rotate_kb_64_inplace(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array, void const *lwe_shift, int8_t *mem_ptr, void *const *bsks,
-    void *const *ksks, uint32_t num_blocks);
-
-void cleanup_cuda_integer_radix_shift_and_rotate(void *const *streams,
-                                                 uint32_t const *gpu_indexes,
-                                                 uint32_t gpu_count,
-                                                 int8_t **mem_ptr_void);
-
-void scratch_cuda_integer_radix_comparison_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t lwe_ciphertext_count,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    COMPARISON_TYPE op_type, bool is_signed, bool allocate_gpu_memory);
-
-void cuda_comparison_integer_radix_ciphertext_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array_out, void const *lwe_array_1, void const *lwe_array_2,
-    int8_t *mem_ptr, void *const *bsks, void *const *ksks,
-    uint32_t lwe_ciphertext_count);
-
-void cuda_scalar_comparison_integer_radix_ciphertext_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array_out, void const *lwe_array_in, void const *scalar_blocks,
-    int8_t *mem_ptr, void *const *bsks, void *const *ksks,
-    uint32_t lwe_ciphertext_count, uint32_t num_scalar_blocks);
-
-void cleanup_cuda_integer_comparison(void *const *streams,
-                                     uint32_t const *gpu_indexes,
-                                     uint32_t gpu_count, int8_t **mem_ptr_void);
-
-void scratch_cuda_integer_radix_bitop_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t lwe_ciphertext_count,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    BITOP_TYPE op_type, bool allocate_gpu_memory);
-
-void cuda_bitop_integer_radix_ciphertext_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array_out, void const *lwe_array_1, void const *lwe_array_2,
-    int8_t *mem_ptr, void *const *bsks, void *const *ksks,
-    uint32_t lwe_ciphertext_count);
-
-void cuda_scalar_bitop_integer_radix_ciphertext_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array_out, void const *lwe_array_input, void const *clear_blocks,
-    uint32_t num_clear_blocks, int8_t *mem_ptr, void *const *bsks,
-    void *const *ksks, uint32_t lwe_ciphertext_count, BITOP_TYPE op);
-
-void cleanup_cuda_integer_bitop(void *const *streams,
-                                uint32_t const *gpu_indexes, uint32_t gpu_count,
-                                int8_t **mem_ptr_void);
-
-void scratch_cuda_integer_radix_cmux_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t lwe_ciphertext_count,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    bool allocate_gpu_memory);
-
-void cuda_cmux_integer_radix_ciphertext_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array_out, void const *lwe_condition, void const *lwe_array_true,
-    void const *lwe_array_false, int8_t *mem_ptr, void *const *bsks,
-    void *const *ksks, uint32_t lwe_ciphertext_count);
-
-void cleanup_cuda_integer_radix_cmux(void *const *streams,
-                                     uint32_t const *gpu_indexes,
-                                     uint32_t gpu_count, int8_t **mem_ptr_void);
-
-void scratch_cuda_integer_radix_scalar_rotate_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
-    uint32_t carry_modulus, PBS_TYPE pbs_type, SHIFT_OR_ROTATE_TYPE shift_type,
-    bool allocate_gpu_memory);
-
-void cuda_integer_radix_scalar_rotate_kb_64_inplace(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array, uint32_t n, int8_t *mem_ptr, void *const *bsks,
-    void *const *ksks, uint32_t num_blocks);
-
-void cleanup_cuda_integer_radix_scalar_rotate(void *const *streams,
-                                              uint32_t const *gpu_indexes,
-                                              uint32_t gpu_count,
-                                              int8_t **mem_ptr_void);
-
-void scratch_cuda_propagate_single_carry_kb_64_inplace(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
-    uint32_t carry_modulus, PBS_TYPE pbs_type, uint32_t requested_flag,
-    uint32_t uses_carry, bool allocate_gpu_memory);
-
-void scratch_cuda_add_and_propagate_single_carry_kb_64_inplace(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
-    uint32_t carry_modulus, PBS_TYPE pbs_type, uint32_t requested_flag,
-    uint32_t uses_carry, bool allocate_gpu_memory);
-
-void cuda_propagate_single_carry_kb_64_inplace(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array, void *carry_out, const void *carry_in, int8_t *mem_ptr,
-    void *const *bsks, void *const *ksks, uint32_t num_blocks,
-    uint32_t requested_flag, uint32_t uses_carry);
-
-void cuda_add_and_propagate_single_carry_kb_64_inplace(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lhs_array, const void *rhs_array, void *carry_out,
-    const void *carry_in, int8_t *mem_ptr, void *const *bsks, void *const *ksks,
-    uint32_t num_blocks, uint32_t requested_flag, uint32_t uses_carry);
-
-void cleanup_cuda_propagate_single_carry(void *const *streams,
-                                         uint32_t const *gpu_indexes,
-                                         uint32_t gpu_count,
-                                         int8_t **mem_ptr_void);
-
-void cleanup_cuda_add_and_propagate_single_carry(void *const *streams,
-                                                 uint32_t const *gpu_indexes,
-                                                 uint32_t gpu_count,
-                                                 int8_t **mem_ptr_void);
-
-void scratch_cuda_integer_overflowing_sub_kb_64_inplace(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
-    uint32_t carry_modulus, PBS_TYPE pbs_type, uint32_t compute_overflow,
-    bool allocate_gpu_memory);
-
-void cuda_integer_overflowing_sub_kb_64_inplace(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lhs_array, const void *rhs_array, void *overflow_block,
-    const void *input_borrow, int8_t *mem_ptr, void *const *bsks,
-    void *const *ksks, uint32_t num_blocks, uint32_t compute_overflow,
-    uint32_t uses_input_borrow);
-
-void cleanup_cuda_integer_overflowing_sub(void *const *streams,
-                                          uint32_t const *gpu_indexes,
-                                          uint32_t gpu_count,
-                                          int8_t **mem_ptr_void);
-
-void scratch_cuda_integer_radix_partial_sum_ciphertexts_vec_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t lwe_dimension, uint32_t ks_level, uint32_t ks_base_log,
-    uint32_t pbs_level, uint32_t pbs_base_log, uint32_t grouping_factor,
-    uint32_t num_blocks_in_radix, uint32_t max_num_radix_in_vec,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    bool allocate_gpu_memory);
-
-void cuda_integer_radix_partial_sum_ciphertexts_vec_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *radix_lwe_out, void *radix_lwe_vec, uint32_t num_radix_in_vec,
-    int8_t *mem_ptr, void *const *bsks, void *const *ksks,
-    uint32_t num_blocks_in_radix);
-
-void cleanup_cuda_integer_radix_partial_sum_ciphertexts_vec(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr_void);
-
-void scratch_cuda_integer_scalar_mul_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t lwe_dimension, uint32_t ks_level, uint32_t ks_base_log,
-    uint32_t pbs_level, uint32_t pbs_base_log, uint32_t grouping_factor,
-    uint32_t num_blocks, uint32_t message_modulus, uint32_t carry_modulus,
-    PBS_TYPE pbs_type, bool allocate_gpu_memory);
-
-void cuda_scalar_multiplication_integer_radix_ciphertext_64_inplace(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array, uint64_t const *decomposed_scalar,
-    uint64_t const *has_at_least_one_set, int8_t *mem_ptr, void *const *bsks,
-    void *const *ksks, uint32_t lwe_dimension, uint32_t polynomial_size,
-    uint32_t message_modulus, uint32_t num_blocks, uint32_t num_scalars);
-
-void cleanup_cuda_integer_radix_scalar_mul(void *const *streams,
-                                           uint32_t const *gpu_indexes,
-                                           uint32_t gpu_count,
-                                           int8_t **mem_ptr_void);
-
-void scratch_cuda_integer_div_rem_radix_ciphertext_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    bool is_signed, int8_t **mem_ptr, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t big_lwe_dimension,
-    uint32_t small_lwe_dimension, uint32_t ks_level, uint32_t ks_base_log,
-    uint32_t pbs_level, uint32_t pbs_base_log, uint32_t grouping_factor,
-    uint32_t num_blocks, uint32_t message_modulus, uint32_t carry_modulus,
-    PBS_TYPE pbs_type, bool allocate_gpu_memory);
-
-void cuda_integer_div_rem_radix_ciphertext_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *quotient, void *remainder, void const *numerator, void const *divisor,
-    bool is_signed, int8_t *mem_ptr, void *const *bsks, void *const *ksks,
-    uint32_t num_blocks_in_radix);
-
-void cleanup_cuda_integer_div_rem(void *const *streams,
-                                  uint32_t const *gpu_indexes,
-                                  uint32_t gpu_count, int8_t **mem_ptr_void);
-
-void scratch_cuda_integer_compute_prefix_sum_hillis_steele_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, void const *input_lut, uint32_t lwe_dimension,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t num_radix_blocks,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    bool allocate_gpu_memory);
-
-void cuda_integer_compute_prefix_sum_hillis_steele_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *output_radix_lwe, void *generates_or_propagates, int8_t *mem_ptr,
-    void *const *ksks, void *const *bsks, uint32_t num_blocks, uint32_t shift);
-
-void cleanup_cuda_integer_compute_prefix_sum_hillis_steele_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr_void);
-
-void cuda_integer_reverse_blocks_64_inplace(void *const *streams,
-                                            uint32_t const *gpu_indexes,
-                                            uint32_t gpu_count, void *lwe_array,
-                                            uint32_t num_blocks,
-                                            uint32_t lwe_size);
-
-void scratch_cuda_integer_abs_inplace_radix_ciphertext_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, bool is_signed, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t big_lwe_dimension,
-    uint32_t small_lwe_dimension, uint32_t ks_level, uint32_t ks_base_log,
-    uint32_t pbs_level, uint32_t pbs_base_log, uint32_t grouping_factor,
-    uint32_t num_blocks, uint32_t message_modulus, uint32_t carry_modulus,
-    PBS_TYPE pbs_type, bool allocate_gpu_memory);
-
-void cuda_integer_abs_inplace_radix_ciphertext_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *ct, int8_t *mem_ptr, bool is_signed, void *const *bsks,
-    void *const *ksks, uint32_t num_blocks);
-
-void cleanup_cuda_integer_abs_inplace(void *const *streams,
-                                      uint32_t const *gpu_indexes,
-                                      uint32_t gpu_count,
-                                      int8_t **mem_ptr_void);
-
-void scratch_cuda_integer_are_all_comparisons_block_true_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t num_radix_blocks,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    bool allocate_gpu_memory);
-
-void cuda_integer_are_all_comparisons_block_true_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array_out, void const *lwe_array_in, int8_t *mem_ptr,
-    void *const *bsks, void *const *ksks, uint32_t num_radix_blocks);
-
-void cleanup_cuda_integer_are_all_comparisons_block_true(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr_void);
-
-void scratch_cuda_integer_is_at_least_one_comparisons_block_true_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
-    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
-    uint32_t grouping_factor, uint32_t num_radix_blocks,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    bool allocate_gpu_memory);
-
-void cuda_integer_is_at_least_one_comparisons_block_true_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array_out, void const *lwe_array_in, int8_t *mem_ptr,
-    void *const *bsks, void *const *ksks, uint32_t num_radix_blocks);
-
-void cleanup_cuda_integer_is_at_least_one_comparisons_block_true(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    int8_t **mem_ptr_void);
-
-} // extern C
-#endif // CUDA_INTEGER_H

backends/tfhe-cuda-backend/cuda/include/keyswitch.h

@@ -1,35 +1,32 @@
 #ifndef CNCRT_KS_H_
 #define CNCRT_KS_H_
 
-#include <stdint.h>
+#include <cstdint>
 
 extern "C" {
 
 void cuda_keyswitch_lwe_ciphertext_vector_32(
     void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void const *lwe_output_indexes, void const *lwe_array_in,
-    void const *lwe_input_indexes, void const *ksk, uint32_t lwe_dimension_in,
-    uint32_t lwe_dimension_out, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples);
+    void *lwe_output_indexes, void *lwe_array_in, void *lwe_input_indexes,
+    void *ksk, uint32_t lwe_dimension_in, uint32_t lwe_dimension_out,
+    uint32_t base_log, uint32_t level_count, uint32_t num_samples);
 
 void cuda_keyswitch_lwe_ciphertext_vector_64(
     void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void const *lwe_output_indexes, void const *lwe_array_in,
-    void const *lwe_input_indexes, void const *ksk, uint32_t lwe_dimension_in,
-    uint32_t lwe_dimension_out, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples);
+    void *lwe_output_indexes, void *lwe_array_in, void *lwe_input_indexes,
+    void *ksk, uint32_t lwe_dimension_in, uint32_t lwe_dimension_out,
+    uint32_t base_log, uint32_t level_count, uint32_t num_samples);
 
 void scratch_packing_keyswitch_lwe_list_to_glwe_64(
     void *stream, uint32_t gpu_index, int8_t **fp_ks_buffer,
-    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t num_lwes, bool allocate_gpu_memory);
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t num_lwes,
+    bool allocate_gpu_memory);
 
 void cuda_packing_keyswitch_lwe_list_to_glwe_64(
-    void *stream, uint32_t gpu_index, void *glwe_array_out,
-    void const *lwe_array_in, void const *fp_ksk_array, int8_t *fp_ks_buffer,
-    uint32_t input_lwe_dimension, uint32_t output_glwe_dimension,
-    uint32_t output_polynomial_size, uint32_t base_log, uint32_t level_count,
-    uint32_t num_lwes);
+    void *stream, uint32_t gpu_index, void *glwe_array_out, void *lwe_array_in,
+    void *fp_ksk_array, int8_t *fp_ks_buffer, uint32_t input_lwe_dimension,
+    uint32_t output_glwe_dimension, uint32_t output_polynomial_size,
+    uint32_t base_log, uint32_t level_count, uint32_t num_lwes);
 
 void cleanup_packing_keyswitch_lwe_list_to_glwe(void *stream,
                                                 uint32_t gpu_index,

backends/tfhe-cuda-backend/cuda/include/linear_algebra.h

@@ -1,49 +1,50 @@
 #ifndef CUDA_LINALG_H_
 #define CUDA_LINALG_H_
 
-#include <stdint.h>
+#include "programmable_bootstrap.h"
+#include <cstdint>
+#include <device.h>
 
 extern "C" {
 
 void cuda_negate_lwe_ciphertext_vector_32(void *stream, uint32_t gpu_index,
                                           void *lwe_array_out,
-                                          void const *lwe_array_in,
+                                          void *lwe_array_in,
                                           uint32_t input_lwe_dimension,
                                           uint32_t input_lwe_ciphertext_count);
 void cuda_negate_lwe_ciphertext_vector_64(void *stream, uint32_t gpu_index,
                                           void *lwe_array_out,
-                                          void const *lwe_array_in,
+                                          void *lwe_array_in,
                                           uint32_t input_lwe_dimension,
                                           uint32_t input_lwe_ciphertext_count);
 void cuda_add_lwe_ciphertext_vector_32(void *stream, uint32_t gpu_index,
                                        void *lwe_array_out,
-                                       void const *lwe_array_in_1,
-                                       void const *lwe_array_in_2,
+                                       void *lwe_array_in_1,
+                                       void *lwe_array_in_2,
                                        uint32_t input_lwe_dimension,
                                        uint32_t input_lwe_ciphertext_count);
 void cuda_add_lwe_ciphertext_vector_64(void *stream, uint32_t gpu_index,
                                        void *lwe_array_out,
-                                       void const *lwe_array_in_1,
-                                       void const *lwe_array_in_2,
+                                       void *lwe_array_in_1,
+                                       void *lwe_array_in_2,
                                        uint32_t input_lwe_dimension,
                                        uint32_t input_lwe_ciphertext_count);
-
 void cuda_add_lwe_ciphertext_vector_plaintext_vector_32(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void const *lwe_array_in, void const *plaintext_array_in,
-    uint32_t input_lwe_dimension, uint32_t input_lwe_ciphertext_count);
+    void *stream, uint32_t gpu_index, void *lwe_array_out, void *lwe_array_in,
+    void *plaintext_array_in, uint32_t input_lwe_dimension,
+    uint32_t input_lwe_ciphertext_count);
 void cuda_add_lwe_ciphertext_vector_plaintext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void const *lwe_array_in, void const *plaintext_array_in,
-    uint32_t input_lwe_dimension, uint32_t input_lwe_ciphertext_count);
+    void *stream, uint32_t gpu_index, void *lwe_array_out, void *lwe_array_in,
+    void *plaintext_array_in, uint32_t input_lwe_dimension,
+    uint32_t input_lwe_ciphertext_count);
 void cuda_mult_lwe_ciphertext_vector_cleartext_vector_32(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void const *lwe_array_in, void const *cleartext_array_in,
-    uint32_t input_lwe_dimension, uint32_t input_lwe_ciphertext_count);
+    void *stream, uint32_t gpu_index, void *lwe_array_out, void *lwe_array_in,
+    void *cleartext_array_in, uint32_t input_lwe_dimension,
+    uint32_t input_lwe_ciphertext_count);
 void cuda_mult_lwe_ciphertext_vector_cleartext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void const *lwe_array_in, void const *cleartext_array_in,
-    uint32_t input_lwe_dimension, uint32_t input_lwe_ciphertext_count);
+    void *stream, uint32_t gpu_index, void *lwe_array_out, void *lwe_array_in,
+    void *cleartext_array_in, uint32_t input_lwe_dimension,
+    uint32_t input_lwe_ciphertext_count);
 }
 
 #endif // CUDA_LINALG_H_

backends/tfhe-cuda-backend/cuda/include/pbs/pbs_enums.h

@@ -1,7 +0,0 @@
-#ifndef CUDA_PBS_ENUMS_H
-#define CUDA_PBS_ENUMS_H
-
-enum PBS_TYPE { MULTI_BIT = 0, CLASSICAL = 1 };
-enum PBS_VARIANT { DEFAULT = 0, CG = 1, TBC = 2 };
-
-#endif // CUDA_PBS_ENUMS_H

backends/tfhe-cuda-backend/cuda/include/pbs/programmable_bootstrap.h

@@ -1,86 +0,0 @@
-#ifndef CUDA_BOOTSTRAP_H
-#define CUDA_BOOTSTRAP_H
-
-#include "pbs_enums.h"
-#include <stdint.h>
-
-extern "C" {
-void cuda_fourier_polynomial_mul(void *stream, uint32_t gpu_index,
-                                 void const *input1, void const *input2,
-                                 void *output, uint32_t polynomial_size,
-                                 uint32_t total_polynomials);
-
-void cuda_convert_lwe_programmable_bootstrap_key_32(
-    void *stream, uint32_t gpu_index, void *dest, void const *src,
-    uint32_t input_lwe_dim, uint32_t glwe_dim, uint32_t level_count,
-    uint32_t polynomial_size);
-
-void cuda_convert_lwe_programmable_bootstrap_key_64(
-    void *stream, uint32_t gpu_index, void *dest, void const *src,
-    uint32_t input_lwe_dim, uint32_t glwe_dim, uint32_t level_count,
-    uint32_t polynomial_size);
-
-void scratch_cuda_programmable_bootstrap_amortized_32(
-    void *stream, uint32_t gpu_index, int8_t **pbs_buffer,
-    uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-void scratch_cuda_programmable_bootstrap_amortized_64(
-    void *stream, uint32_t gpu_index, int8_t **pbs_buffer,
-    uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-void cuda_programmable_bootstrap_amortized_lwe_ciphertext_vector_32(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void const *lwe_output_indexes, void const *lut_vector,
-    void const *lut_vector_indexes, void const *lwe_array_in,
-    void const *lwe_input_indexes, void const *bootstrapping_key,
-    int8_t *pbs_buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples);
-
-void cuda_programmable_bootstrap_amortized_lwe_ciphertext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void const *lwe_output_indexes, void const *lut_vector,
-    void const *lut_vector_indexes, void const *lwe_array_in,
-    void const *lwe_input_indexes, void const *bootstrapping_key,
-    int8_t *pbs_buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples);
-
-void cleanup_cuda_programmable_bootstrap_amortized(void *stream,
-                                                   uint32_t gpu_index,
-                                                   int8_t **pbs_buffer);
-
-void scratch_cuda_programmable_bootstrap_32(
-    void *stream, uint32_t gpu_index, int8_t **buffer, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t level_count,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-void scratch_cuda_programmable_bootstrap_64(
-    void *stream, uint32_t gpu_index, int8_t **buffer, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t level_count,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-void cuda_programmable_bootstrap_lwe_ciphertext_vector_32(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void const *lwe_output_indexes, void const *lut_vector,
-    void const *lut_vector_indexes, void const *lwe_array_in,
-    void const *lwe_input_indexes, void const *bootstrapping_key,
-    int8_t *buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples, uint32_t num_many_lut, uint32_t lut_stride);
-
-void cuda_programmable_bootstrap_lwe_ciphertext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void const *lwe_output_indexes, void const *lut_vector,
-    void const *lut_vector_indexes, void const *lwe_array_in,
-    void const *lwe_input_indexes, void const *bootstrapping_key,
-    int8_t *buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples, uint32_t num_many_lut, uint32_t lut_stride);
-
-void cleanup_cuda_programmable_bootstrap(void *stream, uint32_t gpu_index,
-                                         int8_t **pbs_buffer);
-}
-#endif // CUDA_BOOTSTRAP_H

backends/tfhe-cuda-backend/cuda/include/pbs/programmable_bootstrap_multibit.h

@@ -1,38 +0,0 @@
-#ifndef CUDA_MULTI_BIT_H
-#define CUDA_MULTI_BIT_H
-
-#include "pbs_enums.h"
-#include "stdint.h"
-
-extern "C" {
-
-bool has_support_to_cuda_programmable_bootstrap_cg_multi_bit(
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
-    uint32_t num_samples);
-
-void cuda_convert_lwe_multi_bit_programmable_bootstrap_key_64(
-    void *stream, uint32_t gpu_index, void *dest, void const *src,
-    uint32_t input_lwe_dim, uint32_t glwe_dim, uint32_t level_count,
-    uint32_t polynomial_size, uint32_t grouping_factor);
-
-void scratch_cuda_multi_bit_programmable_bootstrap_64(
-    void *stream, uint32_t gpu_index, int8_t **pbs_buffer,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-void cuda_multi_bit_programmable_bootstrap_lwe_ciphertext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void const *lwe_output_indexes, void const *lut_vector,
-    void const *lut_vector_indexes, void const *lwe_array_in,
-    void const *lwe_input_indexes, void const *bootstrapping_key,
-    int8_t *buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t grouping_factor, uint32_t base_log,
-    uint32_t level_count, uint32_t num_samples, uint32_t num_many_lut,
-    uint32_t lut_stride);
-
-void cleanup_cuda_multi_bit_programmable_bootstrap(void *stream,
-                                                   uint32_t gpu_index,
-                                                   int8_t **pbs_buffer);
-}
-
-#endif // CUDA_MULTI_BIT_H

backends/tfhe-cuda-backend/cuda/include/programmable_bootstrap.h

@@ -1,10 +1,87 @@
-#ifndef CUDA_BOOTSTRAP_UTILITIES_H
-#define CUDA_BOOTSTRAP_UTILITIES_H
+#ifndef CUDA_BOOTSTRAP_H
+#define CUDA_BOOTSTRAP_H
 
 #include "device.h"
-#include "pbs_enums.h"
-#include "vector_types.h"
-#include <stdint.h>
+#include <cstdint>
+
+enum PBS_TYPE { MULTI_BIT = 0, CLASSICAL = 1 };
+enum PBS_VARIANT { DEFAULT = 0, CG = 1, TBC = 2 };
+
+extern "C" {
+void cuda_fourier_polynomial_mul(cudaStream_t stream, uint32_t gpu_index,
+                                 void *input1, void *input2, void *output,
+                                 uint32_t polynomial_size,
+                                 uint32_t total_polynomials);
+
+void cuda_convert_lwe_programmable_bootstrap_key_32(
+    void *stream, uint32_t gpu_index, void *dest, void *src,
+    uint32_t input_lwe_dim, uint32_t glwe_dim, uint32_t level_count,
+    uint32_t polynomial_size);
+
+void cuda_convert_lwe_programmable_bootstrap_key_64(
+    void *stream, uint32_t gpu_index, void *dest, void *src,
+    uint32_t input_lwe_dim, uint32_t glwe_dim, uint32_t level_count,
+    uint32_t polynomial_size);
+
+void scratch_cuda_programmable_bootstrap_amortized_32(
+    void *stream, uint32_t gpu_index, int8_t **pbs_buffer,
+    uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
+
+void scratch_cuda_programmable_bootstrap_amortized_64(
+    void *stream, uint32_t gpu_index, int8_t **pbs_buffer,
+    uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
+
+void cuda_programmable_bootstrap_amortized_lwe_ciphertext_vector_32(
+    void *stream, uint32_t gpu_index, void *lwe_array_out,
+    void *lwe_output_indexes, void *lut_vector, void *lut_vector_indexes,
+    void *lwe_array_in, void *lwe_input_indexes, void *bootstrapping_key,
+    int8_t *pbs_buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
+    uint32_t num_samples);
+
+void cuda_programmable_bootstrap_amortized_lwe_ciphertext_vector_64(
+    void *stream, uint32_t gpu_index, void *lwe_array_out,
+    void *lwe_output_indexes, void *lut_vector, void *lut_vector_indexes,
+    void *lwe_array_in, void *lwe_input_indexes, void *bootstrapping_key,
+    int8_t *pbs_buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
+    uint32_t num_samples);
+
+void cleanup_cuda_programmable_bootstrap_amortized(void *stream,
+                                                   uint32_t gpu_index,
+                                                   int8_t **pbs_buffer);
+
+void scratch_cuda_programmable_bootstrap_32(
+    void *stream, uint32_t gpu_index, int8_t **buffer, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t level_count,
+    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
+
+void scratch_cuda_programmable_bootstrap_64(
+    void *stream, uint32_t gpu_index, int8_t **buffer, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t level_count,
+    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
+
+void cuda_programmable_bootstrap_lwe_ciphertext_vector_32(
+    void *stream, uint32_t gpu_index, void *lwe_array_out,
+    void *lwe_output_indexes, void *lut_vector, void *lut_vector_indexes,
+    void *lwe_array_in, void *lwe_input_indexes, void *bootstrapping_key,
+    int8_t *buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
+    uint32_t num_samples, uint32_t lut_count, uint32_t lut_stride);
+
+void cuda_programmable_bootstrap_lwe_ciphertext_vector_64(
+    void *stream, uint32_t gpu_index, void *lwe_array_out,
+    void *lwe_output_indexes, void *lut_vector, void *lut_vector_indexes,
+    void *lwe_array_in, void *lwe_input_indexes, void *bootstrapping_key,
+    int8_t *buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
+    uint32_t num_samples, uint32_t lut_count, uint32_t lut_stride);
+
+void cleanup_cuda_programmable_bootstrap(void *stream, uint32_t gpu_index,
+                                         int8_t **pbs_buffer);
+}
 
 template <typename Torus>
 uint64_t get_buffer_size_full_sm_programmable_bootstrap_step_one(
@@ -69,7 +146,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
   int8_t *d_mem;
 
   Torus *global_accumulator;
-  double2 *global_join_buffer;
+  double2 *global_accumulator_fft;
 
   PBS_VARIANT pbs_variant;
 
@@ -114,7 +191,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
         // Otherwise, both kernels run all in shared memory
         d_mem = (int8_t *)cuda_malloc_async(device_mem, stream, gpu_index);
 
-        global_join_buffer = (double2 *)cuda_malloc_async(
+        global_accumulator_fft = (double2 *)cuda_malloc_async(
             (glwe_dimension + 1) * level_count * input_lwe_ciphertext_count *
                 (polynomial_size / 2) * sizeof(double2),
             stream, gpu_index);
@@ -147,7 +224,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
         // Otherwise, both kernels run all in shared memory
         d_mem = (int8_t *)cuda_malloc_async(device_mem, stream, gpu_index);
 
-        global_join_buffer = (double2 *)cuda_malloc_async(
+        global_accumulator_fft = (double2 *)cuda_malloc_async(
             (glwe_dimension + 1) * level_count * input_lwe_ciphertext_count *
                 polynomial_size / 2 * sizeof(double2),
             stream, gpu_index);
@@ -194,7 +271,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
         // Otherwise, both kernels run all in shared memory
         d_mem = (int8_t *)cuda_malloc_async(device_mem, stream, gpu_index);
 
-        global_join_buffer = (double2 *)cuda_malloc_async(
+        global_accumulator_fft = (double2 *)cuda_malloc_async(
             (glwe_dimension + 1) * level_count * input_lwe_ciphertext_count *
                 polynomial_size / 2 * sizeof(double2),
             stream, gpu_index);
@@ -208,7 +285,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
 
   void release(cudaStream_t stream, uint32_t gpu_index) {
     cuda_drop_async(d_mem, stream, gpu_index);
-    cuda_drop_async(global_join_buffer, stream, gpu_index);
+    cuda_drop_async(global_accumulator_fft, stream, gpu_index);
 
     if (pbs_variant == DEFAULT)
       cuda_drop_async(global_accumulator, stream, gpu_index);
@@ -250,35 +327,32 @@ bool has_support_to_cuda_programmable_bootstrap_cg(uint32_t glwe_dimension,
 template <typename Torus>
 void cuda_programmable_bootstrap_cg_lwe_ciphertext_vector(
     void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus const *lwe_output_indexes, Torus const *lut_vector,
-    Torus const *lut_vector_indexes, Torus const *lwe_array_in,
-    Torus const *lwe_input_indexes, double2 const *bootstrapping_key,
+    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
+    Torus *lwe_array_in, Torus *lwe_input_indexes, double2 *bootstrapping_key,
     pbs_buffer<Torus, CLASSICAL> *buffer, uint32_t lwe_dimension,
     uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t base_log,
-    uint32_t level_count, uint32_t num_samples, uint32_t num_many_lut,
+    uint32_t level_count, uint32_t num_samples, uint32_t lut_count,
     uint32_t lut_stride);
 
 template <typename Torus>
 void cuda_programmable_bootstrap_lwe_ciphertext_vector(
     void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus const *lwe_output_indexes, Torus const *lut_vector,
-    Torus const *lut_vector_indexes, Torus const *lwe_array_in,
-    Torus const *lwe_input_indexes, double2 const *bootstrapping_key,
+    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
+    Torus *lwe_array_in, Torus *lwe_input_indexes, double2 *bootstrapping_key,
     pbs_buffer<Torus, CLASSICAL> *buffer, uint32_t lwe_dimension,
     uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t base_log,
-    uint32_t level_count, uint32_t num_samples, uint32_t num_many_lut,
+    uint32_t level_count, uint32_t num_samples, uint32_t lut_count,
     uint32_t lut_stride);
 
 #if (CUDA_ARCH >= 900)
 template <typename Torus>
 void cuda_programmable_bootstrap_tbc_lwe_ciphertext_vector(
     void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus const *lwe_output_indexes, Torus const *lut_vector,
-    Torus const *lut_vector_indexes, Torus const *lwe_array_in,
-    Torus const *lwe_input_indexes, double2 const *bootstrapping_key,
+    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
+    Torus *lwe_array_in, Torus *lwe_input_indexes, double2 *bootstrapping_key,
     pbs_buffer<Torus, CLASSICAL> *buffer, uint32_t lwe_dimension,
     uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t base_log,
-    uint32_t level_count, uint32_t num_samples, uint32_t num_many_lut,
+    uint32_t level_count, uint32_t num_samples, uint32_t lut_count,
     uint32_t lut_stride);
 
 template <typename Torus>
@@ -334,4 +408,4 @@ __device__ const T *get_multi_bit_ith_lwe_gth_group_kth_block(
 
 #endif
 
-#endif // CUDA_BOOTSTRAP_UTILITIES_H
+#endif // CUDA_BOOTSTRAP_H

backends/tfhe-cuda-backend/cuda/include/programmable_bootstrap_multibit.h

@@ -1,7 +1,38 @@
-#ifndef CUDA_MULTI_BIT_UTILITIES_H
-#define CUDA_MULTI_BIT_UTILITIES_H
+#ifndef CUDA_MULTI_BIT_H
+#define CUDA_MULTI_BIT_H
 
-#include "pbs_utilities.h"
+#include "programmable_bootstrap.h"
+#include <cstdint>
+
+extern "C" {
+
+bool has_support_to_cuda_programmable_bootstrap_cg_multi_bit(
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
+    uint32_t num_samples);
+
+void cuda_convert_lwe_multi_bit_programmable_bootstrap_key_64(
+    void *stream, uint32_t gpu_index, void *dest, void *src,
+    uint32_t input_lwe_dim, uint32_t glwe_dim, uint32_t level_count,
+    uint32_t polynomial_size, uint32_t grouping_factor);
+
+void scratch_cuda_multi_bit_programmable_bootstrap_64(
+    void *stream, uint32_t gpu_index, int8_t **pbs_buffer,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
+    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
+
+void cuda_multi_bit_programmable_bootstrap_lwe_ciphertext_vector_64(
+    void *stream, uint32_t gpu_index, void *lwe_array_out,
+    void *lwe_output_indexes, void *lut_vector, void *lut_vector_indexes,
+    void *lwe_array_in, void *lwe_input_indexes, void *bootstrapping_key,
+    int8_t *buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t grouping_factor, uint32_t base_log,
+    uint32_t level_count, uint32_t num_samples, uint32_t lut_count,
+    uint32_t lut_stride);
+
+void cleanup_cuda_multi_bit_programmable_bootstrap(void *stream,
+                                                   uint32_t gpu_index,
+                                                   int8_t **pbs_buffer);
+}
 
 template <typename Torus>
 bool supports_distributed_shared_memory_on_multibit_programmable_bootstrap(
@@ -22,13 +53,12 @@ void scratch_cuda_tbc_multi_bit_programmable_bootstrap(
 template <typename Torus>
 void cuda_tbc_multi_bit_programmable_bootstrap_lwe_ciphertext_vector(
     void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus const *lwe_output_indexes, Torus const *lut_vector,
-    Torus const *lut_vector_indexes, Torus const *lwe_array_in,
-    Torus const *lwe_input_indexes, Torus const *bootstrapping_key,
+    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
+    Torus *lwe_array_in, Torus *lwe_input_indexes, Torus *bootstrapping_key,
     pbs_buffer<Torus, MULTI_BIT> *pbs_buffer, uint32_t lwe_dimension,
     uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t grouping_factor,
     uint32_t base_log, uint32_t level_count, uint32_t num_samples,
-    uint32_t num_many_lut, uint32_t lut_stride);
+    uint32_t lut_count, uint32_t lut_stride);
 #endif
 
 template <typename Torus>
@@ -40,13 +70,12 @@ void scratch_cuda_cg_multi_bit_programmable_bootstrap(
 template <typename Torus>
 void cuda_cg_multi_bit_programmable_bootstrap_lwe_ciphertext_vector(
     void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus const *lwe_output_indexes, Torus const *lut_vector,
-    Torus const *lut_vector_indexes, Torus const *lwe_array_in,
-    Torus const *lwe_input_indexes, Torus const *bootstrapping_key,
+    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
+    Torus *lwe_array_in, Torus *lwe_input_indexes, Torus *bootstrapping_key,
     pbs_buffer<Torus, MULTI_BIT> *pbs_buffer, uint32_t lwe_dimension,
     uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t grouping_factor,
     uint32_t base_log, uint32_t level_count, uint32_t num_samples,
-    uint32_t num_many_lut, uint32_t lut_stride);
+    uint32_t lut_count, uint32_t lut_stride);
 
 template <typename Torus>
 void scratch_cuda_multi_bit_programmable_bootstrap(
@@ -57,13 +86,12 @@ void scratch_cuda_multi_bit_programmable_bootstrap(
 template <typename Torus>
 void cuda_multi_bit_programmable_bootstrap_lwe_ciphertext_vector(
     void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus const *lwe_output_indexes, Torus const *lut_vector,
-    Torus const *lut_vector_indexes, Torus const *lwe_array_in,
-    Torus const *lwe_input_indexes, Torus const *bootstrapping_key,
+    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
+    Torus *lwe_array_in, Torus *lwe_input_indexes, Torus *bootstrapping_key,
     pbs_buffer<Torus, MULTI_BIT> *pbs_buffer, uint32_t lwe_dimension,
     uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t grouping_factor,
     uint32_t base_log, uint32_t level_count, uint32_t num_samples,
-    uint32_t num_many_lut, uint32_t lut_stride);
+    uint32_t lut_count, uint32_t lut_stride);
 
 template <typename Torus>
 uint64_t get_buffer_size_full_sm_multibit_programmable_bootstrap_keybundle(
@@ -93,10 +121,6 @@ template <typename Torus>
 uint64_t get_buffer_size_full_sm_tbc_multibit_programmable_bootstrap(
     uint32_t polynomial_size);
 
-template <typename Torus, class params>
-uint32_t get_lwe_chunk_size(uint32_t gpu_index, uint32_t max_num_pbs,
-                            uint32_t polynomial_size);
-
 template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::MULTI_BIT> {
   int8_t *d_mem_keybundle = NULL;
   int8_t *d_mem_acc_step_one = NULL;
@@ -106,7 +130,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::MULTI_BIT> {
   uint32_t lwe_chunk_size;
   double2 *keybundle_fft;
   Torus *global_accumulator;
-  double2 *global_join_buffer;
+  double2 *global_accumulator_fft;
 
   PBS_VARIANT pbs_variant;
 
@@ -225,12 +249,10 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::MULTI_BIT> {
           num_blocks_keybundle * (polynomial_size / 2) * sizeof(double2),
           stream, gpu_index);
       global_accumulator = (Torus *)cuda_malloc_async(
-          input_lwe_ciphertext_count * (glwe_dimension + 1) * polynomial_size *
-              sizeof(Torus),
-          stream, gpu_index);
-      global_join_buffer = (double2 *)cuda_malloc_async(
-          level_count * (glwe_dimension + 1) * input_lwe_ciphertext_count *
-              (polynomial_size / 2) * sizeof(double2),
+          num_blocks_acc_step_one * polynomial_size * sizeof(Torus), stream,
+          gpu_index);
+      global_accumulator_fft = (double2 *)cuda_malloc_async(
+          num_blocks_acc_step_one * (polynomial_size / 2) * sizeof(double2),
           stream, gpu_index);
     }
   }
@@ -262,8 +284,12 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::MULTI_BIT> {
 
     cuda_drop_async(keybundle_fft, stream, gpu_index);
     cuda_drop_async(global_accumulator, stream, gpu_index);
-    cuda_drop_async(global_join_buffer, stream, gpu_index);
+    cuda_drop_async(global_accumulator_fft, stream, gpu_index);
   }
 };
 
-#endif // CUDA_MULTI_BIT_UTILITIES_H
+template <typename Torus, class params>
+uint32_t get_lwe_chunk_size(uint32_t gpu_index, uint32_t max_num_pbs,
+                            uint32_t polynomial_size);
+
+#endif // CUDA_MULTI_BIT_H

backends/tfhe-cuda-backend/cuda/src/crypto/ciphertext.cu

@@ -22,8 +22,8 @@ void cuda_convert_lwe_ciphertext_vector_to_cpu_64(void *stream,
 }
 
 void cuda_glwe_sample_extract_64(void *stream, uint32_t gpu_index,
-                                 void *lwe_array_out, void const *glwe_array_in,
-                                 uint32_t const *nth_array, uint32_t num_nths,
+                                 void *lwe_array_out, void *glwe_array_in,
+                                 uint32_t *nth_array, uint32_t num_nths,
                                  uint32_t glwe_dimension,
                                  uint32_t polynomial_size) {
 
@@ -31,43 +31,43 @@ void cuda_glwe_sample_extract_64(void *stream, uint32_t gpu_index,
   case 256:
     host_sample_extract<uint64_t, AmortizedDegree<256>>(
         static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t const *)glwe_array_in, (uint32_t const *)nth_array, num_nths,
+        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_nths,
         glwe_dimension);
     break;
   case 512:
     host_sample_extract<uint64_t, AmortizedDegree<512>>(
         static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t const *)glwe_array_in, (uint32_t const *)nth_array, num_nths,
+        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_nths,
         glwe_dimension);
     break;
   case 1024:
     host_sample_extract<uint64_t, AmortizedDegree<1024>>(
         static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t const *)glwe_array_in, (uint32_t const *)nth_array, num_nths,
+        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_nths,
         glwe_dimension);
     break;
   case 2048:
     host_sample_extract<uint64_t, AmortizedDegree<2048>>(
         static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t const *)glwe_array_in, (uint32_t const *)nth_array, num_nths,
+        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_nths,
         glwe_dimension);
     break;
   case 4096:
     host_sample_extract<uint64_t, AmortizedDegree<4096>>(
         static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t const *)glwe_array_in, (uint32_t const *)nth_array, num_nths,
+        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_nths,
         glwe_dimension);
     break;
   case 8192:
     host_sample_extract<uint64_t, AmortizedDegree<8192>>(
         static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t const *)glwe_array_in, (uint32_t const *)nth_array, num_nths,
+        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_nths,
         glwe_dimension);
     break;
   case 16384:
     host_sample_extract<uint64_t, AmortizedDegree<16384>>(
         static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t const *)glwe_array_in, (uint32_t const *)nth_array, num_nths,
+        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_nths,
         glwe_dimension);
     break;
   default:

backends/tfhe-cuda-backend/cuda/src/crypto/ciphertext.cuh

@@ -27,9 +27,8 @@ void cuda_convert_lwe_ciphertext_vector_to_cpu(cudaStream_t stream,
 }
 
 template <typename Torus, class params>
-__global__ void sample_extract(Torus *lwe_array_out, Torus const *glwe_array_in,
-                               uint32_t const *nth_array,
-                               uint32_t glwe_dimension) {
+__global__ void sample_extract(Torus *lwe_array_out, Torus *glwe_array_in,
+                               uint32_t *nth_array, uint32_t glwe_dimension) {
 
   const int input_id = blockIdx.x;
 
@@ -51,9 +50,8 @@ __global__ void sample_extract(Torus *lwe_array_out, Torus const *glwe_array_in,
 
 template <typename Torus, class params>
 __host__ void host_sample_extract(cudaStream_t stream, uint32_t gpu_index,
-                                  Torus *lwe_array_out,
-                                  Torus const *glwe_array_in,
-                                  uint32_t const *nth_array, uint32_t num_nths,
+                                  Torus *lwe_array_out, Torus *glwe_array_in,
+                                  uint32_t *nth_array, uint32_t num_nths,
                                   uint32_t glwe_dimension) {
   cudaSetDevice(gpu_index);
 

backends/tfhe-cuda-backend/cuda/src/crypto/fast_packing_keyswitch.cuh

@@ -1,358 +0,0 @@
-#ifndef CNCRT_FAST_KS_CUH
-#define CNCRT_FAST_KS_CUH
-
-#undef NDEBUG
-#include <assert.h>
-
-#include "device.h"
-#include "gadget.cuh"
-#include "helper_multi_gpu.h"
-#include "keyswitch.cuh"
-#include "polynomial/functions.cuh"
-#include "polynomial/polynomial_math.cuh"
-#include "torus.cuh"
-#include "utils/helper.cuh"
-#include "utils/kernel_dimensions.cuh"
-#include <thread>
-#include <vector>
-
-#define CEIL_DIV(M, N) ((M) + (N)-1) / (N)
-
-const int BLOCK_SIZE_GEMM = 64;
-const int THREADS_GEMM = 8;
-const int BLOCK_SIZE_DECOMP = 8;
-
-template <typename Torus> uint64_t get_shared_mem_size_tgemm() {
-  return BLOCK_SIZE_GEMM * THREADS_GEMM * 2 * sizeof(Torus);
-}
-
-__host__ inline bool can_use_pks_fast_path(uint32_t lwe_dimension,
-                                           uint32_t num_lwe,
-                                           uint32_t polynomial_size,
-                                           uint32_t level_count,
-                                           uint32_t glwe_dimension) {
-  // TODO: activate it back, fix tests and extend to level_count > 1
-  return false;
-}
-
-// Initialize decomposition by performing rounding
-// and decomposing one level of an array of Torus LWEs. Only
-// decomposes the mask elements of the incoming LWEs.
-template <typename Torus, typename TorusVec>
-__global__ void decompose_vectorize_init(Torus const *lwe_in, Torus *lwe_out,
-                                         uint32_t lwe_dimension,
-                                         uint32_t num_lwe, uint32_t base_log,
-                                         uint32_t level_count) {
-
-  // index of this LWE ct in the buffer
-  auto lwe_idx = blockIdx.x * blockDim.x + threadIdx.x;
-  // index of the LWE sample in the LWE ct
-  auto lwe_sample_idx = blockIdx.y * blockDim.y + threadIdx.y;
-
-  if (lwe_idx >= num_lwe || lwe_sample_idx >= lwe_dimension)
-    return;
-
-  // Input LWE array is [mask_0, .., mask_lwe_dim, message] and
-  // we only decompose the mask. Thus the stride for reading
-  // is lwe_dimension + 1, while for writing it is lwe_dimension
-  auto read_val_idx = lwe_idx * (lwe_dimension + 1) + lwe_sample_idx;
-  auto write_val_idx = lwe_idx * lwe_dimension + lwe_sample_idx;
-
-  Torus a_i = lwe_in[read_val_idx];
-
-  Torus state = init_decomposer_state(a_i, base_log, level_count);
-
-  Torus mod_b_mask = (1ll << base_log) - 1ll;
-  lwe_out[write_val_idx] = decompose_one<Torus>(state, mod_b_mask, base_log);
-}
-
-// Continue decomposiion of an array of Torus elements in place. Supposes
-// that the array contains already decomposed elements and
-// computes the new decomposed level in place.
-template <typename Torus, typename TorusVec>
-__global__ void
-decompose_vectorize_step_inplace(Torus *buffer_in, uint32_t lwe_dimension,
-                                 uint32_t num_lwe, uint32_t base_log,
-                                 uint32_t level_count) {
-
-  // index of this LWE ct in the buffer
-  auto lwe_idx = blockIdx.x * blockDim.x + threadIdx.x;
-  // index of the LWE sample in the LWE ct
-  auto lwe_sample_idx = blockIdx.y * blockDim.y + threadIdx.y;
-
-  if (lwe_idx >= num_lwe || lwe_sample_idx >= lwe_dimension)
-    return;
-
-  auto val_idx = lwe_idx * lwe_dimension + lwe_sample_idx;
-
-  Torus state = buffer_in[val_idx];
-
-  Torus mod_b_mask = (1ll << base_log) - 1ll;
-
-  buffer_in[val_idx] = decompose_one<Torus>(state, mod_b_mask, base_log);
-}
-
-// Multiply matrices A, B of size (M, K), (K, N) respectively
-// with K as the inner dimension.
-//
-// A block of threads processeds blocks of size (BLOCK_SIZE_GEMM,
-// BLOCK_SIZE_GEMM) splitting them in multiple tiles: (BLOCK_SIZE_GEMM,
-// THREADS_GEMM)-shaped tiles of values from A, and a (THREADS_GEMM,
-// BLOCK_SIZE_GEMM)-shaped tiles of values from B.
-template <typename Torus, typename TorusVec>
-__global__ void tgemm(int M, int N, int K, const Torus *A, const Torus *B,
-                      int stride_B, Torus *C) {
-
-  const int BM = BLOCK_SIZE_GEMM;
-  const int BN = BLOCK_SIZE_GEMM;
-  const int BK = THREADS_GEMM;
-  const int TM = THREADS_GEMM;
-
-  const uint cRow = blockIdx.y;
-  const uint cCol = blockIdx.x;
-
-  const uint totalResultsBlocktile = BM * BN;
-  const int threadCol = threadIdx.x % BN;
-  const int threadRow = threadIdx.x / BN;
-
-  // Allocate space for the current block tile in shared memory
-  __shared__ Torus As[BM * BK];
-  __shared__ Torus Bs[BK * BN];
-
-  // Initialize the pointers to the input blocks from A, B
-  // Tiles from these blocks are loaded to shared memory
-  A += cRow * BM * K;
-  B += cCol * BN;
-
-  // Each thread will handle multiple sub-blocks
-  const uint innerColA = threadIdx.x % BK;
-  const uint innerRowA = threadIdx.x / BK;
-  const uint innerColB = threadIdx.x % BN;
-  const uint innerRowB = threadIdx.x / BN;
-
-  // allocate thread-local cache for results in registerfile
-  Torus threadResults[TM] = {0};
-
-  auto row_A = cRow * BM + innerRowA;
-  auto col_B = cCol * BN + innerColB;
-
-  // For each thread, loop over block tiles
-  for (uint bkIdx = 0; bkIdx < K; bkIdx += BK) {
-    auto col_A = bkIdx + innerColA;
-    auto row_B = bkIdx + innerRowB;
-
-    if (row_A < M && col_A < K) {
-      As[innerRowA * BK + innerColA] = A[innerRowA * K + innerColA];
-    } else {
-      As[innerRowA * BK + innerColA] = 0;
-    }
-
-    if (col_B < N && row_B < K) {
-      Bs[innerRowB * BN + innerColB] = B[innerRowB * stride_B + innerColB];
-    } else {
-      Bs[innerRowB * BN + innerColB] = 0;
-    }
-    __syncthreads();
-
-    // Advance blocktile for the next iteration of this loop
-    A += BK;
-    B += BK * stride_B;
-
-    // calculate per-thread results
-    for (uint dotIdx = 0; dotIdx < BK; ++dotIdx) {
-      // we make the dotproduct loop the outside loop, which facilitates
-      // reuse of the Bs entry, which we can cache in a tmp var.
-      Torus tmp = Bs[dotIdx * BN + threadCol];
-      for (uint resIdx = 0; resIdx < TM; ++resIdx) {
-        threadResults[resIdx] +=
-            As[(threadRow * TM + resIdx) * BK + dotIdx] * tmp;
-      }
-    }
-    __syncthreads();
-  }
-
-  // Initialize the pointer to the output block of size (BLOCK_SIZE_GEMM,
-  // BLOCK_SIZE_GEMM)
-  C += cRow * BM * N + cCol * BN;
-
-  // write out the results
-  for (uint resIdx = 0; resIdx < TM; ++resIdx) {
-    int outRow = cRow * BM + threadRow * TM + resIdx;
-    int outCol = cCol * BN + threadCol;
-
-    if (outRow >= M)
-      continue;
-    if (outCol >= N)
-      continue;
-
-    C[(threadRow * TM + resIdx) * N + threadCol] += threadResults[resIdx];
-  }
-}
-
-// Finish the keyswitching operation and prepare GLWEs for accumulation.
-// 1. Finish the keyswitching computation partially performed with a GEMM:
-//  - negate the dot product between the GLWE and KSK polynomial
-//  - add the GLWE message for the N-th polynomial coeff in the message poly
-// 2. Rotate each of the GLWE . KSK poly dot products to
-//    prepare them for accumulation into a single GLWE
-template <typename Torus>
-__global__ void polynomial_accumulate_monic_monomial_mul_many_neg_and_add_C(
-    Torus *in_glwe_buffer, Torus *out_glwe_buffer, Torus const *lwe_array,
-    uint32_t lwe_dimension, uint32_t num_glwes, uint32_t polynomial_size,
-    uint32_t glwe_dimension) {
-
-  uint32_t glwe_id = blockIdx.x * blockDim.x + threadIdx.x;
-  uint32_t degree = glwe_id; // lwe 0 rotate 0, lwe 1 rotate 1, .. , lwe
-                             // poly_size-1 rotate poly_size-1
-  uint32_t coeffIdx = blockIdx.y * blockDim.y + threadIdx.y;
-
-  if (glwe_id >= num_glwes)
-    return;
-  if (coeffIdx >= polynomial_size)
-    return;
-
-  auto in_poly =
-      in_glwe_buffer + glwe_id * polynomial_size * (glwe_dimension + 1);
-  auto out_result =
-      out_glwe_buffer + glwe_id * polynomial_size * (glwe_dimension + 1);
-  if (coeffIdx == 0) {
-    // Add the message value of the input LWE (`C`) to the N-th coefficient
-    // in the GLWE . KSK dot product
-
-    // The C is added to the first position of the last polynomial in the GLWE
-    // which has (glwe_dimension+1) polynomials
-    // The C value is extracted as the last value of the LWE ct. (of index
-    // glwe_id) the LWEs have (polynomial_size + 1) values
-    in_poly[polynomial_size * glwe_dimension] =
-        lwe_array[glwe_id * (lwe_dimension + 1) + lwe_dimension] -
-        in_poly[polynomial_size * glwe_dimension];
-
-    for (int gi = 1; gi < glwe_dimension; ++gi)
-      in_poly[coeffIdx + gi * polynomial_size] =
-          -in_poly[coeffIdx + gi * polynomial_size];
-
-  } else {
-    // Otherwise simply negate the input coefficient
-    for (int gi = 1; gi < glwe_dimension + 1; ++gi)
-      in_poly[coeffIdx + gi * polynomial_size] =
-          -in_poly[coeffIdx + gi * polynomial_size];
-  }
-  // Negate all the coefficients for rotation for the first poly
-  in_poly[coeffIdx] = -in_poly[coeffIdx];
-
-  // rotate the body
-  polynomial_accumulate_monic_monomial_mul<Torus>(
-      out_result, in_poly, degree, coeffIdx, polynomial_size, 1, true);
-  // rotate the mask too
-  for (int gi = 1; gi < glwe_dimension + 1; ++gi)
-    polynomial_accumulate_monic_monomial_mul<Torus>(
-        out_result + gi * polynomial_size, in_poly + gi * polynomial_size,
-        degree, coeffIdx, polynomial_size, 1, true);
-}
-
-template <typename Torus, typename TorusVec>
-__host__ void host_fast_packing_keyswitch_lwe_list_to_glwe(
-    cudaStream_t stream, uint32_t gpu_index, Torus *glwe_out,
-    Torus const *lwe_array_in, Torus const *fp_ksk_array, int8_t *fp_ks_buffer,
-    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t base_log, uint32_t level_count, uint32_t num_lwes) {
-
-  // Optimization of packing keyswitch when packing many LWEs
-
-  if (level_count > 1) {
-    PANIC("Fast path PKS only supports level_count==1");
-  }
-
-  cudaSetDevice(gpu_index);
-  check_cuda_error(cudaGetLastError());
-
-  int glwe_accumulator_size = (glwe_dimension + 1) * polynomial_size;
-
-  // The fast path of PKS uses the scratch buffer (d_mem) differently than the
-  // old path: it needs to store the decomposed masks in the first half of this
-  // buffer and the keyswitched GLWEs in the second half of the buffer. Thus the
-  // scratch buffer for the fast path must determine the half-size of the
-  // scratch buffer as the max between the size of the GLWE and the size of the
-  // LWE-mask
-  int memory_unit = glwe_accumulator_size > lwe_dimension
-                        ? glwe_accumulator_size
-                        : lwe_dimension;
-
-  // ping pong the buffer between successive calls
-  // split the buffer in two parts of this size
-  auto d_mem_0 = (Torus *)fp_ks_buffer;
-  auto d_mem_1 = d_mem_0 + num_lwes * memory_unit;
-
-  // Set the scratch buffer to 0 as it is used to accumulate
-  // decomposition temporary results
-  cuda_memset_async(d_mem_1, 0, num_lwes * memory_unit * sizeof(Torus), stream,
-                    gpu_index);
-  check_cuda_error(cudaGetLastError());
-
-  // decompose LWEs
-  // don't decompose LWE body - the LWE has lwe_size + 1 elements. The last
-  // element, the body is ignored by rounding down the number of blocks assuming
-  // here that the LWE dimension is a multiple of the block size
-  dim3 grid_decomp(CEIL_DIV(num_lwes, BLOCK_SIZE_DECOMP),
-                   CEIL_DIV(lwe_dimension, BLOCK_SIZE_DECOMP));
-  dim3 threads_decomp(BLOCK_SIZE_DECOMP, BLOCK_SIZE_DECOMP);
-
-  // decompose first level
-  decompose_vectorize_init<Torus, TorusVec>
-      <<<grid_decomp, threads_decomp, 0, stream>>>(lwe_array_in, d_mem_0,
-                                                   lwe_dimension, num_lwes,
-                                                   base_log, level_count);
-  check_cuda_error(cudaGetLastError());
-
-  // gemm to ks the individual LWEs to GLWEs
-  dim3 grid_gemm(CEIL_DIV(glwe_accumulator_size, BLOCK_SIZE_GEMM),
-                 CEIL_DIV(num_lwes, BLOCK_SIZE_GEMM));
-  dim3 threads_gemm(BLOCK_SIZE_GEMM * THREADS_GEMM);
-
-  auto stride_KSK_buffer = glwe_accumulator_size;
-
-  uint32_t shared_mem_size = get_shared_mem_size_tgemm<Torus>();
-  tgemm<Torus, TorusVec><<<grid_gemm, threads_gemm, shared_mem_size, stream>>>(
-      num_lwes, glwe_accumulator_size, lwe_dimension, d_mem_0, fp_ksk_array,
-      stride_KSK_buffer, d_mem_1);
-  check_cuda_error(cudaGetLastError());
-
-  /*
-    TODO: transpose key to generalize to level_count > 1
-
-    for (int li = 1; li < level_count; ++li) {
-      decompose_vectorize_step_inplace<Torus, TorusVec>
-          <<<grid_decomp, threads_decomp, 0, stream>>>(
-              d_mem_0, lwe_dimension, num_lwes, base_log, level_count);
-      check_cuda_error(cudaGetLastError());
-
-      tgemm<Torus, TorusVec><<<grid_gemm, threads_gemm, shared_mem_size,
-    stream>>>( num_lwes, glwe_accumulator_size, lwe_dimension, d_mem_0,
-          fp_ksk_array + li * ksk_block_size, stride_KSK_buffer, d_mem_1);
-      check_cuda_error(cudaGetLastError());
-    }
-  */
-
-  // should we include the mask in the rotation ??
-  dim3 grid_rotate(CEIL_DIV(num_lwes, BLOCK_SIZE_DECOMP),
-                   CEIL_DIV(polynomial_size, BLOCK_SIZE_DECOMP));
-  dim3 threads_rotate(BLOCK_SIZE_DECOMP, BLOCK_SIZE_DECOMP);
-  // rotate the GLWEs
-  polynomial_accumulate_monic_monomial_mul_many_neg_and_add_C<Torus>
-      <<<grid_rotate, threads_rotate, 0, stream>>>(
-          d_mem_1, d_mem_0, lwe_array_in, lwe_dimension, num_lwes,
-          polynomial_size, glwe_dimension);
-  check_cuda_error(cudaGetLastError());
-
-  dim3 grid_accumulate(
-      CEIL_DIV(polynomial_size * (glwe_dimension + 1), BLOCK_SIZE_DECOMP));
-  dim3 threads_accum(BLOCK_SIZE_DECOMP);
-
-  // accumulate to a single glwe
-  accumulate_glwes<Torus><<<grid_accumulate, threads_accum, 0, stream>>>(
-      glwe_out, d_mem_0, glwe_dimension, polynomial_size, num_lwes);
-
-  check_cuda_error(cudaGetLastError());
-}
-
-#endif

backends/tfhe-cuda-backend/cuda/src/crypto/gadget.cuh

@@ -1,7 +1,6 @@
 #ifndef CNCRT_CRYPTO_CUH
 #define CNCRT_CRPYTO_CUH
 
-#include "crypto/torus.cuh"
 #include "device.h"
 #include <cstdint>
 
@@ -22,6 +21,7 @@ private:
   uint32_t base_log;
   uint32_t mask;
   uint32_t num_poly;
+  int current_level;
   T mask_mod_b;
   T *state;
 
@@ -32,6 +32,13 @@ public:
         state(state) {
 
     mask_mod_b = (1ll << base_log) - 1ll;
+    current_level = level_count;
+    int tid = threadIdx.x;
+    for (int i = 0; i < num_poly * params::opt; i++) {
+      state[tid] >>= (sizeof(T) * 8 - base_log * level_count);
+      tid += params::degree / params::opt;
+    }
+    synchronize_threads_in_block();
   }
 
   // Decomposes all polynomials at once
@@ -45,30 +52,28 @@ public:
   // Decomposes a single polynomial
   __device__ void decompose_and_compress_next_polynomial(double2 *result,
                                                          int j) {
-    uint32_t tid = threadIdx.x;
-    auto state_slice = &state[j * params::degree];
+    if (j == 0)
+      current_level -= 1;
+
+    int tid = threadIdx.x;
+    auto state_slice = state + j * params::degree;
     for (int i = 0; i < params::opt / 2; i++) {
-      auto input1 = &state_slice[tid];
-      auto input2 = &state_slice[tid + params::degree / 2];
-      T res_re = *input1 & mask_mod_b;
-      T res_im = *input2 & mask_mod_b;
-
-      *input1 >>= base_log; // Update state
-      *input2 >>= base_log; // Update state
-
-      T carry_re = ((res_re - 1ll) | *input1) & res_re;
-      T carry_im = ((res_im - 1ll) | *input2) & res_im;
+      T res_re = state_slice[tid] & mask_mod_b;
+      T res_im = state_slice[tid + params::degree / 2] & mask_mod_b;
+      state_slice[tid] >>= base_log;
+      state_slice[tid + params::degree / 2] >>= base_log;
+      T carry_re = ((res_re - 1ll) | state_slice[tid]) & res_re;
+      T carry_im =
+          ((res_im - 1ll) | state_slice[tid + params::degree / 2]) & res_im;
       carry_re >>= (base_log - 1);
       carry_im >>= (base_log - 1);
-
-      *input1 += carry_re; // Update state
-      *input2 += carry_im; // Update state
-
+      state_slice[tid] += carry_re;
+      state_slice[tid + params::degree / 2] += carry_im;
       res_re -= carry_re << base_log;
       res_im -= carry_im << base_log;
 
-      typecast_torus_to_double(res_re, result[tid].x);
-      typecast_torus_to_double(res_im, result[tid].y);
+      result[tid].x = (int32_t)res_re;
+      result[tid].y = (int32_t)res_im;
 
       tid += params::degree / params::opt;
     }

backends/tfhe-cuda-backend/cuda/src/crypto/keyswitch.cu

@@ -1,8 +1,6 @@
-#include "fast_packing_keyswitch.cuh"
 #include "keyswitch.cuh"
 #include "keyswitch.h"
 #include <cstdint>
-#include <stdio.h>
 
 /* Perform keyswitch on a batch of 32 bits input LWE ciphertexts.
  * Head out to the equivalent operation on 64 bits for more details.
@@ -39,58 +37,42 @@ void cuda_keyswitch_lwe_ciphertext_vector_32(
  */
 void cuda_keyswitch_lwe_ciphertext_vector_64(
     void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void const *lwe_output_indexes, void const *lwe_array_in,
-    void const *lwe_input_indexes, void const *ksk, uint32_t lwe_dimension_in,
-    uint32_t lwe_dimension_out, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples) {
+    void *lwe_output_indexes, void *lwe_array_in, void *lwe_input_indexes,
+    void *ksk, uint32_t lwe_dimension_in, uint32_t lwe_dimension_out,
+    uint32_t base_log, uint32_t level_count, uint32_t num_samples) {
   host_keyswitch_lwe_ciphertext_vector<uint64_t>(
       static_cast<cudaStream_t>(stream), gpu_index,
       static_cast<uint64_t *>(lwe_array_out),
-      static_cast<const uint64_t *>(lwe_output_indexes),
-      static_cast<const uint64_t *>(lwe_array_in),
-      static_cast<const uint64_t *>(lwe_input_indexes),
-      static_cast<const uint64_t *>(ksk), lwe_dimension_in, lwe_dimension_out,
-      base_log, level_count, num_samples);
+      static_cast<uint64_t *>(lwe_output_indexes),
+      static_cast<uint64_t *>(lwe_array_in),
+      static_cast<uint64_t *>(lwe_input_indexes), static_cast<uint64_t *>(ksk),
+      lwe_dimension_in, lwe_dimension_out, base_log, level_count, num_samples);
 }
 
 void scratch_packing_keyswitch_lwe_list_to_glwe_64(
     void *stream, uint32_t gpu_index, int8_t **fp_ks_buffer,
-    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t num_lwes, bool allocate_gpu_memory) {
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t num_lwes,
+    bool allocate_gpu_memory) {
   scratch_packing_keyswitch_lwe_list_to_glwe<uint64_t>(
-      static_cast<cudaStream_t>(stream), gpu_index, fp_ks_buffer, lwe_dimension,
+      static_cast<cudaStream_t>(stream), gpu_index, fp_ks_buffer,
       glwe_dimension, polynomial_size, num_lwes, allocate_gpu_memory);
 }
-
 /* Perform functional packing keyswitch on a batch of 64 bits input LWE
  * ciphertexts.
  */
-
 void cuda_packing_keyswitch_lwe_list_to_glwe_64(
-    void *stream, uint32_t gpu_index, void *glwe_array_out,
-    void const *lwe_array_in, void const *fp_ksk_array, int8_t *fp_ks_buffer,
-    uint32_t input_lwe_dimension, uint32_t output_glwe_dimension,
-    uint32_t output_polynomial_size, uint32_t base_log, uint32_t level_count,
-    uint32_t num_lwes) {
+    void *stream, uint32_t gpu_index, void *glwe_array_out, void *lwe_array_in,
+    void *fp_ksk_array, int8_t *fp_ks_buffer, uint32_t input_lwe_dimension,
+    uint32_t output_glwe_dimension, uint32_t output_polynomial_size,
+    uint32_t base_log, uint32_t level_count, uint32_t num_lwes) {
 
-  if (can_use_pks_fast_path(input_lwe_dimension, num_lwes,
-                            output_polynomial_size, level_count,
-                            output_glwe_dimension)) {
-    host_fast_packing_keyswitch_lwe_list_to_glwe<uint64_t, ulonglong4>(
-        static_cast<cudaStream_t>(stream), gpu_index,
-        static_cast<uint64_t *>(glwe_array_out),
-        static_cast<const uint64_t *>(lwe_array_in),
-        static_cast<const uint64_t *>(fp_ksk_array), fp_ks_buffer,
-        input_lwe_dimension, output_glwe_dimension, output_polynomial_size,
-        base_log, level_count, num_lwes);
-  } else
-    host_packing_keyswitch_lwe_list_to_glwe<uint64_t>(
-        static_cast<cudaStream_t>(stream), gpu_index,
-        static_cast<uint64_t *>(glwe_array_out),
-        static_cast<const uint64_t *>(lwe_array_in),
-        static_cast<const uint64_t *>(fp_ksk_array), fp_ks_buffer,
-        input_lwe_dimension, output_glwe_dimension, output_polynomial_size,
-        base_log, level_count, num_lwes);
+  host_packing_keyswitch_lwe_list_to_glwe<uint64_t>(
+      static_cast<cudaStream_t>(stream), gpu_index,
+      static_cast<uint64_t *>(glwe_array_out),
+      static_cast<uint64_t *>(lwe_array_in),
+      static_cast<uint64_t *>(fp_ksk_array), fp_ks_buffer, input_lwe_dimension,
+      output_glwe_dimension, output_polynomial_size, base_log, level_count,
+      num_lwes);
 }
 
 void cleanup_packing_keyswitch_lwe_list_to_glwe(void *stream,

backends/tfhe-cuda-backend/cuda/src/crypto/keyswitch.cuh

@@ -71,8 +71,9 @@ keyswitch(Torus *lwe_array_out, const Torus *__restrict__ lwe_output_indexes,
 
     // This loop distribution seems to benefit the global mem reads
     for (int i = start_i; i < end_i; i++) {
-      Torus state =
-          init_decomposer_state(block_lwe_array_in[i], base_log, level_count);
+      Torus a_i = round_to_closest_multiple(block_lwe_array_in[i], base_log,
+                                            level_count);
+      Torus state = a_i >> (sizeof(Torus) * 8 - base_log * level_count);
 
       for (int j = 0; j < level_count; j++) {
         auto ksk_block =
@@ -100,10 +101,9 @@ keyswitch(Torus *lwe_array_out, const Torus *__restrict__ lwe_output_indexes,
 template <typename Torus>
 __host__ void host_keyswitch_lwe_ciphertext_vector(
     cudaStream_t stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus const *lwe_output_indexes, Torus const *lwe_array_in,
-    Torus const *lwe_input_indexes, Torus const *ksk, uint32_t lwe_dimension_in,
-    uint32_t lwe_dimension_out, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples) {
+    Torus *lwe_output_indexes, Torus *lwe_array_in, Torus *lwe_input_indexes,
+    Torus *ksk, uint32_t lwe_dimension_in, uint32_t lwe_dimension_out,
+    uint32_t base_log, uint32_t level_count, uint32_t num_samples) {
 
   cudaSetDevice(gpu_index);
 
@@ -124,13 +124,13 @@ __host__ void host_keyswitch_lwe_ciphertext_vector(
 }
 
 template <typename Torus>
-void execute_keyswitch_async(cudaStream_t const *streams,
-                             uint32_t const *gpu_indexes, uint32_t gpu_count,
+void execute_keyswitch_async(cudaStream_t *streams, uint32_t *gpu_indexes,
+                             uint32_t gpu_count,
                              const LweArrayVariant<Torus> &lwe_array_out,
                              const LweArrayVariant<Torus> &lwe_output_indexes,
                              const LweArrayVariant<Torus> &lwe_array_in,
                              const LweArrayVariant<Torus> &lwe_input_indexes,
-                             Torus *const *ksks, uint32_t lwe_dimension_in,
+                             Torus **ksks, uint32_t lwe_dimension_in,
                              uint32_t lwe_dimension_out, uint32_t base_log,
                              uint32_t level_count, uint32_t num_samples) {
 
@@ -158,20 +158,16 @@ void execute_keyswitch_async(cudaStream_t const *streams,
 template <typename Torus>
 __host__ void scratch_packing_keyswitch_lwe_list_to_glwe(
     cudaStream_t stream, uint32_t gpu_index, int8_t **fp_ks_buffer,
-    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t num_lwes, bool allocate_gpu_memory) {
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t num_lwes,
+    bool allocate_gpu_memory) {
   cudaSetDevice(gpu_index);
 
   int glwe_accumulator_size = (glwe_dimension + 1) * polynomial_size;
 
-  int memory_unit = glwe_accumulator_size > lwe_dimension
-                        ? glwe_accumulator_size
-                        : lwe_dimension;
-
-  if (allocate_gpu_memory) {
+  if (allocate_gpu_memory)
     *fp_ks_buffer = (int8_t *)cuda_malloc_async(
-        2 * num_lwes * memory_unit * sizeof(Torus), stream, gpu_index);
-  }
+        2 * num_lwes * glwe_accumulator_size * sizeof(Torus), stream,
+        gpu_index);
 }
 
 // public functional packing keyswitch for a single LWE ciphertext
@@ -180,9 +176,9 @@ __host__ void scratch_packing_keyswitch_lwe_list_to_glwe(
 // different thread blocks at the x-axis to work on that input.
 template <typename Torus>
 __device__ void packing_keyswitch_lwe_ciphertext_into_glwe_ciphertext(
-    Torus *glwe_out, Torus const *lwe_in, Torus const *fp_ksk,
-    uint32_t lwe_dimension_in, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count) {
+    Torus *glwe_out, Torus *lwe_in, Torus *fp_ksk, uint32_t lwe_dimension_in,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t base_log,
+    uint32_t level_count) {
 
   const int tid = threadIdx.x + blockIdx.x * blockDim.x;
   size_t glwe_size = (glwe_dimension + 1);
@@ -204,8 +200,9 @@ __device__ void packing_keyswitch_lwe_ciphertext_into_glwe_ciphertext(
     // Iterate through all lwe elements
     for (int i = 0; i < lwe_dimension_in; i++) {
       // Round and prepare decomposition
-      Torus state = init_decomposer_state(lwe_in[i], base_log, level_count);
+      Torus a_i = round_to_closest_multiple(lwe_in[i], base_log, level_count);
 
+      Torus state = a_i >> (sizeof(Torus) * 8 - base_log * level_count);
       Torus mod_b_mask = (1ll << base_log) - 1ll;
 
       // block of key for current lwe coefficient (cur_input_lwe[i])
@@ -228,11 +225,12 @@ __device__ void packing_keyswitch_lwe_ciphertext_into_glwe_ciphertext(
 // Assumes there are (glwe_dimension+1) * polynomial_size threads split through
 // different thread blocks at the x-axis to work on that input.
 template <typename Torus>
-__global__ void packing_keyswitch_lwe_list_to_glwe(
-    Torus *glwe_array_out, Torus const *lwe_array_in, Torus const *fp_ksk,
-    uint32_t lwe_dimension_in, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
-    Torus *d_mem) {
+__global__ void
+packing_keyswitch_lwe_list_to_glwe(Torus *glwe_array_out, Torus *lwe_array_in,
+                                   Torus *fp_ksk, uint32_t lwe_dimension_in,
+                                   uint32_t glwe_dimension,
+                                   uint32_t polynomial_size, uint32_t base_log,
+                                   uint32_t level_count, Torus *d_mem) {
   const int tid = threadIdx.x + blockIdx.x * blockDim.x;
 
   const int glwe_accumulator_size = (glwe_dimension + 1) * polynomial_size;
@@ -245,18 +243,17 @@ __global__ void packing_keyswitch_lwe_list_to_glwe(
   auto lwe_in = lwe_array_in + input_id * lwe_size;
   auto ks_glwe_out = d_mem + input_id * glwe_accumulator_size;
   auto glwe_out = glwe_array_out + input_id * glwe_accumulator_size;
-
   // KS LWE to GLWE
-  packing_keyswitch_lwe_ciphertext_into_glwe_ciphertext<Torus>(
+  packing_keyswitch_lwe_ciphertext_into_glwe_ciphertext(
       ks_glwe_out, lwe_in, fp_ksk, lwe_dimension_in, glwe_dimension,
       polynomial_size, base_log, level_count);
 
   // P * x ^degree
   auto in_poly = ks_glwe_out + (tid / polynomial_size) * polynomial_size;
   auto out_result = glwe_out + (tid / polynomial_size) * polynomial_size;
-  polynomial_accumulate_monic_monomial_mul<Torus>(out_result, in_poly, degree,
-                                                  tid % polynomial_size,
-                                                  polynomial_size, 1, true);
+  polynomial_accumulate_monic_monomial_mul(out_result, in_poly, degree,
+                                           tid % polynomial_size,
+                                           polynomial_size, 1, true);
 }
 
 /// To-do: Rewrite this kernel for efficiency
@@ -279,7 +276,7 @@ __global__ void accumulate_glwes(Torus *glwe_out, Torus *glwe_array_in,
 template <typename Torus>
 __host__ void host_packing_keyswitch_lwe_list_to_glwe(
     cudaStream_t stream, uint32_t gpu_index, Torus *glwe_out,
-    Torus const *lwe_array_in, Torus const *fp_ksk_array, int8_t *fp_ks_buffer,
+    Torus *lwe_array_in, Torus *fp_ksk_array, int8_t *fp_ks_buffer,
     uint32_t lwe_dimension_in, uint32_t glwe_dimension,
     uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
     uint32_t num_lwes) {
@@ -298,27 +295,17 @@ __host__ void host_packing_keyswitch_lwe_list_to_glwe(
   dim3 grid(num_blocks, num_lwes);
   dim3 threads(num_threads);
 
-  // The fast path of PKS uses the scratch buffer (d_mem) differently:
-  // it needs to store the decomposed masks in the first half of this buffer
-  // and the keyswitched GLWEs in the second half of the buffer. Thus the
-  // scratch buffer for the fast path must determine the half-size of the
-  // scratch buffer as the max between the size of the GLWE and the size of the
-  // LWE-mask
-  int memory_unit = glwe_accumulator_size > lwe_dimension_in
-                        ? glwe_accumulator_size
-                        : lwe_dimension_in;
-
   auto d_mem = (Torus *)fp_ks_buffer;
-  auto d_tmp_glwe_array_out = d_mem + num_lwes * memory_unit;
+  auto d_tmp_glwe_array_out = d_mem + num_lwes * glwe_accumulator_size;
 
   // individually keyswitch each lwe
-  packing_keyswitch_lwe_list_to_glwe<Torus><<<grid, threads, 0, stream>>>(
+  packing_keyswitch_lwe_list_to_glwe<<<grid, threads, 0, stream>>>(
       d_tmp_glwe_array_out, lwe_array_in, fp_ksk_array, lwe_dimension_in,
       glwe_dimension, polynomial_size, base_log, level_count, d_mem);
   check_cuda_error(cudaGetLastError());
 
   // accumulate to a single glwe
-  accumulate_glwes<Torus><<<num_blocks, threads, 0, stream>>>(
+  accumulate_glwes<<<num_blocks, threads, 0, stream>>>(
       glwe_out, d_tmp_glwe_array_out, glwe_dimension, polynomial_size,
       num_lwes);
   check_cuda_error(cudaGetLastError());