chore(core): remove log and modular standard_dev

Bumps [tj-actions/changed-files](https://github.com/tj-actions/changed-files) from 45.0.6 to 45.0.7.
- [Release notes](https://github.com/tj-actions/changed-files/releases)
- [Changelog](https://github.com/tj-actions/changed-files/blob/main/HISTORY.md)
- [Commits](d6e91a2266...dcc7a0cba8)

---
updated-dependencies:
- dependency-name: tj-actions/changed-files
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <support@github.com>

.editorconfig

@@ -8,8 +8,14 @@ root = true
 end_of_line = lf
 insert_final_newline = true
 
-# 4 space indentation
-[*.rs]
+# 4 space indentation for rust and toml
+[*.{rs,toml}]
 charset = utf-8
 indent_style = space
 indent_size = 4
+
+# 2 for c and js
+[*.{js,json,c,h}]
+charset = utf-8
+indent_style = space
+indent_size = 2

.github/actionlint.yaml

@@ -5,6 +5,7 @@ self-hosted-runner:
     - 4090-desktop
     - large_windows_16_latest
     - large_ubuntu_16
+    - large_ubuntu_16-22.04
 # Configuration variables in array of strings defined in your repository or
 # organization. `null` means disabling configuration variables check.
 # Empty array means no configuration variable is allowed.

.github/actions/hyperstack_setup/action.yml

@@ -0,0 +1,53 @@
+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

@@ -11,22 +11,54 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
 
 jobs:
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (backward-compat-tests)
+    needs: check-user-permission
     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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -39,7 +71,7 @@ jobs:
     name: Backward compatibility tests
     needs: [ setup-instance ]
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: true
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
@@ -47,10 +79,11 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: stable
 
@@ -76,7 +109,7 @@ jobs:
         with:
           persist-credentials: 'false'
           repository: zama-ai/tfhe-backward-compat-data
-          path: tfhe/tfhe-backward-compat-data
+          path: tests/tfhe-backward-compat-data
           lfs: 'true'
           ref: ${{ steps.backward_compat_branch.outputs.branch }}
 
@@ -90,17 +123,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Backward compatibility tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Backward compatibility tests finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (backward-compat-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, backward-compat-tests ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -114,4 +147,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (backward-compat-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (backward-compat-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_fast_tests.yml

@@ -11,18 +11,32 @@ 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' }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' || github.event_name == 'pull_request_target' }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
 
 jobs:
   should-run:
     runs-on: ubuntu-latest
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       csprng_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.csprng_any_changed }}
       zk_pok_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.zk_pok_any_changed }}
@@ -54,22 +68,24 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             dependencies:
               - tfhe/Cargo.toml
-              - concrete-csprng/**
+              - tfhe-csprng/**
               - tfhe-fft/**
               - tfhe-zk-pok/**
               - utils/tfhe-versionable/**
               - utils/tfhe-versionable-derive/**
             csprng:
-              - concrete-csprng/**
+              - tfhe-csprng/**
             zk_pok:
               - tfhe-zk-pok/**
             versionable:
@@ -102,7 +118,7 @@ jobs:
             user_docs:
               - tfhe/src/**
               - '!tfhe/src/c_api/**'
-              - 'tfhe/docs/**.md'
+              - 'tfhe/docs/**/**.md'
               - README.md
 
       - name: Aggregate file changes
@@ -121,18 +137,35 @@ jobs:
         run: |
           echo "any_changed=true" >> "$GITHUB_OUTPUT"
 
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (fast-tests)
-    if: github.event_name != 'pull_request' ||
-      needs.should-run.outputs.any_file_changed == 'true'
-    needs: should-run
+    if: github.event_name == 'workflow_dispatch' ||
+      (github.event_name != 'workflow_dispatch' && needs.should-run.outputs.any_file_changed == 'true')
+    needs: [ should-run, check-user-permission ]
     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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -143,11 +176,9 @@ jobs:
 
   fast-tests:
     name: Fast CPU tests
-    if: github.event_name != 'pull_request' ||
-      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
     needs: [ should-run, setup-instance ]
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: true
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
@@ -155,17 +186,18 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: stable
 
-      - name: Run concrete-csprng tests
+      - name: Run tfhe-csprng tests
         if: needs.should-run.outputs.csprng_test == 'true'
         run: |
-          make test_concrete_csprng
+          make test_tfhe_csprng
 
       - name: Run tfhe-zk-pok tests
         if: needs.should-run.outputs.zk_pok_test == 'true'
@@ -198,7 +230,7 @@ jobs:
 
       - name: Node cache restoration
         id: node-cache
-        uses: actions/cache/restore@6849a6489940f00c2f30c0fb92c6274307ccb58a #v4.1.2
+        uses: actions/cache/restore@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
         with:
           path: |
             ~/.nvm
@@ -211,7 +243,7 @@ jobs:
           make install_node
 
       - name: Node cache save
-        uses: actions/cache/save@6849a6489940f00c2f30c0fb92c6274307ccb58a #v4.1.2
+        uses: actions/cache/save@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
         if: steps.node-cache.outputs.cache-hit != 'true'
         with:
           path: |
@@ -259,17 +291,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Fast AWS tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Fast AWS tests finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (fast-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, fast-tests ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -283,4 +315,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (fast-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (fast-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_integer_tests.yml

@@ -10,16 +10,31 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
   # We clear the cache to reduce memory pressure because of the numerous processes of cargo
   # nextest
   TFHE_RS_CLEAR_IN_MEMORY_KEY_CACHE: "1"
   NO_BIG_PARAMS: FALSE
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
-    types: [labeled]
+    types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
   push:
     branches:
       - main
@@ -28,12 +43,11 @@ jobs:
   should-run:
     if:
       (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs') ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
-      (github.event_name == 'pull_request' && contains(github.event.label.name, 'approved')) ||
+      (github.event_name == 'pull_request_target' && contains(github.event.label.name, 'approved')) ||
       github.event_name == 'workflow_dispatch'
     runs-on: ubuntu-latest
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       integer_test: ${{ github.event_name == 'workflow_dispatch' ||
         steps.changed-files.outputs.integer_any_changed }}
@@ -42,30 +56,49 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          persist-credentials: "false"
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             integer:
               - tfhe/Cargo.toml
-              - concrete-csprng/**
+              - tfhe-csprng/**
               - tfhe-fft/**
               - tfhe-zk-pok/**
               - tfhe/src/core_crypto/**
               - tfhe/src/shortint/**
               - tfhe/src/integer/**
+              - .github/workflows/aws_tfhe_integer_tests.yml
+
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
 
   setup-instance:
     name: Setup instance (unsigned-integer-tests)
-    needs: should-run
+    needs: [ should-run, check-user-permission ]
     if:
       (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs' && needs.should-run.outputs.integer_test == 'true') ||
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
-      (github.event_name == 'pull_request' && contains(github.event.label.name, 'approved')) ||
+      (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.integer_test == 'true') ||
       github.event_name == 'workflow_dispatch'
     runs-on: ubuntu-latest
     outputs:
@@ -73,7 +106,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -86,7 +119,7 @@ jobs:
     name: Unsigned integer tests
     needs: setup-instance
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}${{ github.ref == 'refs/heads/main' && github.sha || '' }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
@@ -94,15 +127,16 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: "false"
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: stable
 
       - name: Should skip big parameters set
-        if: github.event_name == 'pull_request'
+        if: github.event_name == 'pull_request_target'
         run: |
           echo "NO_BIG_PARAMS=TRUE" >> "${GITHUB_ENV}"
 
@@ -128,17 +162,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Unsigned Integer tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Unsigned Integer tests finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (unsigned-integer-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [setup-instance, unsigned-integer-tests]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -152,4 +186,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (unsigned-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (unsigned-integer-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_signed_integer_tests.yml

@@ -10,16 +10,31 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
   # We clear the cache to reduce memory pressure because of the numerous processes of cargo
   # nextest
   TFHE_RS_CLEAR_IN_MEMORY_KEY_CACHE: "1"
   NO_BIG_PARAMS: FALSE
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
-    types: [labeled]
+    types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
   push:
     branches:
       - main
@@ -29,11 +44,11 @@ jobs:
     if:
       (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs') ||
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
-      (github.event_name == 'pull_request' && contains(github.event.label.name, 'approved')) ||
+      ((github.event_name == 'pull_request_target' || github.event_name == 'pull_request_target') && contains(github.event.label.name, 'approved')) ||
       github.event_name == 'workflow_dispatch'
     runs-on: ubuntu-latest
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       integer_test: ${{ github.event_name == 'workflow_dispatch' ||
         steps.changed-files.outputs.integer_any_changed }}
@@ -42,30 +57,49 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          persist-credentials: "false"
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             integer:
               - tfhe/Cargo.toml
-              - concrete-csprng/**
+              - tfhe-csprng/**
               - tfhe-fft/**
               - tfhe-zk-pok/**
               - tfhe/src/core_crypto/**
               - tfhe/src/shortint/**
               - tfhe/src/integer/**
+              - .github/workflows/aws_tfhe_signed_integer_tests.yml
+
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
 
   setup-instance:
     name: Setup instance (unsigned-integer-tests)
-    needs: should-run
+    needs: [ should-run, check-user-permission ]
     if:
       (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs' && needs.should-run.outputs.integer_test == 'true') ||
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
-      (github.event_name == 'pull_request' && contains(github.event.label.name, 'approved')) ||
+      (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.integer_test == 'true') ||
       github.event_name == 'workflow_dispatch'
     runs-on: ubuntu-latest
     outputs:
@@ -73,7 +107,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -86,7 +120,7 @@ jobs:
     name: Signed integer tests
     needs: setup-instance
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}${{ github.ref == 'refs/heads/main' && github.sha || '' }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
@@ -94,15 +128,16 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: "false"
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: stable
 
       - name: Should skip big parameters set
-        if: github.event_name == 'pull_request'
+        if: github.event_name == 'pull_request_target'
         run: |
           echo "NO_BIG_PARAMS=TRUE" >> "${GITHUB_ENV}"
 
@@ -132,17 +167,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Signed Integer tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Signed Integer tests finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (signed-integer-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [setup-instance, signed-integer-tests]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -156,4 +191,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (signed-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (signed-integer-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_tests.yml

@@ -10,13 +10,28 @@ 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' }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' || github.event_name == 'pull_request_target' }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
     types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
   schedule:
     # Nightly tests @ 1AM after each work day
     - cron: "0 1 * * MON-FRI"
@@ -27,7 +42,7 @@ jobs:
     if: github.event_name != 'schedule' ||
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       csprng_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.csprng_any_changed }}
       zk_pok_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.zk_pok_any_changed }}
@@ -40,6 +55,9 @@ 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 }}
@@ -60,20 +78,22 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             dependencies:
               - tfhe/Cargo.toml
-              - concrete-csprng/**
+              - tfhe-csprng/**
               - tfhe-fft/**
               - tfhe-zk-pok/**
             csprng:
-              - concrete-csprng/**
+              - tfhe-csprng/**
             zk_pok:
               - tfhe-zk-pok/**
             core_crypto:
@@ -84,6 +104,11 @@ 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/**'
@@ -102,7 +127,7 @@ jobs:
             user_docs:
               - tfhe/src/**
               - '!tfhe/src/c_api/**'
-              - 'tfhe/docs/**.md'
+              - 'tfhe/docs/**/**.md'
               - README.md
 
       - name: Aggregate file changes
@@ -113,6 +138,7 @@ 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' ||
@@ -121,18 +147,35 @@ jobs:
         run: |
           echo "any_changed=true" >> "$GITHUB_OUTPUT"
 
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (cpu-tests)
-    if: github.event_name != 'pull_request' ||
+    if: github.event_name != 'pull_request_target' ||
       (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.any_file_changed == 'true')
-    needs: should-run
+    needs: [ should-run, check-user-permission ]
     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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -143,11 +186,11 @@ jobs:
 
   cpu-tests:
     name: CPU tests
-    if: github.event_name != 'pull_request' ||
-      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.setup-instance.result != 'skipped')
     needs: [ should-run, setup-instance ]
     concurrency:
-      group: ${{ github.workflow }}_${{github.event_name}}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{github.event_name}}_${{ github.head_ref || github.ref }}
       cancel-in-progress: true
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
@@ -155,17 +198,18 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: stable
 
-      - name: Run concrete-csprng tests
+      - name: Run tfhe-csprng tests
         if: needs.should-run.outputs.csprng_test == 'true'
         run: |
-          make test_concrete_csprng
+          make test_tfhe_csprng
 
       - name: Run tfhe-zk-pok tests
         if: needs.should-run.outputs.zk_pok_test == 'true'
@@ -202,6 +246,11 @@ 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: |
@@ -225,17 +274,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "CPU tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "CPU tests finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (cpu-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, cpu-tests ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -249,4 +298,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cpu-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (cpu-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_wasm_tests.yml

@@ -10,16 +10,49 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
     types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
 
 jobs:
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (wasm-tests)
+    needs: check-user-permission
     if: ${{ github.event_name == 'workflow_dispatch' || contains(github.event.label.name, 'approved') }}
     runs-on: ubuntu-latest
     outputs:
@@ -27,7 +60,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -40,7 +73,7 @@ jobs:
     name: WASM tests
     needs: setup-instance
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: true
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
@@ -48,10 +81,11 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: stable
 
@@ -61,7 +95,7 @@ jobs:
 
       - name: Node cache restoration
         id: node-cache
-        uses: actions/cache/restore@6849a6489940f00c2f30c0fb92c6274307ccb58a #v4.1.2
+        uses: actions/cache/restore@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
         with:
           path: |
             ~/.nvm
@@ -74,7 +108,7 @@ jobs:
           make install_node
 
       - name: Node cache save
-        uses: actions/cache/save@6849a6489940f00c2f30c0fb92c6274307ccb58a #v4.1.2
+        uses: actions/cache/save@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
         if: steps.node-cache.outputs.cache-hit != 'true'
         with:
           path: |
@@ -109,17 +143,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "WASM tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "WASM tests finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (wasm-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, wasm-tests ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -133,4 +167,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (wasm-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (wasm-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ 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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -51,7 +51,8 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -62,7 +63,7 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: nightly
 
@@ -80,8 +81,7 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
-          --name-suffix avx512 \
-          --throughput
+          --name-suffix avx512
 
       - name: Measure key sizes
         run: |
@@ -94,7 +94,7 @@ jobs:
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
           name: ${{ github.sha }}_boolean
           path: ${{ env.RESULTS_FILENAME }}
@@ -104,7 +104,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -122,13 +123,13 @@ jobs:
 
   teardown-instance:
     name: Teardown instance (boolean-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, boolean-benchmarks ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}

.github/workflows/benchmark_core_crypto.yml

@@ -3,6 +3,9 @@ name: Core crypto benchmarks
 
 on:
   workflow_dispatch:
+  schedule:
+    # Weekly benchmarks will be triggered each Saturday at 5a.m.
+    - cron: '0 5 * * 6'
 
 env:
   CARGO_TERM_COLOR: always
@@ -26,7 +29,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -47,7 +50,8 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -58,7 +62,7 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: nightly
 
@@ -78,11 +82,10 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --name-suffix avx512 \
-          --walk-subdirs \
-          --throughput
+          --walk-subdirs
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
           name: ${{ github.sha }}_core_crypto
           path: ${{ env.RESULTS_FILENAME }}
@@ -92,7 +95,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -110,13 +114,13 @@ jobs:
 
   teardown-instance:
     name: Teardown instance (core-crypto-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, core-crypto-benchmarks ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}

.github/workflows/benchmark_erc20.yml

@@ -29,7 +29,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -52,7 +52,8 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -63,7 +64,7 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: nightly
 
@@ -72,7 +73,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Run benchmarks
         run: |
@@ -97,7 +99,7 @@ jobs:
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
           name: ${{ github.sha }}_erc20
           path: ${{ env.RESULTS_FILENAME }}
@@ -118,13 +120,13 @@ jobs:
 
   teardown-instance:
     name: Teardown instance (erc20-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, erc20-benchmarks ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}

.github/workflows/benchmark_gpu_4090.yml

@@ -11,20 +11,53 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
   FAST_BENCH: TRUE
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
-    types: [labeled]
+    types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
   schedule:
     # Weekly benchmarks will be triggered each Friday at 9p.m.
     - cron: "0 21 * * 5"
 
 jobs:
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   cuda-integer-benchmarks:
     name: Cuda integer benchmarks (RTX 4090)
+    needs: check-user-permission
     if: ${{ github.event_name == 'workflow_dispatch' ||
       github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs' ||
       contains(github.event.label.name, '4090_bench') }}
@@ -33,16 +66,14 @@ jobs:
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ["self-hosted", "4090-desktop"]
     timeout-minutes: 1440 # 24 hours
-    strategy:
-      fail-fast: false
-      max-parallel: 1
-
     steps:
       - name: Checkout tfhe-rs
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Get benchmark details
         run: |
@@ -54,7 +85,7 @@ jobs:
           echo "FAST_BENCH=TRUE" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: nightly
 
@@ -63,7 +94,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Run integer benchmarks
         run: |
@@ -79,11 +111,10 @@ jobs:
           --branch ${{ github.ref_name }} \
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
-          --walk-subdirs \
-          --throughput
+          --walk-subdirs
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
           name: ${{ github.sha }}_integer_multi_bit_gpu_default
           path: ${{ env.RESULTS_FILENAME }}
@@ -100,7 +131,7 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Integer RTX 4090 full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Integer RTX 4090 full benchmarks finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   cuda-core-crypto-benchmarks:
     name: Cuda core crypto benchmarks  (RTX 4090)
@@ -117,6 +148,8 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -127,7 +160,7 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: nightly
 
@@ -136,7 +169,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Run core crypto benchmarks
         run: |
@@ -154,10 +188,10 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
-          --throughput
+      
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
           name: ${{ github.sha }}_core_crypto
           path: ${{ env.RESULTS_FILENAME }}
@@ -182,11 +216,11 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Core crypto RTX 4090 full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Core crypto RTX 4090 full benchmarks finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   remove_github_label:
     name: Remove 4090 bench label
-    if: ${{ always() && github.event_name == 'pull_request' }}
+    if: ${{ always() && github.event_name == 'pull_request_target' }}
     needs: [cuda-integer-benchmarks, cuda-core-crypto-benchmarks]
     runs-on: ubuntu-latest
     steps:

.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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -48,28 +48,19 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ 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
-          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
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+      - name: Setup Hyperstack dependencies
+        uses: ./.github/actions/hyperstack_setup
+        with:
+          cuda-version: ${{ matrix.cuda }}
+          gcc-version: ${{ matrix.gcc }}
 
       - name: Get benchmark details
         run: |
@@ -85,31 +76,10 @@ jobs:
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         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";
-          } >> "${GITHUB_ENV}"
-
       - name: Run benchmarks with AVX512
         run: |
           make bench_pbs_gpu
@@ -126,11 +96,10 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --name-suffix avx512 \
-          --walk-subdirs \
-          --throughput
+          --walk-subdirs
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
           name: ${{ github.sha }}_core_crypto
           path: ${{ env.RESULTS_FILENAME }}
@@ -140,7 +109,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -163,13 +133,13 @@ jobs:
 
   teardown-instance:
     name: Teardown instance (cuda-integer-full-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, cuda-core-crypto-benchmarks, slack-notify ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}

.github/workflows/benchmark_gpu_erc20.yml

@@ -1,195 +1,44 @@
-# Run ERC20 benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
-name: ERC20 GPU H100 benchmarks
+# Run CUDA ERC20 benchmarks on a Hyperstack VM and return parsed results to Slab CI bot.
+name: Cuda 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_${{ github.sha }}.json
-  PARSE_INTEGER_BENCH_CSV_FILE: tfhe_rs_integer_benches_${{ 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 }}
+    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:
-  setup-instance:
-    name: Setup instance (cuda-erc20-benchmarks)
+  parse-inputs:
     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 }}
+      profile: ${{ steps.parse_profile.outputs.profile }}
+      hardware_name: ${{ steps.parse_hardware_name.outputs.name }}
     steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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-erc20-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
+      - name: Parse profile
+        id: parse_profile
         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"$(nproc)"
-          sudo make install
+          echo "profile=$(echo '${{ inputs.profile }}' | sed 's|\(.*\)[[:space:]](.*)|\1|')" >> "${GITHUB_OUTPUT}"
 
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Get benchmark details
+      - name: Parse hardware name
+        id: parse_hardware_name
         run: |
-          {
-            echo "BENCH_DATE=$(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)";
-          } >> "${GITHUB_ENV}"
+          echo "name=$(echo '${{ inputs.profile }}' | sed 's|.*[[:space:]](\(.*\))|\1|')" >> "${GITHUB_OUTPUT}"
 
-      - 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
-        run: |
-          make bench_hlapi_erc20_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
-
-      - 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@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
-        with:
-          name: ${{ github.sha }}_erc20
-          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: "Integer GPU benchmarks finished with status: ${{ needs.cuda-erc20-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-erc20-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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+  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_8h100.yml

@@ -1,195 +0,0 @@
-# Run ERC20 benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
-name: ERC20 GPU 8xH100 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_${{ github.sha }}.json
-  PARSE_INTEGER_BENCH_CSV_FILE: tfhe_rs_integer_benches_${{ 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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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-erc20-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: |
-          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
-        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
-        run: |
-          make bench_hlapi_erc20_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
-
-      - 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@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
-        with:
-          name: ${{ github.sha }}_erc20
-          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: "ERC20 8xH100 benchmarks finished with status: ${{ needs.cuda-erc20-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-erc20-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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_erc20_common.yml

@@ -1,11 +1,35 @@
 # Run ERC20 benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
-name: ERC20 GPU 2xH100 benchmarks
+name: Cuda ERC20 benchmarks - common
 
 on:
-  workflow_dispatch:
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 5a.m.
-    - cron: '0 5 * * 6'
+  workflow_call:
+    inputs:
+      backend:
+        type: string
+        default: hyperstack
+      profile:
+        type: string
+        required: true
+      hardware_name:
+        type: string
+        required: true
+    secrets:
+      REPO_CHECKOUT_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
@@ -30,17 +54,17 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        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: 2-h100
+          backend: ${{ inputs.backend }}
+          profile: ${{ inputs.profile }}
 
   cuda-erc20-benchmarks:
-    name: Execute GPU integer benchmarks
+    name: Cuda ERC20 benchmarks (${{ inputs.profile }})
     needs: setup-instance
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     strategy:
@@ -51,27 +75,19 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ 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
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+      - name: Setup Hyperstack dependencies
+        uses: ./.github/actions/hyperstack_setup
+        with:
+          cuda-version: ${{ matrix.cuda }}
+          gcc-version: ${{ matrix.gcc }}
 
       - name: Get benchmark details
         run: |
@@ -87,34 +103,10 @@ jobs:
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         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
         run: |
           make bench_hlapi_erc20_gpu
@@ -123,7 +115,7 @@ jobs:
         run: |
           python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
           --database tfhe_rs \
-          --hardware "n3-H100x2" \
+          --hardware "${{ inputs.hardware_name }}" \
           --backend gpu \
           --project-version "${{ env.COMMIT_HASH }}" \
           --branch ${{ github.ref_name }} \
@@ -132,16 +124,10 @@ jobs:
           --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@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
-          name: ${{ github.sha }}_erc20
+          name: ${{ github.sha }}_erc20_${{ inputs.profile }}
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
@@ -149,7 +135,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -168,17 +155,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ needs.cuda-erc20-benchmarks.result }}
-          SLACK_MESSAGE: "ERC20 2xH100 benchmarks finished with status: ${{ needs.cuda-erc20-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
+          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-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    name: Teardown instance (cuda-erc20-${{ inputs.profile }}-benchmarks)
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -192,4 +179,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-erc20-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          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

@@ -0,0 +1,35 @@
+# 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

@@ -1,201 +1,79 @@
-# Run integer benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
-name: Integer GPU benchmarks
+# Run CUDA benchmarks on a Hyperstack VM and return parsed results to Slab CI bot.
+name: Cuda benchmarks
 
 on:
   workflow_dispatch:
-  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 }}
+    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
 
 jobs:
-  setup-instance:
-    name: Setup instance (cuda-integer-benchmarks)
+  parse-inputs:
     runs-on: ubuntu-latest
-    if:  github.event_name == 'workflow_dispatch' ||
-      (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs')
     outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
+      profile: ${{ steps.parse_profile.outputs.profile }}
+      hardware_name: ${{ steps.parse_hardware_name.outputs.name }}
     steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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-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
+      - name: Parse profile
+        id: parse_profile
         run: |
-          sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev libclang-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"$(nproc)"
-          sudo make install
+          echo "profile=$(echo '${{ inputs.profile }}' | sed 's|\(.*\)[[:space:]](.*)|\1|')" >> "${GITHUB_OUTPUT}"
 
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
-        with:
-          fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Get benchmark details
+      - name: Parse hardware name
+        id: parse_hardware_name
         run: |
-          {
-            echo "BENCH_DATE=$(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)";
-          } >> "${GITHUB_ENV}"
+          echo "name=$(echo '${{ inputs.profile }}' | sed 's|.*[[:space:]](\(.*\))|\1|')" >> "${GITHUB_OUTPUT}"
 
-      - 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@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
-        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@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
-        with:
-          name: ${{ github.sha }}_integer
-          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-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@c33737706dea87cd7784c687dadc9adf1be59990
-        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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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-integer-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+  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

.github/workflows/benchmark_gpu_integer_2H100_full.yml

@@ -1,194 +0,0 @@
-# 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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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 libclang-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
-        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@11bd71901bbe5b1630ceea73d27597364c9af683
-        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@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
-        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@c33737706dea87cd7784c687dadc9adf1be59990
-        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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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-integer-full-2-gpu-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_integer_common.yml

@@ -0,0 +1,258 @@
+# 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:
+      REPO_CHECKOUT_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
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_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@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
+        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
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_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 == 'success' }}
+    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

@@ -1,200 +0,0 @@
-# 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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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 libclang-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
-        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@11bd71901bbe5b1630ceea73d27597364c9af683
-        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@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
-        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@c33737706dea87cd7784c687dadc9adf1be59990
-        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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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-integer-full-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_integer_multi_bit.yml

@@ -1,224 +0,0 @@
-# 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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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 libclang-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
-        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@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
-        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@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
-        with:
-          name: ${{ github.sha }}_integer
-          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-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@c33737706dea87cd7784c687dadc9adf1be59990
-        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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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-integer-multi-bit-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_integer_multi_bit_multi_gpu.yml

@@ -1,214 +0,0 @@
-# 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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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 libclang-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
-        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@11bd71901bbe5b1630ceea73d27597364c9af683
-        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@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
-        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@c33737706dea87cd7784c687dadc9adf1be59990
-        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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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-integer-multi-bit-multi-gpu-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_integer_multi_gpu_full.yml

@@ -1,194 +0,0 @@
-# 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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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 libclang-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
-        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@11bd71901bbe5b1630ceea73d27597364c9af683
-        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@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
-        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@c33737706dea87cd7784c687dadc9adf1be59990
-        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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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-integer-full-multi-gpu-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_integer_weekly.yml

@@ -0,0 +1,60 @@
+# 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

@@ -1,206 +0,0 @@
-# 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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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 libclang-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
-        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@11bd71901bbe5b1630ceea73d27597364c9af683
-        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@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
-        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@c33737706dea87cd7784c687dadc9adf1be59990
-        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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
-        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-l40-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_integer.yml

@@ -8,6 +8,14 @@ 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.
@@ -36,10 +44,10 @@ jobs:
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
     outputs:
       op_flavor: ${{ steps.set_op_flavor.outputs.op_flavor }}
+      bench_type: ${{ steps.set_bench_type.outputs.bench_type }}
     steps:
       - name: Weekly benchmarks
-        if: github.event_name == 'workflow_dispatch' ||
-          github.event.schedule == '0 1 * * 6'
+        if: github.event.schedule == '0 1 * * 6'
         run: |
           echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
 
@@ -48,11 +56,31 @@ jobs:
         run: |
           echo "OP_FLAVOR=[\"default\", \"smart\", \"unchecked\", \"misc\"]" >> "${GITHUB_ENV}"
 
-      -  name: Set operation flavor output
-         id: set_op_flavor
-         run: |
+      - 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 (integer-benchmarks)
     needs: prepare-matrix
@@ -62,7 +90,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -85,12 +113,14 @@ 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
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -101,7 +131,7 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: nightly
 
@@ -110,7 +140,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Should run benchmarks with all precisions
         if: inputs.all_precisions
@@ -119,13 +150,13 @@ jobs:
 
       - name: Run benchmarks with AVX512
         run: |
-          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_${{ matrix.command }}
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} BENCH_TYPE=${{ matrix.bench_type }} bench_${{ matrix.command }}
 
-      # Run these benchmarks only once
+      # Run these benchmarks only once per benchmark type
       - name: Run compression benchmarks with AVX512
         if: matrix.op_flavor == 'default' && matrix.command == 'integer'
         run: |
-          make bench_integer_compression
+          make BENCH_TYPE=${{ matrix.bench_type }} bench_integer_compression
 
       - name: Parse results
         run: |
@@ -138,12 +169,12 @@ jobs:
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
           --name-suffix avx512 \
-          --throughput
+          --bench-type ${{ matrix.bench_type }}
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
-          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}_${{ matrix.bench_type }}
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Send data to Slab
@@ -162,13 +193,13 @@ jobs:
 
   teardown-instance:
     name: Teardown instance (integer-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, integer-benchmarks ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}

.github/workflows/benchmark_shortint.yml

@@ -56,7 +56,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -82,7 +82,8 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -93,7 +94,7 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: nightly
 
@@ -102,7 +103,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Run benchmarks with AVX512
         run: |
@@ -120,8 +122,7 @@ jobs:
           --commit-date "${COMMIT_DATE}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
-          --name-suffix avx512 \
-          --throughput
+          --name-suffix avx512
 
       # This small benchmark needs to be executed only once.
       - name: Measure key sizes
@@ -137,7 +138,7 @@ jobs:
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
           name: ${{ github.sha }}_shortint_${{ matrix.op_flavor }}
           path: ${{ env.RESULTS_FILENAME }}
@@ -158,13 +159,13 @@ jobs:
 
   teardown-instance:
     name: Teardown instance (shortint-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, shortint-benchmarks ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}

.github/workflows/benchmark_signed_integer.yml

@@ -8,6 +8,14 @@ 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.
@@ -36,10 +44,10 @@ jobs:
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
     outputs:
       op_flavor: ${{ steps.set_op_flavor.outputs.op_flavor }}
+      bench_type: ${{ steps.set_bench_type.outputs.bench_type }}
     steps:
       - name: Weekly benchmarks
-        if: github.event_name == 'workflow_dispatch' ||
-          github.event.schedule == '0 1 * * 6'
+        if: github.event.schedule == '0 1 * * 6'
         run: |
           echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
 
@@ -48,11 +56,31 @@ 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
@@ -62,7 +90,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -84,13 +112,15 @@ jobs:
       max-parallel: 1
       matrix:
         command: [ integer, integer_multi_bit ]
-        op_flavor: [ default, unchecked ]
+        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
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -101,7 +131,7 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: nightly
 
@@ -110,7 +140,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Should run benchmarks with all precisions
         if: inputs.all_precisions
@@ -119,7 +150,7 @@ jobs:
 
       - name: Run benchmarks with AVX512
         run: |
-          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_signed_${{ matrix.command }}
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} BENCH_TYPE=${{ matrix.bench_type }} bench_signed_${{ matrix.command }}
 
       - name: Parse results
         run: |
@@ -132,12 +163,12 @@ jobs:
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
           --name-suffix avx512 \
-          --throughput
+          --bench-type ${{ matrix.bench_type }}
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
-          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}_${{ matrix.bench_type }}
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Send data to Slab
@@ -156,13 +187,13 @@ jobs:
 
   teardown-instance:
     name: Teardown instance (integer-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, signed-integer-benchmarks ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}

.github/workflows/benchmark_tfhe_fft.yml

@@ -16,6 +16,9 @@ on:
   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'
@@ -29,7 +32,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@1dced74825027fe3d481392163ed8fc56813fb5d
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -47,7 +50,7 @@ jobs:
     runs-on: ${{ needs.setup-ec2.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@ac593985615ec2ede58e132d2e21d2b1cbd6127c
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
 
@@ -81,17 +84,18 @@ jobs:
           --name-suffix avx512
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b7f8abb1508181956e8e162db84b466c27e18ce
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
           name: ${{ github.sha }}_fft
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@ac593985615ec2ede58e132d2e21d2b1cbd6127c
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -110,7 +114,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "tfhe-fft benchmarks failed. (${{ env.ACTION_RUN_URL }})"
@@ -123,7 +127,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@1dced74825027fe3d481392163ed8fc56813fb5d
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -134,7 +138,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
+        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

@@ -0,0 +1,144 @@
+# 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@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
+        with:
+          name: ${{ github.sha }}_ntt
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          repository: zama-ai/slab
+          path: slab
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          echo "Computing HMac on downloaded artifact"
+          SIGNATURE="$(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

@@ -36,9 +36,8 @@ jobs:
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@c3a1bb2c992d77180ae65be6ae6c166cf40f857c
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             zk_pok:
               - tfhe-zk-pok/**
@@ -58,7 +57,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -80,7 +79,8 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -91,7 +91,7 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: nightly
 
@@ -100,7 +100,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Run benchmarks
         run: |
@@ -118,11 +119,10 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
-          --name-suffix avx512 \
-          --throughput
+          --name-suffix avx512
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
           name: ${{ github.sha }}_tfhe_zk_pok
           path: ${{ env.RESULTS_FILENAME }}
@@ -132,7 +132,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -150,13 +151,13 @@ jobs:
 
   teardown-instance:
     name: Teardown instance (tfhe-zk-pok-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, tfhe-zk-pok-benchmarks ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}

.github/workflows/benchmark_wasm_client.yml

@@ -28,7 +28,7 @@ jobs:
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
       (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs')
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       wasm_bench: ${{ steps.changed-files.outputs.wasm_bench_any_changed }}
     steps:
@@ -36,16 +36,17 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             wasm_bench:
               - tfhe/Cargo.toml
-              - concrete-csprng/**
+              - tfhe-csprng/**
               - tfhe-zk-pok/**
               - tfhe/src/**
               - '!tfhe/src/c_api/**'
@@ -64,7 +65,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -87,7 +88,8 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -98,7 +100,7 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: nightly
 
@@ -108,7 +110,7 @@ jobs:
 
       - name: Node cache restoration
         id: node-cache
-        uses: actions/cache/restore@6849a6489940f00c2f30c0fb92c6274307ccb58a #v4.1.2
+        uses: actions/cache/restore@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
         with:
           path: |
             ~/.nvm
@@ -121,7 +123,7 @@ jobs:
           make install_node
 
       - name: Node cache save
-        uses: actions/cache/save@6849a6489940f00c2f30c0fb92c6274307ccb58a #v4.1.2
+        uses: actions/cache/save@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
         if: steps.node-cache.outputs.cache-hit != 'true'
         with:
           path: |
@@ -165,7 +167,7 @@ jobs:
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
           name: ${{ github.sha }}_wasm_${{ matrix.browser }}
           path: ${{ env.RESULTS_FILENAME }}
@@ -175,7 +177,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -193,13 +196,13 @@ jobs:
 
   teardown-instance:
     name: Teardown instance (wasm-client-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, wasm-client-benchmarks ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}

.github/workflows/benchmark_zk_pke.yml

@@ -3,6 +3,16 @@ name: PKE ZK benchmarks
 
 on:
   workflow_dispatch:
+    inputs:
+      bench_type:
+        description: "Benchmarks type"
+        type: choice
+        default: latency
+        options:
+          - latency
+          - throughput
+          - both
+
   push:
     branches:
       - main
@@ -33,16 +43,17 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             zk_pok:
               - tfhe/Cargo.toml
-              - concrete-csprng/**
+              - tfhe-csprng/**
               - tfhe-fft/**
               - tfhe-zk-pok/**
               - tfhe/src/core_crypto/**
@@ -52,10 +63,37 @@ 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
+    needs: [ should-run, prepare-matrix ]
     if: github.event_name == 'workflow_dispatch' ||
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
       (github.event_name == 'push' &&
@@ -66,7 +104,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -78,17 +116,22 @@ jobs:
   pke-zk-benchmarks:
     name: Execute PKE ZK benchmarks
     if: needs.setup-instance.result != 'skipped'
-    needs: setup-instance
+    needs: [ prepare-matrix, 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
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -99,7 +142,7 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: nightly
 
@@ -108,11 +151,12 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Run benchmarks with AVX512
         run: |
-          make bench_integer_zk
+          make BENCH_TYPE=${{ matrix.bench_type }} bench_integer_zk
 
       - name: Parse results
         run: |
@@ -126,7 +170,7 @@ jobs:
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
           --name-suffix avx512 \
-          --throughput
+          --bench-type ${{ matrix.bench_type }}
 
       - name: Parse CRS sizes results
         run: |
@@ -135,7 +179,7 @@ jobs:
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08
         with:
           name: ${{ github.sha }}_integer_zk
           path: ${{ env.RESULTS_FILENAME }}
@@ -145,7 +189,8 @@ jobs:
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -163,13 +208,13 @@ jobs:
 
   teardown-instance:
     name: Teardown instance (pke-zk-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, pke-zk-benchmarks ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}

.github/workflows/cargo_build.yml

@@ -28,7 +28,7 @@ jobs:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: stable
 
@@ -47,10 +47,10 @@ jobs:
         run: |
           make pcc
 
-      - name: Build concrete-csprng
+      - name: Build tfhe-csprng
         if: ${{ contains(matrix.os, 'ubuntu') }}
         run: |
-          make build_concrete_csprng
+          make build_tfhe_csprng
 
       - name: Build Release core
         if: ${{ contains(matrix.os, 'ubuntu') }}

.github/workflows/cargo_build_tfhe_fft.yml

@@ -2,9 +2,6 @@
 name: Cargo Build tfhe-fft
 
 on:
-  push:
-    branches:
-      - 'main'
   pull_request:
 
 env:
@@ -24,7 +21,7 @@ jobs:
       fail-fast: false
 
     steps:
-      - uses: actions/checkout@ac593985615ec2ede58e132d2e21d2b1cbd6127c
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
 
       - name: Install Rust
         uses: actions-rs/toolchain@16499b5e05bf2e26879000db0c1d13f7e13fa3af

.github/workflows/cargo_build_tfhe_ntt.yml

@@ -0,0 +1,40 @@
+# 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-ntt:
+    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

@@ -2,9 +2,6 @@
 name: Cargo Test tfhe-fft
 
 on:
-  push:
-    branches:
-      - "main"
   pull_request:
 
 env:
@@ -15,14 +12,14 @@ concurrency:
   cancel-in-progress: true
 
 jobs:
-  cargo-tests:
+  cargo-tests-fft:
     runs-on: ${{ matrix.runner_type }}
     strategy:
       matrix:
         runner_type: [ubuntu-latest, macos-latest, windows-latest]
       fail-fast: false
     steps:
-      - uses: actions/checkout@ac593985615ec2ede58e132d2e21d2b1cbd6127c
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
 
       - name: Install Rust
         uses: actions-rs/toolchain@16499b5e05bf2e26879000db0c1d13f7e13fa3af
@@ -41,13 +38,13 @@ jobs:
         run: |
           make test_fft_no_std
 
-  cargo-tests-nightly:
+  cargo-tests-fft-nightly:
     runs-on: ${{ matrix.runner_type }}
     strategy:
       matrix:
         runner_type: [ubuntu-latest, macos-latest, windows-latest]
     steps:
-      - uses: actions/checkout@ac593985615ec2ede58e132d2e21d2b1cbd6127c
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
 
       - name: Install Rust
         uses: actions-rs/toolchain@16499b5e05bf2e26879000db0c1d13f7e13fa3af
@@ -63,10 +60,10 @@ jobs:
         run: |
           make test_fft_no_std_nightly
 
-  cargo-tests-node-js:
+  cargo-tests-fft-node-js:
     runs-on: "ubuntu-latest"
     steps:
-      - uses: actions/checkout@ac593985615ec2ede58e132d2e21d2b1cbd6127c
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
 
       - name: Test node js
         run: |

.github/workflows/cargo_test_ntt.yml

@@ -0,0 +1,54 @@
+# 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-ntt:
+    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-ntt-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_actor_permissions.yml

@@ -0,0 +1,39 @@
+# Check if an actor is a collaborator and has write access
+name: Check Actor Permissions
+
+on:
+  workflow_call:
+    inputs:
+      username:
+        type: string
+        default: ${{ github.triggering_actor }}
+    outputs:
+      is_authorized:
+        value: ${{ jobs.check-actor-permission.outputs.actor_authorized }}
+    secrets:
+      TOKEN:
+        required: true
+
+jobs:
+  check-actor-permission:
+    runs-on: ubuntu-latest
+    outputs:
+      actor_authorized: ${{ steps.check-access.outputs.require-result }}
+    steps:
+      - name: Get User Permission
+        id: check-access
+        uses: actions-cool/check-user-permission@7b90a27f92f3961b368376107661682c441f6103 # v2.3.0
+        with:
+          require: write
+          username: ${{ inputs.username }}
+        env:
+          GITHUB_TOKEN: ${{ secrets.TOKEN }}
+
+      - name: Check User Permission
+        if: ${{ !(inputs.username == 'dependabot[bot]' || inputs.username == 'cla-bot[bot]') &&
+          steps.check-access.outputs.require-result == 'false' }}
+        run: |
+          echo "${{ inputs.username }} 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/check_ci_files_change.yml

@@ -0,0 +1,40 @@
+# Check if there is any change in CI files since last commit
+name: Check changes in CI files
+
+on:
+  workflow_call:
+    inputs:
+      checkout_ref:
+        type: string
+        required: true
+    outputs:
+      ci_file_changed:
+        value: ${{ jobs.check-changes.outputs.ci_file_changed }}
+    secrets:
+      REPO_CHECKOUT_TOKEN:
+        required: true
+
+jobs:
+  check-changes:
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: read
+    outputs:
+      ci_file_changed: ${{ steps.changed-files.outputs.ci_any_changed }}
+    steps:
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ inputs.checkout_ref }}
+
+      - name: Check for file changes
+        id: changed-files
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
+        with:
+          files_yaml: |
+            ci:
+              - .github/**
+              - ci/**

.github/workflows/check_commit.yml

@@ -2,6 +2,7 @@
 name: Check commit and PR compliance
 on:
   pull_request:
+
 jobs:
   check-commit-pr:
     name: Check commit and PR

.github/workflows/check_external_pr.yml

@@ -0,0 +1,32 @@
+# Check if a pull request fulfill pre-conditions to be accepted
+name: Check PR from fork
+
+on:
+  pull_request_target:
+    paths:
+      - '.github/**'
+      - 'ci/**'
+
+jobs:
+  # Fail if the triggering actor is not part of Zama organization.
+  check-user-permission:
+    name: Check event user permissions
+    uses: ./.github/workflows/check_actor_permissions.yml
+    with:
+      username: ${{ github.event.pull_request.user.login }}
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
+  write-comment:
+    name: Write PR comment
+    if: ${{ always() && needs.check-user-permission.outputs.is_authorized == 'false' }}
+    needs: check-user-permission
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: write
+    steps:
+      - name: Write warning
+        uses: thollander/actions-comment-pull-request@24bffb9b452ba05a4f3f77933840a6a841d1b32b
+        with:
+          message: |
+            CI files have changed. Only Zama organization members are authorized to modify these files.

.github/workflows/ci_lint.yml

@@ -14,6 +14,9 @@ jobs:
     steps:
       - name: Checkout tfhe-rs
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Get actionlint
         run: |
@@ -27,7 +30,8 @@ jobs:
           make lint_workflow
 
       - name: Ensure SHA pinned actions
-        uses: zgosalvez/github-actions-ensure-sha-pinned-actions@38608ef4fb69adae7f1eac6eeb88e67b7d083bfd # v3.0.16
+        uses: zgosalvez/github-actions-ensure-sha-pinned-actions@c3a2b64f69b7a1542a68f44d9edbd9ec3fc1455e # v3.0.20
         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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -47,19 +47,19 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: stable
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
           files_yaml: |
             tfhe:
               - tfhe/src/**
-            concrete_csprng:
-              - concrete-csprng/src/**
+            tfhe_csprng:
+              - tfhe-csprng/src/**
 
       - name: Generate Keys
         if: steps.changed-files.outputs.tfhe_any_changed == 'true'
@@ -83,7 +83,7 @@ jobs:
           make test_shortint_cov
 
       - name: Upload tfhe coverage to Codecov
-        uses: codecov/codecov-action@5c47607acb93fed5485fdbf7232e8a31425f672a
+        uses: codecov/codecov-action@13ce06bfc6bbe3ecf90edbbf1bc32fe5978ca1d3
         if: steps.changed-files.outputs.tfhe_any_changed == 'true'
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
@@ -97,7 +97,7 @@ jobs:
           make test_integer_cov
 
       - name: Upload tfhe coverage to Codecov
-        uses: codecov/codecov-action@5c47607acb93fed5485fdbf7232e8a31425f672a
+        uses: codecov/codecov-action@13ce06bfc6bbe3ecf90edbbf1bc32fe5978ca1d3
         if: steps.changed-files.outputs.tfhe_any_changed == 'true'
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
@@ -115,13 +115,13 @@ jobs:
 
   teardown-instance:
     name: Teardown instance (code-coverage)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, code-coverage ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}

.github/workflows/csprng_randomness_tests.yml

@@ -10,16 +10,49 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
     types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
 
 jobs:
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (csprng-randomness-tests)
+    needs: check-user-permission
     if: ${{ github.event_name == 'workflow_dispatch' || contains(github.event.label.name, 'approved') }}
     runs-on: ubuntu-latest
     outputs:
@@ -27,7 +60,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -40,7 +73,7 @@ jobs:
     name: CSPRNG randomness tests
     needs: setup-instance
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: true
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
@@ -48,10 +81,11 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: stable
 
@@ -65,17 +99,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "concrete-csprng randomness check finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "tfhe-csprng randomness check finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (csprng-randomness-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, csprng-randomness-tests ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -89,4 +123,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (csprng-randomness-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (csprng-randomness-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/data_pr_close.yml

@@ -8,6 +8,8 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
   PR_BRANCH: ${{ github.head_ref || github.ref_name }}
   CLOSE_TYPE: ${{ github.event.pull_request.merged && 'merge' || 'close' }}
 
@@ -15,6 +17,8 @@ env:
 on:
   pull_request:
     types: [ closed ]
+  pull_request_target:
+    types: [ closed ]
 
 # The same pattern is used for jobs that use the github api:
 # - save the result of the API call in the env var "GH_API_RES". Since the var is multiline

.github/workflows/gpu_4090_tests.yml

@@ -1,5 +1,5 @@
 # Compile and test tfhe-cuda-backend on an RTX 4090 machine
-name: TFHE Cuda Backend - 4090 full tests
+name: Cuda - 4090 full tests
 
 env:
   CARGO_TERM_COLOR: always
@@ -11,24 +11,57 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
     types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
   schedule:
     # Nightly tests @ 1AM after each work day
     - cron: "0 1 * * MON-FRI"
 
 jobs:
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   cuda-tests-linux:
     name: CUDA tests (RTX 4090)
+    needs: check-user-permission
     if: github.event_name == 'workflow_dispatch' ||
       contains(github.event.label.name, '4090_test') ||
       (github.event_name == 'schedule' &&  github.repository == 'zama-ai/tfhe-rs')
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: true
     runs-on: ["self-hosted", "4090-desktop"]
 
@@ -37,10 +70,11 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: stable
 
@@ -69,7 +103,7 @@ jobs:
           make test_high_level_api_gpu
 
       - uses: actions-ecosystem/action-remove-labels@2ce5d41b4b6aa8503e285553f75ed56e0a40bae0
-        if: ${{ always() && github.event_name == 'pull_request' }}
+        if: ${{ always() && github.event_name == 'pull_request_target' }}
         with:
           labels: 4090_test
           github_token: ${{ secrets.GITHUB_TOKEN }}
@@ -80,4 +114,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "CUDA RTX 4090 tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "CUDA RTX 4090 tests finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_fast_h100_tests.yml

@@ -1,5 +1,5 @@
 # Compile and test tfhe-cuda-backend on an H100 VM on hyperstack
-name: TFHE Cuda Backend - Fast tests on H100
+name: Cuda - Fast tests on H100
 
 env:
   CARGO_TERM_COLOR: always
@@ -11,19 +11,34 @@ 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' }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' || github.event_name == 'pull_request_target' }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
-      types: [ labeled ]
+    types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
 
 jobs:
   should-run:
     runs-on: ubuntu-latest
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
@@ -31,12 +46,14 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             gpu:
               - tfhe/Cargo.toml
@@ -50,15 +67,32 @@ jobs:
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
+              - 'tfhe/docs/**/**.md'
               - '.github/workflows/gpu_fast_h100_tests.yml'
               - scripts/integer-tests.sh
               - ci/slab.toml
 
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (cuda-h100-tests)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
+    needs: [ should-run, check-user-permission ]
+    if: github.event_name != 'pull_request_target' ||
       (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
@@ -67,7 +101,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -79,10 +113,10 @@ jobs:
   cuda-tests-linux:
     name: CUDA H100 tests
     needs: [ should-run, setup-instance ]
-    if: github.event_name != 'pull_request' ||
-      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.setup-instance.result != 'skipped')
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     strategy:
@@ -93,60 +127,29 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ 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
-          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
         with:
           persist-credentials: 'false'
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
+
+      - 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@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         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
@@ -176,17 +179,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         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 }})"
+          SLACK_MESSAGE: "Fast H100 tests finished with status: ${{ needs.cuda-tests-linux.result }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (cuda-h100-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, cuda-tests-linux ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -200,4 +203,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_fast_tests.yml

@@ -1,5 +1,5 @@
 # Compile and test tfhe-cuda-backend on an AWS instance
-name: TFHE Cuda Backend - Fast tests
+name: Cuda - Fast tests
 
 env:
   CARGO_TERM_COLOR: always
@@ -11,18 +11,32 @@ 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' }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' || github.event_name == 'pull_request_target' }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
 
 jobs:
   should-run:
     runs-on: ubuntu-latest
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
@@ -30,12 +44,14 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             gpu:
               - tfhe/Cargo.toml
@@ -49,15 +65,32 @@ jobs:
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
+              - 'tfhe/docs/**/**.md'
               - '.github/workflows/gpu_fast_tests.yml'
               - scripts/integer-tests.sh
               - ci/slab.toml
 
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (cuda-tests)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
+    needs: [ should-run, check-user-permission ]
+    if: github.event_name == 'workflow_dispatch' ||
       needs.should-run.outputs.gpu_test == 'true'
     runs-on: ubuntu-latest
     outputs:
@@ -65,7 +98,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -77,10 +110,10 @@ jobs:
   cuda-tests-linux:
     name: CUDA tests
     needs: [ should-run, setup-instance ]
-    if: github.event_name != 'pull_request' ||
-      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.setup-instance.result != 'skipped')
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     strategy:
@@ -91,60 +124,29 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ 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
-          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
         with:
           persist-credentials: 'false'
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
+
+      - 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@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         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
@@ -174,17 +176,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         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 }})"
+          SLACK_MESSAGE: "Base GPU tests finished with status: ${{ needs.cuda-tests-linux.result }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (cuda-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, cuda-tests-linux ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -198,4 +200,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (cuda-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_full_h100_tests.yml

@@ -1,5 +1,5 @@
 # Compile and test tfhe-cuda-backend on an H100 VM on hyperstack
-name: TFHE Cuda Backend - Full tests on H100
+name: Cuda - Full tests on H100
 
 env:
   CARGO_TERM_COLOR: always
@@ -11,7 +11,6 @@ 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' }}
 
 on:
   workflow_dispatch:
@@ -25,7 +24,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -49,9 +48,6 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
     steps:
       # Mandatory on hyperstack since a bootable volume is not re-usable yet.
       - name: Install dependencies
@@ -69,40 +65,23 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_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@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         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
@@ -130,7 +109,7 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         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 }})"
+          SLACK_MESSAGE: "Full H100 tests finished with status: ${{ needs.cuda-tests-linux.result }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (cuda-h100-tests)
@@ -139,7 +118,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -153,4 +132,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_full_multi_gpu_tests.yml

@@ -1,5 +1,5 @@
 # Compile and test tfhe-cuda-backend on an AWS instance
-name: TFHE Cuda Backend - Full tests multi-GPU
+name: Cuda - Full tests multi-GPU
 
 env:
   CARGO_TERM_COLOR: always
@@ -11,19 +11,34 @@ 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' }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' || github.event_name == 'pull_request_target' }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
     types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
 
 jobs:
   should-run:
     runs-on: ubuntu-latest
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
@@ -31,12 +46,14 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             gpu:
               - tfhe/Cargo.toml
@@ -50,15 +67,32 @@ jobs:
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
+              - 'tfhe/docs/**/**.md'
               - '.github/workflows/**_multi_gpu_tests.yml'
               - scripts/integer-tests.sh
               - ci/slab.toml
 
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (cuda-tests-multi-gpu)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
+    needs: [ should-run, check-user-permission ]
+    if: github.event_name != 'pull_request_target' ||
       (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
@@ -67,7 +101,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -79,10 +113,10 @@ jobs:
   cuda-tests-linux:
     name: CUDA multi-GPU tests
     needs: [ should-run, setup-instance ]
-    if: github.event_name != 'pull_request' ||
-      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.setup-instance.result != 'skipped')
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     strategy:
@@ -93,60 +127,29 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ 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
-          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
         with:
           persist-credentials: 'false'
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
+
+      - 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@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         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
@@ -179,17 +182,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         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 }})"
+          SLACK_MESSAGE: "Multi-GPU tests finished with status: ${{ needs.cuda-tests-linux.result }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (cuda-tests-multi-gpu)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, cuda-tests-linux ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -203,4 +206,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-tests-multi-gpu) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (cuda-tests-multi-gpu) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_integer_long_run_tests.yml

@@ -0,0 +1,115 @@
+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 == 'success' }}
+    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

@@ -1,5 +1,5 @@
 # Perfom tfhe-cuda-backend post-commit checks on an AWS instance
-name: TFHE Cuda Backend - Post-commit Checks
+name: Cuda - Post-commit Checks
 
 env:
   CARGO_TERM_COLOR: always
@@ -11,20 +11,52 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
 
 jobs:
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (cuda-pcc)
+    needs: check-user-permission
     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@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -37,7 +69,7 @@ jobs:
     name: CUDA post-commit checks
     needs: setup-instance
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: true
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     strategy:
@@ -56,14 +88,15 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: stable
 
@@ -100,17 +133,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "CUDA AWS post-commit checks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "CUDA AWS post-commit checks finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (cuda-pcc)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, cuda-pcc ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -124,4 +157,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-pcc) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (cuda-pcc) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_signed_integer_classic_tests.yml

@@ -1,5 +1,5 @@
 # Signed integer GPU tests on an RTXA6000 VM on hyperstack with classical PBS
-name: TFHE Cuda Backend - Signed integer tests with classical PBS
+name: Cuda - Signed integer tests with classical PBS
 
 env:
   CARGO_TERM_COLOR: always
@@ -11,19 +11,34 @@ 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' }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' || github.event_name == 'pull_request_target' }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
-      types: [ labeled ]
+    types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
 
 jobs:
   should-run:
     runs-on: ubuntu-latest
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
@@ -31,13 +46,14 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             gpu:
               - tfhe/Cargo.toml
@@ -51,15 +67,32 @@ jobs:
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
+              - 'tfhe/docs/**/**.md'
               - '.github/workflows/gpu_signed_integer_classic_tests.yml'
               - scripts/integer-tests.sh
               - ci/slab.toml
 
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (cuda-signed-classic-tests)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
+    needs: [ should-run, check-user-permission ]
+    if: github.event_name != 'pull_request_target' ||
       (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
@@ -68,7 +101,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -80,10 +113,10 @@ jobs:
   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')
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.setup-instance.result != 'skipped')
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     strategy:
@@ -94,58 +127,29 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ 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
-          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
+        with:
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
+
+      - 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@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         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
@@ -161,17 +165,17 @@ jobs:
         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 }})"
+          SLACK_MESSAGE: "Integer GPU signed integer tests with classical PBS finished with status: ${{ needs.cuda-tests-linux.result }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (cuda-signed-classic-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, cuda-tests-linux ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -185,4 +189,4 @@ jobs:
         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 }})"
+          SLACK_MESSAGE: "Instance teardown (cuda-signed-classic-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_signed_integer_h100_tests.yml

@@ -1,5 +1,5 @@
 # Signed integer GPU tests on an H100 VM on hyperstack
-name: TFHE Cuda Backend - Signed integer tests on H100
+name: Cuda - Signed integer tests on H100
 
 env:
   CARGO_TERM_COLOR: always
@@ -11,19 +11,34 @@ 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' }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' || github.event_name == 'pull_request_target' }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
-      types: [ labeled ]
+    types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
 
 jobs:
   should-run:
     runs-on: ubuntu-latest
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
@@ -31,13 +46,14 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             gpu:
               - tfhe/Cargo.toml
@@ -51,15 +67,32 @@ jobs:
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
+              - 'tfhe/docs/**/**.md'
               - '.github/workflows/gpu_signed_integer_h100_tests.yml'
               - scripts/integer-tests.sh
               - ci/slab.toml
 
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (cuda-h100-tests)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
+    needs: [ should-run, check-user-permission ]
+    if: github.event_name != 'pull_request_target' ||
       (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
@@ -68,7 +101,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -80,10 +113,10 @@ jobs:
   cuda-tests-linux:
     name: CUDA H100 signed integer tests
     needs: [ should-run, setup-instance ]
-    if: github.event_name != 'pull_request' ||
-      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.setup-instance.result != 'skipped')
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     strategy:
@@ -94,58 +127,29 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ 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
-          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
+        with:
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
+
+      - 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@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         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 multi-bit tests
         run: |
           BIG_TESTS_INSTANCE=TRUE make test_signed_integer_multi_bit_gpu_ci
@@ -161,17 +165,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         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 }})"
+          SLACK_MESSAGE: "Integer GPU H100 tests finished with status: ${{ needs.cuda-tests-linux.result }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (cuda-h100-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, cuda-tests-linux ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -185,4 +189,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_signed_integer_tests.yml

@@ -1,5 +1,5 @@
 # Compile and test tfhe-cuda-backend signed integer on an AWS instance
-name: TFHE Cuda Backend - Signed integer tests
+name: Cuda - Signed integer tests
 
 env:
   CARGO_TERM_COLOR: always
@@ -11,17 +11,28 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
   FAST_TESTS: TRUE
   NIGHTLY_TESTS: FALSE
-  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' }}
+  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' || github.event_name == 'pull_request_target' }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
-    types:
-      - opened
-      - synchronize
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
   schedule:
     # Nightly tests @ 1AM after each work day
     - cron: "0 1 * * MON-FRI"
@@ -30,7 +41,7 @@ jobs:
   should-run:
     runs-on: ubuntu-latest
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
@@ -38,12 +49,14 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             gpu:
               - tfhe/Cargo.toml
@@ -57,15 +70,32 @@ jobs:
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
+              - 'tfhe/docs/**/**.md'
               - '.github/workflows/gpu_signed_integer_tests.yml'
               - scripts/integer-tests.sh
               - ci/slab.toml
 
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (cuda-signed-integer-tests)
     runs-on: ubuntu-latest
-    needs: should-run
+    needs: [ should-run, check-user-permission ]
     if: (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
       github.event_name == 'workflow_dispatch' ||
       needs.should-run.outputs.gpu_test == 'true'
@@ -74,7 +104,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -86,10 +116,10 @@ jobs:
   cuda-signed-integer-tests:
     name: CUDA signed integer tests
     needs: [ should-run, setup-instance ]
-    if: github.event_name != 'pull_request' ||
-      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.setup-instance.result != 'skipped')
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     strategy:
@@ -100,57 +130,29 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ 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
-          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
         with:
           persist-credentials: 'false'
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
+
+      - 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@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         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";
-          } >> "${GITHUB_ENV}"
-
       - name: Should run nightly tests
         if: github.event_name == 'schedule'
         run: |
@@ -159,10 +161,6 @@ jobs:
             echo "NIGHTLY_TESTS=TRUE";
           } >> "${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
@@ -178,17 +176,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         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 }})"
+          SLACK_MESSAGE: "Base GPU tests finished with status: ${{ needs.cuda-signed-integer-tests.result }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (cuda-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, cuda-signed-integer-tests ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -202,4 +200,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-signed-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (cuda-signed-integer-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_unsigned_integer_classic_tests.yml

@@ -1,5 +1,5 @@
 # Test unsigned integers on an RTXA6000 VM on hyperstack with the classical PBS
-name: TFHE Cuda Backend - Unsigned integer tests with classical PBS
+name: Cuda - Unsigned integer tests with classical PBS
 
 env:
   CARGO_TERM_COLOR: always
@@ -11,19 +11,34 @@ 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' }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' || github.event_name == 'pull_request_target' }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
-      types: [ labeled ]
+    types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
 
 jobs:
   should-run:
     runs-on: ubuntu-latest
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
@@ -31,13 +46,14 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             gpu:
               - tfhe/Cargo.toml
@@ -51,15 +67,32 @@ jobs:
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
+              - 'tfhe/docs/**/**.md'
               - '.github/workflows/gpu_unsigned_integer_classic_tests.yml'
               - scripts/integer-tests.sh
               - ci/slab.toml
 
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (cuda-unsigned-classic-tests)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
+    needs: [ should-run, check-user-permission ]
+    if: github.event_name == 'workflow_dispatch' ||
       (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
@@ -68,7 +101,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -80,10 +113,10 @@ jobs:
   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')
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.setup-instance.result != 'skipped')
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     strategy:
@@ -94,58 +127,29 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ 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
-          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
+        with:
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
+
+      - 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@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         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
@@ -161,17 +165,17 @@ jobs:
         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 }})"
+          SLACK_MESSAGE: "Unsigned integer GPU classic tests finished with status: ${{ needs.cuda-tests-linux.result }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (cuda-unsigned-classic-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, cuda-tests-linux ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -185,4 +189,4 @@ jobs:
         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 }})"
+          SLACK_MESSAGE: "Instance teardown (cuda-unsigned-classic-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_unsigned_integer_h100_tests.yml

@@ -1,5 +1,5 @@
 # Test unsigned integers on an H100 VM on hyperstack
-name: TFHE Cuda Backend - Unsigned integer tests on H100
+name: Cuda - Unsigned integer tests on H100
 
 env:
   CARGO_TERM_COLOR: always
@@ -11,19 +11,34 @@ 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' }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
+  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' || github.event_name == 'pull_request_target' }}
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
-      types: [ labeled ]
+    types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
 
 jobs:
   should-run:
     runs-on: ubuntu-latest
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
@@ -31,13 +46,14 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             gpu:
               - tfhe/Cargo.toml
@@ -51,15 +67,32 @@ jobs:
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
+              - 'tfhe/docs/**/**.md'
               - '.github/workflows/gpu_unsigned_integer_h100_tests.yml'
               - scripts/integer-tests.sh
               - ci/slab.toml
 
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (cuda-h100-tests)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
+    needs: [ should-run, check-user-permission ]
+    if: github.event_name == 'workflow_dispatch' ||
       (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
@@ -68,7 +101,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -80,10 +113,10 @@ jobs:
   cuda-tests-linux:
     name: CUDA H100 unsigned integer tests
     needs: [ should-run, setup-instance ]
-    if: github.event_name != 'pull_request' ||
-      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.setup-instance.result != 'skipped')
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     strategy:
@@ -94,58 +127,29 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ 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
-          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
+        with:
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
+
+      - 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@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         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 multi-bit tests
         run: |
           BIG_TESTS_INSTANCE=TRUE make test_unsigned_integer_multi_bit_gpu_ci
@@ -161,17 +165,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         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 }})"
+          SLACK_MESSAGE: "Unsigned integer GPU H100 tests finished with status: ${{ needs.cuda-tests-linux.result }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (cuda-h100-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, cuda-tests-linux ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -185,4 +189,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_unsigned_integer_tests.yml

@@ -1,5 +1,5 @@
 # Compile and test tfhe-cuda-backend unsigned integer on an AWS instance
-name: TFHE Cuda Backend - Unsigned integer tests
+name: Cuda - Unsigned integer tests
 
 env:
   CARGO_TERM_COLOR: always
@@ -11,16 +11,29 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  MSG_MINIMAL: event,action url,commit
+  BRANCH: ${{ github.head_ref || github.ref }}
   FAST_TESTS: TRUE
   NIGHTLY_TESTS: FALSE
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
-    types:
-      - opened
-      - synchronize
+    types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
   schedule:
     # Nightly tests @ 1AM after each work day
     - cron: "0 1 * * MON-FRI"
@@ -29,7 +42,7 @@ jobs:
   should-run:
     runs-on: ubuntu-latest
     permissions:
-      pull-requests: write
+      pull-requests: read
     outputs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
@@ -37,13 +50,14 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@4edd678ac3f81e2dc578756871e4d00c19191daf
+        uses: tj-actions/changed-files@dcc7a0cba800f454d79fff4b993e8c3555bcc0a8
         with:
-          since_last_remote_commit: true
           files_yaml: |
             gpu:
               - tfhe/Cargo.toml
@@ -57,14 +71,31 @@ jobs:
               - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
+              - 'tfhe/docs/**/**.md'
               - '.github/workflows/gpu_unsigned_integer_tests.yml'
               - scripts/integer-tests.sh
               - ci/slab.toml
 
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   setup-instance:
     name: Setup instance (cuda-unsigned-integer-tests)
-    needs: should-run
+    needs: [ should-run, check-user-permission ]
     if: (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
       github.event_name == 'workflow_dispatch' ||
       needs.should-run.outputs.gpu_test == 'true'
@@ -74,7 +105,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -86,10 +117,10 @@ jobs:
   cuda-unsigned-integer-tests:
     name: CUDA unsigned integer tests
     needs: [ should-run, setup-instance ]
-    if: github.event_name != 'pull_request' ||
-      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.setup-instance.result != 'skipped')
     concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     strategy:
@@ -100,54 +131,29 @@ jobs:
           - os: ubuntu-22.04
             cuda: "12.2"
             gcc: 11
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ 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
-          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
+        with:
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
+
+      - 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@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         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";
-          } >> "${GITHUB_ENV}"
-
       - name: Should run nightly tests
         if: github.event_name == 'schedule'
         run: |
@@ -156,10 +162,6 @@ jobs:
             echo "NIGHTLY_TESTS=TRUE";
           } >> "${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
@@ -175,17 +177,17 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         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 }})"
+          SLACK_MESSAGE: "Unsigned integer GPU tests finished with status: ${{ needs.cuda-unsigned-integer-tests.result }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (cuda-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, cuda-unsigned-integer-tests ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -199,4 +201,4 @@ jobs:
         uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-unsigned-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (cuda-unsigned-integer-tests) finished with status: ${{ job.status }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"

.github/workflows/integer_long_run_tests.yml

@@ -0,0 +1,95 @@
+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.REPO_CHECKOUT_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 == 'success' }}
+    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

@@ -2,8 +2,20 @@ name: Tests on M1 CPU
 
 on:
   workflow_dispatch:
+  # Trigger pull_request event on CI files to be able to test changes before merging to main branch.
+  # Workflow would fail if changes come from a forked repository since secrets are not available with this event.
   pull_request:
-    types: [labeled]
+    types: [ labeled ]
+    paths:
+      - '.github/**'
+      - 'ci/**'
+  # General entry point for Zama's pull request as well as contribution from forks.
+  pull_request_target:
+    types: [ labeled ]
+    paths:
+      - '**'
+      - '!.github/**'
+      - '!ci/**'
   # Have a nightly build for M1 tests
   schedule:
     # * is a special character in YAML so you have to quote this string
@@ -21,14 +33,35 @@ env:
   # We clear the cache to reduce memory pressure because of the numerous processes of cargo
   # nextest
   TFHE_RS_CLEAR_IN_MEMORY_KEY_CACHE: "1"
+  REF: ${{ github.event.pull_request.head.sha || github.sha }}
 
 concurrency:
-  group: ${{ github.workflow }}-${{ github.head_ref }}
+  group: ${{ github.workflow }}_${{ github.head_ref || github.ref }}
   cancel-in-progress: true
 
 jobs:
+  check-ci-files:
+    uses: ./.github/workflows/check_ci_files_change.yml
+    with:
+      checkout_ref: ${{ github.event.pull_request.head.sha || github.sha }}
+    secrets:
+      REPO_CHECKOUT_TOKEN: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+
+  # Fail if the triggering actor is not part of Zama organization.
+  # If pull_request_target is emitted and CI files have changed, skip this job. This would skip following jobs.
+  check-user-permission:
+    needs: check-ci-files
+    if: github.event_name != 'pull_request_target' ||
+      (github.event_name == 'pull_request_target' && needs.check-ci-files.outputs.ci_file_changed == 'false')
+    uses: ./.github/workflows/check_actor_permissions.yml
+    secrets:
+      TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
   cargo-builds-m1:
-    if: ${{ (github.event_name == 'schedule' &&  github.repository == 'zama-ai/tfhe-rs') || github.event_name == 'workflow_dispatch' || contains(github.event.label.name, 'm1_test') }}
+    needs: check-user-permission
+    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
@@ -37,9 +70,11 @@ jobs:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: "false"
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+          ref: ${{ env.REF }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: stable
 
@@ -64,13 +99,29 @@ jobs:
           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 concrete-csprng
+      - name: Build tfhe-csprng
         run: |
-          make build_concrete_csprng
+          make build_tfhe_csprng
 
       - name: Build Release core
         run: |
@@ -96,9 +147,9 @@ jobs:
         run: |
           make build_c_api
 
-      - name: Run concrete-csprng tests
+      - name: Run tfhe-csprng tests
         run: |
-          make test_concrete_csprng
+          make test_tfhe_csprng
 
       - name: Run tfhe-zk-pok tests
         run: |
@@ -162,7 +213,7 @@ jobs:
     if: ${{ always() }}
     steps:
       - uses: actions-ecosystem/action-remove-labels@2ce5d41b4b6aa8503e285553f75ed56e0a40bae0
-        if: ${{ github.event_name == 'pull_request' }}
+        if: ${{ github.event_name == 'pull_request_target' }}
         with:
           labels: m1_test
           github_token: ${{ secrets.GITHUB_TOKEN }}
@@ -175,6 +226,8 @@ jobs:
           SLACK_COLOR: ${{ needs.cargo-builds-m1.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: ${{ needs.cargo-builds-m1.result }} on '${{ env.BRANCH }}'. (${{ env.ACTION_RUN_URL }})"
           SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
           SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+          MSG_MINIMAL: event,action url,commit
+          BRANCH: ${{ github.head_ref || github.ref }}

.github/workflows/make_release.yml

@@ -43,13 +43,15 @@ jobs:
       hash: ${{ steps.hash.outputs.hash }}
     steps:
       - name: Checkout
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
         with:
           fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
       - name: Prepare package
         run: |
           cargo package -p tfhe
-      - uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882 # v4.4.3
+      - uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08 # v4.6.0
         with:
           name: crate
           path: target/package/*.crate
@@ -76,14 +78,13 @@ jobs:
     name: Publish Release
     needs: [package] # for comparing hashes
     runs-on: ubuntu-latest
-    permissions:
-      contents: read
-      id-token: write
     steps:
       - name: Checkout
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
         with:
           fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
       - name: Create NPM version tag
         if: ${{ inputs.npm_latest_tag }}
         run: |
@@ -108,7 +109,7 @@ 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@c33737706dea87cd7784c687dadc9adf1be59990 # v2.3.2
         env:
           SLACK_COLOR: failure
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
@@ -153,7 +154,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990 # v2.3.2
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/make_release_concrete_csprng.yml

@@ -1,48 +0,0 @@
-name: Publish concrete-csprng 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 concrete-csprng Release
-    needs: verify_tag
-    runs-on: ubuntu-latest
-    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 concrete-csprng --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: "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

@@ -1,4 +1,3 @@
-# Publish new release of tfhe-cuda-backend on crates.io.
 name: Publish CUDA release
 
 on:
@@ -8,10 +7,6 @@ on:
         description: "Dry-run"
         type: boolean
         default: true
-      push_to_crates:
-        description: "Push to crate"
-        type: boolean
-        default: true
 
 env:
   ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
@@ -36,7 +31,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -45,10 +40,12 @@ jobs:
           backend: aws
           profile: gpu-build
 
-  publish-cuda-release:
-    name: Publish CUDA Release
+  package:
+    name: Package CUDA Release for provenance
     needs: setup-instance
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    outputs:
+      hash: ${{ steps.hash.outputs.hash }}
     strategy:
       fail-fast: false
       # explicit include-based build matrix, of known valid options
@@ -61,16 +58,18 @@ jobs:
       CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
     steps:
       - name: Checkout
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
         with:
           fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@a54c7afa936fefeb4456b2dd8068152669aa8203
         with:
           toolchain: stable
 
@@ -94,32 +93,81 @@ jobs:
             echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
             echo "HOME=/home/ubuntu";
           } >> "${GITHUB_ENV}"
+      - name: Prepare package
+        run: |
+          cargo package -p tfhe-cuda-backend
+      - name: generate hash
+        id: hash
+        run: cd target/package && echo "hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
 
+  provenance:
+    if: ${{ !inputs.dry_run  }}
+    needs: [package]
+    uses: slsa-framework/slsa-github-generator/.github/workflows/generator_generic_slsa3.yml@v2.0.0
+    permissions:
+      # Needed to detect the GitHub Actions environment
+      actions: read
+      # Needed to create the provenance via GitHub OIDC
+      id-token: write
+      # Needed to upload assets/artifacts
+      contents: write
+    with:
+      # SHA-256 hashes of the Crate package.
+      base64-subjects: ${{ needs.package.outputs.hash }}
+
+  publish-cuda-release:
+    name: Publish CUDA Release
+    needs: [setup-instance, package] # for comparing hashes
+    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: 9
+    steps:
       - name: Publish crate.io package
-        if: ${{ inputs.push_to_crates }}
         env:
           CRATES_TOKEN: ${{ secrets.CARGO_REGISTRY_TOKEN }}
           DRY_RUN: ${{ inputs.dry_run && '--dry-run' || '' }}
         run: |
           cargo publish -p tfhe-cuda-backend --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
 
+      - name: Generate hash
+        id: published_hash
+        run: cd target/package && echo "pub_hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
+
+      - 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 # v2.3.2
+        env:
+          SLACK_COLOR: failure
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "SLSA tfhe-cuda-backend crate - hash comparison failure: (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990 # v2.3.2
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "tfhe-cuda-backend release finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (publish-release)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    if: ${{ always() && needs.setup-instance.result == 'success' }}
     needs: [ setup-instance, publish-cuda-release ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@801df0b8db5ea2b06128b7476c652f5ed5f193a8
+        uses: zama-ai/slab-github-runner@79939325c3c429837c10d6041e4fd8589d328bac
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}

.github/workflows/make_release_tfhe_csprng.yml

@@ -0,0 +1,103 @@
+name: Publish tfhe-csprng 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 }}
+
+  package:
+    runs-on: ubuntu-latest
+    outputs:
+      hash: ${{ steps.hash.outputs.hash }}
+    steps:
+      - name: Checkout
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+        with:
+          fetch-depth: 0
+      - name: Prepare package
+        run: |
+          cargo package -p tfhe-csprng
+      - uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08 # v4.6.0
+        with:
+          name: crate-tfhe-csprng
+          path: target/package/*.crate
+      - name: generate hash
+        id: hash
+        run: cd target/package && echo "hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
+
+
+  provenance:
+    if: ${{ !inputs.dry_run  }}
+    needs: [package]
+    uses: slsa-framework/slsa-github-generator/.github/workflows/generator_generic_slsa3.yml@v2.0.0
+    permissions:
+      # Needed to detect the GitHub Actions environment
+      actions: read
+      # Needed to create the provenance via GitHub OIDC
+      id-token: write
+      # Needed to upload assets/artifacts
+      contents: write
+    with:
+      # SHA-256 hashes of the Crate package.
+      base64-subjects: ${{ needs.package.outputs.hash }}
+
+
+  publish_release:
+    name: Publish tfhe-csprng Release
+    needs: [verify_tag, package]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+      - name: Download artifact
+        uses: actions/download-artifact@fa0a91b85d4f404e444e00e005971372dc801d16 # v4.1.8
+        with:
+          name: crate-tfhe-csprng
+          path: target/package
+      - 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 }} ${{ env.DRY_RUN }}
+      - name: Generate hash
+        id: published_hash
+        run: cd target/package && echo "pub_hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
+      - 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 # v2.3.2
+        env:
+          SLACK_COLOR: failure
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "SLSA tfhe-csprng - hash comparison failure: (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990 # v2.3.2
+        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_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/make_release_tfhe_fft.yml

@@ -19,26 +19,81 @@ jobs:
       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 # v4.2.2
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+      - name: Prepare package
+        run: |
+          cargo package -p tfhe-fft
+      - uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08 # v4.6.0
+        with:
+          name: crate
+          path: target/package/*.crate
+      - name: generate hash
+        id: hash
+        run: cd target/package && echo "hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
+
+  provenance:
+    if: ${{ !inputs.dry_run  }}
+    needs: [package]
+    uses: slsa-framework/slsa-github-generator/.github/workflows/generator_generic_slsa3.yml@v2.0.0
+    permissions:
+      # Needed to detect the GitHub Actions environment
+      actions: read
+      # Needed to create the provenance via GitHub OIDC
+      id-token: write
+      # Needed to upload assets/artifacts
+      contents: write
+    with:
+      # SHA-256 hashes of the Crate package.
+      base64-subjects: ${{ needs.package.outputs.hash }}
+
   publish_release:
     name: Publish tfhe-fft Release
     runs-on: ubuntu-latest
-    needs: verify_tag
+    needs: [verify_tag, package] # for comparing hashes
     steps:
       - name: Checkout
-        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
         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-fft --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
 
+      - name: Generate hash
+        id: published_hash
+        run: cd target/package && echo "pub_hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
+
+      - 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 # v2.3.2
+        env:
+          SLACK_COLOR: failure
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "SLSA tfhe-fft crate - hash comparison failure: (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990 # v2.3.2
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/make_release_tfhe_ntt.yml

@@ -0,0 +1,102 @@
+# 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 }}
+
+  package:
+    runs-on: ubuntu-latest
+    needs: verify_tag
+    outputs:
+      hash: ${{ steps.hash.outputs.hash }}
+    steps:
+      - name: Checkout
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+      - name: Prepare package
+        run: |
+          cargo package -p tfhe-ntt
+      - uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08 # v4.6.0
+        with:
+          name: crate
+          path: target/package/*.crate
+      - name: generate hash
+        id: hash
+        run: cd target/package && echo "hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
+
+  provenance:
+    if: ${{ !inputs.dry_run  }}
+    needs: [package]
+    uses: slsa-framework/slsa-github-generator/.github/workflows/generator_generic_slsa3.yml@v2.0.0
+    permissions:
+      # Needed to detect the GitHub Actions environment
+      actions: read
+      # Needed to create the provenance via GitHub OIDC
+      id-token: write
+      # Needed to upload assets/artifacts
+      contents: write
+    with:
+      # SHA-256 hashes of the Crate package.
+      base64-subjects: ${{ needs.package.outputs.hash }}
+
+  publish_release:
+    name: Publish tfhe-ntt Release
+    runs-on: ubuntu-latest
+    needs: [verify_tag, package] # for comparing hashes
+    steps:
+      - name: Checkout
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+        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 }} ${{ env.DRY_RUN }}
+
+      - name: Generate hash
+        id: published_hash
+        run: cd target/package && echo "pub_hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
+
+      - 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 # v2.3.2
+        env:
+          SLACK_COLOR: failure
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "SLSA tfhe-ntt crate - hash comparison failure: (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990 # v2.3.2
+        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

@@ -18,34 +18,161 @@ jobs:
       RELEASE_TEAM: ${{ secrets.RELEASE_TEAM }}
       READ_ORG_TOKEN: ${{ secrets.READ_ORG_TOKEN }}
 
+  package-derive:
+    runs-on: ubuntu-latest
+    outputs:
+      hash: ${{ steps.hash.outputs.hash }}
+    steps:
+      - name: Checkout
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+        with:
+          fetch-depth: 0
+      - name: Prepare package
+        run: |
+          cargo package -p tfhe-versionable-derive
+      - uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08 # v4.6.0
+        with:
+          name: crate-tfhe-versionable-derive
+          path: target/package/*.crate
+      - name: generate hash
+        id: hash
+        run: cd target/package && echo "hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
+
+  provenance-derive:
+    needs: [package-derive]
+    uses: slsa-framework/slsa-github-generator/.github/workflows/generator_generic_slsa3.yml@v2.0.0
+    permissions:
+      # Needed to detect the GitHub Actions environment
+      actions: read
+      # Needed to create the provenance via GitHub OIDC
+      id-token: write
+      # Needed to upload assets/artifacts
+      contents: write
+    with:
+      # SHA-256 hashes of the Crate package.
+      base64-subjects: ${{ needs.package-derive.outputs.hash }}
+
+  publish_release-derive:
+    name: Publish tfhe-versionable Release
+    needs: [verify_tag, package-derive] # for comparing hashes
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+        with:
+          fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+      - name: Download artifact
+        uses: actions/download-artifact@fa0a91b85d4f404e444e00e005971372dc801d16 # v4.1.8
+        with:
+          name: crate-tfhe-versionable-derive
+          path: target/package
+      - name: Publish crate.io package
+        env:
+          CRATES_TOKEN: ${{ secrets.CARGO_REGISTRY_TOKEN }}
+        run: |
+          cargo publish -p tfhe-versionable-derive --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
+      - name: Generate hash
+        id: published_hash
+        run: cd target/package && echo "pub_hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
+      - name: Slack notification (hashes comparison)
+        if: ${{ needs.package-derive.outputs.hash != steps.published_hash.outputs.pub_hash }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990 # v2.3.2
+        env:
+          SLACK_COLOR: failure
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "SLSA tfhe-versionable-derive - hash comparison failure: (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990 # v2.3.2
+        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-versionable-derive release finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
+  package:
+    runs-on: ubuntu-latest
+    outputs:
+      hash: ${{ steps.hash.outputs.hash }}
+    steps:
+      - name: Checkout
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+      - name: Prepare package
+        run: |
+          cargo package -p tfhe-versionable
+      - uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08 # v4.6.0
+        with:
+          name: crate-tfhe-versionable
+          path: target/package/*.crate
+      - name: generate hash
+        id: hash
+        run: cd target/package && echo "hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
+
+  provenance:
+    needs: [package]
+    uses: slsa-framework/slsa-github-generator/.github/workflows/generator_generic_slsa3.yml@v2.0.0
+    permissions:
+      # Needed to detect the GitHub Actions environment
+      actions: read
+      # Needed to create the provenance via GitHub OIDC
+      id-token: write
+      # Needed to upload assets/artifacts
+      contents: write
+    with:
+      # SHA-256 hashes of the Crate package.
+      base64-subjects: ${{ needs.package.outputs.hash }}
+
   publish_release:
     name: Publish tfhe-versionable Release
-    needs: verify_tag
+    needs: [package] # for comparing hashes
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
-
-      - name: Publish proc-macro crate
-        env:
-          CRATES_TOKEN: ${{ secrets.CARGO_REGISTRY_TOKEN }}
-          DRY_RUN: ${{ inputs.dry_run && '--dry-run' || '' }}
-        run: |
-          cargo publish -p tfhe-versionable-derive --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
-
-      - name: Publish main crate
-        if: ${{ ! inputs.dry_run }}
+      - name: Download artifact
+        uses: actions/download-artifact@fa0a91b85d4f404e444e00e005971372dc801d16 # v4.1.8
+        with:
+          name: crate-tfhe-versionable
+          path: target/package
+      - name: Publish crate.io package
         env:
           CRATES_TOKEN: ${{ secrets.CARGO_REGISTRY_TOKEN }}
         run: |
-          cargo publish -p tfhe-versionable --token ${{ env.CRATES_TOKEN }}
+          cargo publish -p tfhe-versionable --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
+
+      - name: Generate hash
+        id: published_hash
+        run: cd target/package && echo "pub_hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
+
+      - 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 # v2.3.2
+        env:
+          SLACK_COLOR: failure
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "SLSA tfhe-versionable - hash comparison failure: (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990 # v2.3.2
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/make_release_zk_pok.yml

@@ -1,4 +1,3 @@
-# Publish new release of tfhe-zk-pok on crates.io.
 name: Publish tfhe-zk-pok release
 
 on:
@@ -13,6 +12,40 @@ env:
   ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
 
 jobs:
+  package:
+      runs-on: ubuntu-latest
+      outputs:
+        hash: ${{ steps.hash.outputs.hash }}
+      steps:
+        - name: Checkout
+          uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
+          with:
+            fetch-depth: 0
+        - name: Prepare package
+          run: |
+            cargo package -p tfhe-zk-pok
+        - uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08 # v4.6.0
+          with:
+            name: crate-zk-pok
+            path: target/package/*.crate
+        - name: generate hash
+          id: hash
+          run: cd target/package && echo "hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
+  provenance:
+    if: ${{ !inputs.dry_run  }}
+    needs: [package]
+    uses: slsa-framework/slsa-github-generator/.github/workflows/generator_generic_slsa3.yml@v2.0.0
+    permissions:
+      # Needed to detect the GitHub Actions environment
+      actions: read
+      # Needed to create the provenance via GitHub OIDC
+      id-token: write
+      # Needed to upload assets/artifacts
+      contents: write
+    with:
+      # SHA-256 hashes of the Crate package.
+      base64-subjects: ${{ needs.package.outputs.hash }}
+
   verify_tag:
     uses: ./.github/workflows/verify_tagged_commit.yml
     secrets:
@@ -21,25 +54,44 @@ jobs:
 
   publish_release:
     name: Publish tfhe-zk-pok Release
-    needs: verify_tag
+    needs: [verify_tag, package] # for comparing hashes
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
-        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
         with:
           fetch-depth: 0
-
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
+      - name: Download artifact
+        uses: actions/download-artifact@fa0a91b85d4f404e444e00e005971372dc801d16 # v4.1.8
+        with:
+          name: crate-zk-pok
+          path: target/package
       - name: Publish crate.io package
         env:
           CRATES_TOKEN: ${{ secrets.CARGO_REGISTRY_TOKEN }}
           DRY_RUN: ${{ inputs.dry_run && '--dry-run' || '' }}
         run: |
           cargo publish -p tfhe-zk-pok --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
-
+      - name: Verify hash
+        id: published_hash
+        run: cd target/package && echo "pub_hash=$(sha256sum ./*.crate | base64 -w0)" >> "${GITHUB_OUTPUT}"
+      - 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 # v2.3.2
+        env:
+          SLACK_COLOR: failure
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "SLSA tfhe-zk-pok crate - hash comparison failure: (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990 # v2.3.2
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/parameters_check.yml

@@ -14,7 +14,7 @@ on:
 
 jobs:
   params-curves-security-check:
-    runs-on: large_ubuntu_16
+    runs-on: large_ubuntu_16-22.04
     steps:
       - name: Checkout tfhe-rs
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683

.github/workflows/sync_on_push.yml

@@ -16,6 +16,8 @@ jobs:
         uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          persist-credentials: 'false'
+          token: ${{ secrets.REPO_CHECKOUT_TOKEN }}
       - name: git-sync
         uses: wei/git-sync@55c6b63b4f21607da0e9877ca9b4d11a29fc6d83
         with:

.gitignore

@@ -33,4 +33,4 @@ node_modules/
 package-lock.json
 
 # Dir used for backward compatibility test data
-tfhe/tfhe-backward-compat-data/
+tests/tfhe-backward-compat-data/

.linelint.yml

@@ -1,11 +1,15 @@
 ignore:
   - .git
   - target
+  - tfhe/build
+  - venv
+  - web-test-runner
   - tfhe/benchmarks_parameters
   - tfhe/web_wasm_parallel_tests/node_modules
   - tfhe/web_wasm_parallel_tests/dist
   - keys
   - coverage
+  - utils/tfhe-lints/ui/main.stderr
 
 rules:
   # checks if file ends in a newline character

Cargo.toml

@@ -3,28 +3,29 @@ resolver = "2"
 members = [
     "tfhe",
     "tfhe-fft",
+    "tfhe-ntt",
     "tfhe-zk-pok",
     "tasks",
     "apps/trivium",
-    "concrete-csprng",
+    "tfhe-csprng",
     "backends/tfhe-cuda-backend",
     "utils/tfhe-versionable",
     "utils/tfhe-versionable-derive",
+    "tests",
 ]
 
-exclude = [
-    "tfhe/backward_compatibility_tests",
-    "utils/cargo-tfhe-lints-inner",
-    "utils/cargo-tfhe-lints"
-]
+exclude = ["tests/backward_compatibility_tests", "utils/tfhe-lints"]
 [workspace.dependencies]
-aligned-vec = { version = "0.5", default-features = false }
+aligned-vec = { version = "0.6", default-features = false }
 bytemuck = "1.14.3"
-dyn-stack = { version = "0.10", default-features = false }
+dyn-stack = { version = "0.11", default-features = false }
+itertools = "0.14"
 num-complex = "0.4"
-pulp = { version = "0.18.22", default-features = false }
+pulp = { version = "0.20", default-features = false }
+rand = "0.8"
+rayon = "1"
 serde = { version = "1.0", default-features = false }
-wasm-bindgen = ">=0.2.86,<0.2.94"
+wasm-bindgen = "0.2.100"
 
 [profile.bench]
 lto = "fat"
@@ -42,3 +43,6 @@ inherits = "dev"
 opt-level = 3
 lto = "off"
 debug-assertions = false
+
+[workspace.metadata.dylint]
+libraries = [{ path = "utils/tfhe-lints" }]

Makefile

@@ -2,7 +2,6 @@ SHELL:=$(shell /usr/bin/env which bash)
 OS:=$(shell uname)
 RS_CHECK_TOOLCHAIN:=$(shell cat toolchain.txt | tr -d '\n')
 CARGO_RS_CHECK_TOOLCHAIN:=+$(RS_CHECK_TOOLCHAIN)
-TARGET_ARCH_FEATURE:=$(shell ./scripts/get_arch_feature.sh)
 CPU_COUNT=$(shell ./scripts/cpu_count.sh)
 RS_BUILD_TOOLCHAIN:=stable
 CARGO_RS_BUILD_TOOLCHAIN:=+$(RS_BUILD_TOOLCHAIN)
@@ -18,12 +17,14 @@ FAST_TESTS?=FALSE
 FAST_BENCH?=FALSE
 NIGHTLY_TESTS?=FALSE
 BENCH_OP_FLAVOR?=DEFAULT
+BENCH_TYPE?=latency
 NODE_VERSION=22.6
 BACKWARD_COMPAT_DATA_URL=https://github.com/zama-ai/tfhe-backward-compat-data.git
-BACKWARD_COMPAT_DATA_BRANCH?=v0.4
+BACKWARD_COMPAT_DATA_BRANCH?=$(shell ./scripts/backward_compat_data_version.py)
 BACKWARD_COMPAT_DATA_PROJECT=tfhe-backward-compat-data
 BACKWARD_COMPAT_DATA_DIR=$(BACKWARD_COMPAT_DATA_PROJECT)
 TFHE_SPEC:=tfhe
+WASM_PACK_VERSION="0.13.1"
 # We are kind of hacking the cut here, the version cannot contain a quote '"'
 WASM_BINDGEN_VERSION:=$(shell grep '^wasm-bindgen[[:space:]]*=' Cargo.toml | cut -d '"' -f 2 | xargs)
 WEB_RUNNER_DIR=web-test-runner
@@ -115,8 +116,8 @@ install_wasm_bindgen_cli: install_rs_build_toolchain
 
 .PHONY: install_wasm_pack # Install wasm-pack to build JS packages
 install_wasm_pack: install_rs_build_toolchain
-	@wasm-pack --version > /dev/null 2>&1 || \
-	cargo $(CARGO_RS_BUILD_TOOLCHAIN) install --locked wasm-pack@0.13.0 || \
+	@wasm-pack --version | grep "$(WASM_PACK_VERSION)" > /dev/null 2>&1 || \
+	cargo $(CARGO_RS_BUILD_TOOLCHAIN) install --locked wasm-pack@$(WASM_PACK_VERSION) || \
 	( echo "Unable to install cargo wasm-pack, unknown error." && exit 1 )
 
 .PHONY: install_node # Install last version of NodeJS via nvm
@@ -150,10 +151,9 @@ install_tarpaulin: install_rs_build_toolchain
 	cargo $(CARGO_RS_BUILD_TOOLCHAIN) install cargo-tarpaulin --locked || \
 	( echo "Unable to install cargo tarpaulin, unknown error." && exit 1 )
 
-.PHONY: install_tfhe_lints # Install custom tfhe-rs lints
-install_tfhe_lints:
-	(cd utils/cargo-tfhe-lints-inner && cargo install --path .) && \
-	cd utils/cargo-tfhe-lints && cargo install --path .
+.PHONY: install_cargo_dylint # Install custom tfhe-rs lints
+install_cargo_dylint:
+	cargo install cargo-dylint dylint-link
 
 .PHONY: install_typos_checker # Install typos checker
 install_typos_checker: install_rs_build_toolchain
@@ -242,7 +242,8 @@ fmt_js: check_nvm_installed
 	source ~/.nvm/nvm.sh && \
 	nvm install $(NODE_VERSION) && \
 	nvm use $(NODE_VERSION) && \
-	$(MAKE) -C tfhe/web_wasm_parallel_tests fmt
+	$(MAKE) -C tfhe/web_wasm_parallel_tests fmt && \
+	$(MAKE) -C tfhe/js_on_wasm_tests fmt
 
 .PHONY: fmt_gpu # Format rust and cuda code
 fmt_gpu: install_rs_check_toolchain
@@ -271,7 +272,8 @@ check_fmt_js: check_nvm_installed
 	source ~/.nvm/nvm.sh && \
 	nvm install $(NODE_VERSION) && \
 	nvm use $(NODE_VERSION) && \
-	$(MAKE) -C tfhe/web_wasm_parallel_tests check_fmt
+	$(MAKE) -C tfhe/web_wasm_parallel_tests check_fmt && \
+	$(MAKE) -C tfhe/js_on_wasm_tests check_fmt
 
 .PHONY: check_typos # Check for typos in codebase
 check_typos: install_typos_checker
@@ -280,14 +282,14 @@ check_typos: install_typos_checker
 .PHONY: clippy_gpu # Run clippy lints on tfhe with "gpu" enabled
 clippy_gpu: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache,gpu \
+		--features=boolean,shortint,integer,internal-keycache,gpu \
 		--all-targets \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 
 .PHONY: check_gpu # Run check on tfhe with "gpu" enabled
 check_gpu: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" check \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache,gpu \
+		--features=boolean,shortint,integer,internal-keycache,gpu \
 		--all-targets \
 		-p $(TFHE_SPEC)
 
@@ -306,52 +308,51 @@ lint_workflow: check_actionlint_installed
 .PHONY: clippy_core # Run clippy lints on core_crypto with and without experimental features
 clippy_core: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=$(TARGET_ARCH_FEATURE) \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=$(TARGET_ARCH_FEATURE),experimental \
+		--features=experimental \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=$(TARGET_ARCH_FEATURE),nightly-avx512 \
+		--features=nightly-avx512 \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=$(TARGET_ARCH_FEATURE),experimental,nightly-avx512 \
+		--features=experimental,nightly-avx512 \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=$(TARGET_ARCH_FEATURE),zk-pok \
+		--features=zk-pok \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 
 .PHONY: clippy_boolean # Run clippy lints enabling the boolean features
 clippy_boolean: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=$(TARGET_ARCH_FEATURE),boolean \
+		--features=boolean \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 
 .PHONY: clippy_shortint # Run clippy lints enabling the shortint features
 clippy_shortint: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=$(TARGET_ARCH_FEATURE),shortint \
+		--features=shortint \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=$(TARGET_ARCH_FEATURE),shortint,experimental \
+		--features=shortint,experimental \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=$(TARGET_ARCH_FEATURE),zk-pok,shortint \
+		--features=zk-pok,shortint \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 
 .PHONY: clippy_integer # Run clippy lints enabling the integer features
 clippy_integer: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=$(TARGET_ARCH_FEATURE),integer \
+		--features=integer \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=$(TARGET_ARCH_FEATURE),integer,experimental \
+		--features=integer,experimental \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 
 .PHONY: clippy # Run clippy lints enabling the boolean, shortint, integer
 clippy: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy --all-targets \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer \
+		--features=boolean,shortint,integer \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 
 .PHONY: clippy_rustdoc # Run clippy lints on doctests enabling the boolean, shortint, integer and zk-pok
@@ -362,13 +363,13 @@ clippy_rustdoc: install_rs_check_toolchain
 	fi && \
 	CLIPPYFLAGS="-D warnings" RUSTDOCFLAGS="--no-run --nocapture --test-builder ./scripts/clippy_driver.sh -Z unstable-options" \
 		cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" test --doc \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,zk-pok,pbs-stats \
+		--features=boolean,shortint,integer,zk-pok,pbs-stats,strings \
 		-p $(TFHE_SPEC)
 
 .PHONY: clippy_c_api # Run clippy lints enabling the boolean, shortint and the C API
 clippy_c_api: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=$(TARGET_ARCH_FEATURE),boolean-c-api,shortint-c-api,high-level-c-api \
+		--features=boolean-c-api,shortint-c-api,high-level-c-api \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 
 .PHONY: clippy_js_wasm_api # Run clippy lints enabling the boolean, shortint, integer and the js wasm API
@@ -393,17 +394,16 @@ clippy_trivium: install_rs_check_toolchain
 .PHONY: clippy_all_targets # Run clippy lints on all targets (benches, examples, etc.)
 clippy_all_targets: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy --all-targets \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache,zk-pok \
+		--features=boolean,shortint,integer,internal-keycache,zk-pok,strings \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy --all-targets \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache,zk-pok,experimental \
+		--features=boolean,shortint,integer,internal-keycache,zk-pok,strings,experimental \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 
-.PHONY: clippy_concrete_csprng # Run clippy lints on concrete-csprng
-clippy_concrete_csprng: install_rs_check_toolchain
+.PHONY: clippy_tfhe_csprng # Run clippy lints on tfhe-csprng
+clippy_tfhe_csprng: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy --all-targets \
-		--features=$(TARGET_ARCH_FEATURE) \
-		-p concrete-csprng -- --no-deps -D warnings
+		--features=parallel,software-prng -p tfhe-csprng -- --no-deps -D warnings
 
 .PHONY: clippy_zk_pok # Run clippy lints on tfhe-zk-pok
 clippy_zk_pok: install_rs_check_toolchain
@@ -417,14 +417,19 @@ clippy_versionable: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy --all-targets \
 		-p tfhe-versionable -- --no-deps -D warnings
 
+.PHONY: clippy_tfhe_lints # Run clippy lints on tfhe-lints
+clippy_tfhe_lints: install_cargo_dylint # the toolchain is selected with toolchain.toml
+	cd utils/tfhe-lints && \
+	cargo clippy --all-targets -- --no-deps -D warnings
+
 .PHONY: clippy_all # Run all clippy targets
 clippy_all: clippy_rustdoc clippy clippy_boolean clippy_shortint clippy_integer clippy_all_targets \
-clippy_c_api clippy_js_wasm_api clippy_tasks clippy_core clippy_concrete_csprng clippy_zk_pok clippy_trivium \
-clippy_versionable
+clippy_c_api clippy_js_wasm_api clippy_tasks clippy_core clippy_tfhe_csprng clippy_zk_pok clippy_trivium \
+clippy_versionable clippy_tfhe_lints
 
 .PHONY: clippy_fast # Run main clippy targets
 clippy_fast: clippy_rustdoc clippy clippy_all_targets clippy_c_api clippy_js_wasm_api clippy_tasks \
-clippy_core clippy_concrete_csprng
+clippy_core clippy_tfhe_csprng
 
 .PHONY: clippy_cuda_backend # Run clippy lints on the tfhe-cuda-backend
 clippy_cuda_backend: install_rs_check_toolchain
@@ -436,73 +441,73 @@ check_rust_bindings_did_not_change:
 	cargo build -p tfhe-cuda-backend && "$(MAKE)" fmt_gpu && \
 	git diff --quiet HEAD -- backends/tfhe-cuda-backend/src/bindings.rs || \
 	( echo "Generated bindings have changed! Please run 'git add backends/tfhe-cuda-backend/src/bindings.rs' \
-	and commit the changes." && exit 1 ) 
+	and commit the changes." && exit 1 )
 
 
 .PHONY: tfhe_lints # Run custom tfhe-rs lints
-tfhe_lints: install_tfhe_lints
-	cd tfhe && RUSTFLAGS="$(RUSTFLAGS)" cargo tfhe-lints \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,zk-pok -- -D warnings
+tfhe_lints: install_cargo_dylint
+	RUSTFLAGS="$(RUSTFLAGS)" cargo dylint --all -p tfhe --no-deps -- \
+		--features=boolean,shortint,integer,strings,zk-pok
 
 .PHONY: build_core # Build core_crypto without experimental features
 build_core: install_rs_build_toolchain install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) build --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE) -p $(TFHE_SPEC)
+		-p $(TFHE_SPEC)
 	@if [[ "$(AVX512_SUPPORT)" == "ON" ]]; then \
 		RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) build --profile $(CARGO_PROFILE) \
-			--features=$(TARGET_ARCH_FEATURE),nightly-avx512 -p $(TFHE_SPEC); \
+			--features=nightly-avx512 -p $(TFHE_SPEC); \
 	fi
 
 .PHONY: build_core_experimental # Build core_crypto with experimental features
 build_core_experimental: install_rs_build_toolchain install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) build --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),experimental -p $(TFHE_SPEC)
+		--features=experimental -p $(TFHE_SPEC)
 	@if [[ "$(AVX512_SUPPORT)" == "ON" ]]; then \
 		RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) build --profile $(CARGO_PROFILE) \
-			--features=$(TARGET_ARCH_FEATURE),experimental,nightly-avx512 -p $(TFHE_SPEC); \
+			--features=experimental,nightly-avx512 -p $(TFHE_SPEC); \
 	fi
 
 .PHONY: build_boolean # Build with boolean enabled
 build_boolean: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) build --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),boolean -p $(TFHE_SPEC) --all-targets
+		--features=boolean -p $(TFHE_SPEC) --all-targets
 
 .PHONY: build_shortint # Build with shortint enabled
 build_shortint: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) build --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),shortint -p $(TFHE_SPEC) --all-targets
+		--features=shortint -p $(TFHE_SPEC) --all-targets
 
 .PHONY: build_integer # Build with integer enabled
 build_integer: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) build --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),integer -p $(TFHE_SPEC) --all-targets
+		--features=integer -p $(TFHE_SPEC) --all-targets
 
 .PHONY: build_tfhe_full # Build with boolean, shortint and integer enabled
 build_tfhe_full: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) build --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer -p $(TFHE_SPEC) --all-targets
+		--features=boolean,shortint,integer -p $(TFHE_SPEC) --all-targets
 
 .PHONY: build_tfhe_coverage # Build with test coverage enabled
 build_tfhe_coverage: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS) --cfg tarpaulin" cargo $(CARGO_RS_BUILD_TOOLCHAIN) build --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache -p $(TFHE_SPEC) --tests
+		--features=boolean,shortint,integer,internal-keycache -p $(TFHE_SPEC) --tests
 
 .PHONY: build_c_api # Build the C API for boolean, shortint and integer
 build_c_api: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) build --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),boolean-c-api,shortint-c-api,high-level-c-api,zk-pok \
+		--features=boolean-c-api,shortint-c-api,high-level-c-api,zk-pok \
 		-p $(TFHE_SPEC)
 
 .PHONY: build_c_api_gpu # Build the C API for boolean, shortint and integer
 build_c_api_gpu: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) build --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),boolean-c-api,shortint-c-api,high-level-c-api,zk-pok,gpu \
+		--features=boolean-c-api,shortint-c-api,high-level-c-api,zk-pok,gpu \
 		-p $(TFHE_SPEC)
 
 .PHONY: build_c_api_experimental_deterministic_fft # Build the C API for boolean, shortint and integer with experimental deterministic FFT
 build_c_api_experimental_deterministic_fft: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) build --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),boolean-c-api,shortint-c-api,high-level-c-api,zk-pok,experimental-force_fft_algo_dif4 \
+		--features=boolean-c-api,shortint-c-api,high-level-c-api,zk-pok,experimental-force_fft_algo_dif4 \
 		-p $(TFHE_SPEC)
 
 .PHONY: build_web_js_api # Build the js API targeting the web browser
@@ -516,11 +521,11 @@ build_web_js_api: install_rs_build_toolchain install_wasm_pack
 build_web_js_api_parallel: install_rs_check_toolchain install_wasm_pack
 	cd tfhe && \
 	rustup component add rust-src --toolchain $(RS_CHECK_TOOLCHAIN) && \
-	RUSTFLAGS="$(WASM_RUSTFLAGS) -C target-feature=+atomics,+bulk-memory,+mutable-globals" rustup run $(RS_CHECK_TOOLCHAIN) \
+	RUSTFLAGS="$(WASM_RUSTFLAGS) -C target-feature=+atomics,+bulk-memory" rustup run $(RS_CHECK_TOOLCHAIN) \
 		wasm-pack build --release --target=web \
 		-- --features=boolean-client-js-wasm-api,shortint-client-js-wasm-api,integer-client-js-wasm-api,parallel-wasm-api,zk-pok \
 		-Z build-std=panic_abort,std && \
-	find pkg/snippets -type f -iname workerHelpers.worker.js -exec sed -i "s|from '..\/..\/..\/';|from '..\/..\/..\/tfhe.js';|" {} \;
+	find pkg/snippets -type f -iname workerHelpers.js -exec sed -i "s|const pkg = await import('..\/..\/..');|const pkg = await import('..\/..\/..\/tfhe.js');|" {} \;
 	jq '.files += ["snippets"]' tfhe/pkg/package.json > tmp_pkg.json && mv -f tmp_pkg.json tfhe/pkg/package.json
 
 .PHONY: build_node_js_api # Build the js API targeting nodejs
@@ -530,18 +535,18 @@ build_node_js_api: install_rs_build_toolchain install_wasm_pack
 		wasm-pack build --release --target=nodejs \
 		-- --features=boolean-client-js-wasm-api,shortint-client-js-wasm-api,integer-client-js-wasm-api,zk-pok
 
-.PHONY: build_concrete_csprng # Build concrete_csprng
-build_concrete_csprng: install_rs_build_toolchain
+.PHONY: build_tfhe_csprng # Build tfhe_csprng
+build_tfhe_csprng: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) build --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE) -p concrete-csprng --all-targets
+		-p tfhe-csprng --all-targets
 
 .PHONY: test_core_crypto # Run the tests of the core_crypto module including experimental ones
 test_core_crypto: install_rs_build_toolchain install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),experimental,zk-pok -p $(TFHE_SPEC) -- core_crypto::
+		--features=experimental,zk-pok -p $(TFHE_SPEC) -- core_crypto::
 	@if [[ "$(AVX512_SUPPORT)" == "ON" ]]; then \
 		RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-			--features=$(TARGET_ARCH_FEATURE),experimental,zk-pok,nightly-avx512 -p $(TFHE_SPEC) -- core_crypto::; \
+			--features=experimental,zk-pok,nightly-avx512 -p $(TFHE_SPEC) -- core_crypto::; \
 	fi
 
 .PHONY: test_core_crypto_cov # Run the tests of the core_crypto module with code coverage
@@ -549,13 +554,13 @@ test_core_crypto_cov: install_rs_build_toolchain install_rs_check_toolchain inst
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) tarpaulin --profile $(CARGO_PROFILE) \
 		--out xml --output-dir coverage/core_crypto --line --engine llvm --timeout 500 \
 		--implicit-test-threads $(COVERAGE_EXCLUDED_FILES) \
-		--features=$(TARGET_ARCH_FEATURE),experimental,internal-keycache \
+		--features=experimental,internal-keycache \
 		-p $(TFHE_SPEC) -- core_crypto::
 	@if [[ "$(AVX512_SUPPORT)" == "ON" ]]; then \
 		RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) tarpaulin --profile $(CARGO_PROFILE) \
 			--out xml --output-dir coverage/core_crypto_avx512 --line --engine llvm --timeout 500 \
 			--implicit-test-threads $(COVERAGE_EXCLUDED_FILES) \
-			--features=$(TARGET_ARCH_FEATURE),experimental,internal-keycache,nightly-avx512 \
+			--features=experimental,internal-keycache,nightly-avx512 \
 			-p $(TFHE_SPEC) -- -Z unstable-options --report-time core_crypto::; \
 	fi
 
@@ -573,30 +578,38 @@ test_gpu: test_core_crypto_gpu test_integer_gpu test_cuda_backend
 .PHONY: test_core_crypto_gpu # Run the tests of the core_crypto module including experimental on the gpu backend
 test_core_crypto_gpu: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),gpu -p $(TFHE_SPEC) -- core_crypto::gpu::
+		--features=gpu -p $(TFHE_SPEC) -- core_crypto::gpu::
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --doc --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),gpu -p $(TFHE_SPEC) -- core_crypto::gpu::
+		--features=gpu -p $(TFHE_SPEC) -- core_crypto::gpu::
 
 .PHONY: test_integer_gpu # Run the tests of the integer module including experimental on the gpu backend
 test_integer_gpu: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),integer,gpu -p $(TFHE_SPEC) -- integer::gpu::server_key:: --test-threads=6
+		--features=integer,gpu -p $(TFHE_SPEC) -- integer::gpu::server_key:: --test-threads=6
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --doc --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),integer,gpu -p $(TFHE_SPEC) -- integer::gpu::server_key::
+		--features=integer,gpu -p $(TFHE_SPEC) -- integer::gpu::server_key::
+
+.PHONY: test_integer_long_run_gpu # Run the long run integer tests on the gpu backend
+test_integer_long_run_gpu: install_rs_check_toolchain install_cargo_nextest
+	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
+	LONG_TESTS=TRUE \
+		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_BUILD_TOOLCHAIN) \
+		--cargo-profile "$(CARGO_PROFILE)" --avx512-support "$(AVX512_SUPPORT)" \
+		--tfhe-package "$(TFHE_SPEC)" --backend "gpu"
 
 .PHONY: test_integer_compression
 test_integer_compression: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),integer -p $(TFHE_SPEC) -- integer::ciphertext::compressed_ciphertext_list::tests::
+		--features=integer -p $(TFHE_SPEC) -- integer::ciphertext::compressed_ciphertext_list::tests::
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --doc --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),integer -p $(TFHE_SPEC) -- integer::ciphertext::compress
+		--features=integer -p $(TFHE_SPEC) -- integer::ciphertext::compress
 
 .PHONY: test_integer_compression_gpu
 test_integer_compression_gpu: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),integer,gpu -p $(TFHE_SPEC) -- integer::gpu::ciphertext::compressed_ciphertext_list::tests::
+		--features=integer,gpu -p $(TFHE_SPEC) -- integer::gpu::ciphertext::compressed_ciphertext_list::tests::
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --doc --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),integer,gpu -p $(TFHE_SPEC) -- integer::gpu::ciphertext::compress
+		--features=integer,gpu -p $(TFHE_SPEC) -- integer::gpu::ciphertext::compress
 
 .PHONY: test_integer_gpu_ci # Run the tests for integer ci on gpu backend
 test_integer_gpu_ci: install_rs_check_toolchain install_cargo_nextest
@@ -655,20 +668,20 @@ test_signed_integer_multi_bit_gpu_ci: install_rs_check_toolchain install_cargo_n
 .PHONY: test_boolean # Run the tests of the boolean module
 test_boolean: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),boolean -p $(TFHE_SPEC) -- boolean::
+		--features=boolean -p $(TFHE_SPEC) -- boolean::
 
 .PHONY: test_boolean_cov # Run the tests of the boolean module with code coverage
 test_boolean_cov: install_rs_check_toolchain install_tarpaulin
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) tarpaulin --profile $(CARGO_PROFILE) \
 		--out xml --output-dir coverage/boolean --line --engine llvm --timeout 500 \
 		$(COVERAGE_EXCLUDED_FILES) \
-		--features=$(TARGET_ARCH_FEATURE),boolean,internal-keycache \
+		--features=boolean,internal-keycache \
 		-p $(TFHE_SPEC) -- -Z unstable-options --report-time boolean::
 
 .PHONY: test_c_api_rs # Run the rust tests for the C API
 test_c_api_rs: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),boolean-c-api,shortint-c-api,high-level-c-api \
+		--features=boolean-c-api,shortint-c-api,high-level-c-api \
 		-p $(TFHE_SPEC) \
 		c_api
 
@@ -700,14 +713,14 @@ test_shortint_multi_bit_ci: install_rs_build_toolchain install_cargo_nextest
 .PHONY: test_shortint # Run all the tests for shortint
 test_shortint: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),shortint,internal-keycache -p $(TFHE_SPEC) -- shortint::
+		--features=shortint,internal-keycache -p $(TFHE_SPEC) -- shortint::
 
 .PHONY: test_shortint_cov # Run the tests of the shortint module with code coverage
 test_shortint_cov: install_rs_check_toolchain install_tarpaulin
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) tarpaulin --profile $(CARGO_PROFILE) \
 		--out xml --output-dir coverage/shortint --line --engine llvm --timeout 500 \
 		$(COVERAGE_EXCLUDED_FILES) \
-		--features=$(TARGET_ARCH_FEATURE),shortint,internal-keycache \
+		--features=shortint,internal-keycache \
 		-p $(TFHE_SPEC) -- -Z unstable-options --report-time shortint::
 
 .PHONY: test_integer_ci # Run the tests for integer ci
@@ -764,20 +777,28 @@ test_signed_integer_multi_bit_ci: install_rs_check_toolchain install_cargo_nexte
 		--cargo-profile "$(CARGO_PROFILE)" --multi-bit --avx512-support "$(AVX512_SUPPORT)" \
 		--signed-only --tfhe-package "$(TFHE_SPEC)"
 
+.PHONY: test_integer_long_run # Run the long run integer tests
+test_integer_long_run: install_rs_check_toolchain install_cargo_nextest
+	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
+	LONG_TESTS=TRUE \
+		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_BUILD_TOOLCHAIN) \
+		--cargo-profile "$(CARGO_PROFILE)" --avx512-support "$(AVX512_SUPPORT)" \
+		--tfhe-package "$(TFHE_SPEC)"
+
 .PHONY: test_safe_serialization # Run the tests for safe serialization
 test_safe_serialization: install_rs_build_toolchain install_cargo_nextest
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache -p $(TFHE_SPEC) -- safe_serialization::
+		--features=boolean,shortint,integer,internal-keycache -p $(TFHE_SPEC) -- safe_serialization::
 
 .PHONY: test_zk # Run the tests for the zk module of the TFHE-rs crate
 test_zk: install_rs_build_toolchain install_cargo_nextest
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),shortint,zk-pok -p $(TFHE_SPEC) -- zk::
+		--features=shortint,zk-pok -p $(TFHE_SPEC) -- zk::
 
 .PHONY: test_integer # Run all the tests for integer
 test_integer: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache -p $(TFHE_SPEC) -- integer::
+		--features=integer,internal-keycache -p $(TFHE_SPEC) -- integer::
 
 .PHONY: test_integer_cov # Run the tests of the integer module with code coverage
 test_integer_cov: install_rs_check_toolchain install_tarpaulin
@@ -785,38 +806,38 @@ test_integer_cov: install_rs_check_toolchain install_tarpaulin
 		--out xml --output-dir coverage/integer --line --engine llvm --timeout 500 \
 		--implicit-test-threads \
 		--exclude-files $(COVERAGE_EXCLUDED_FILES) \
-		--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache \
+		--features=integer,internal-keycache \
 		-p $(TFHE_SPEC) -- -Z unstable-options --report-time integer::
 
 .PHONY: test_high_level_api # Run all the tests for high_level_api
 test_high_level_api: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache,zk-pok -p $(TFHE_SPEC) \
+		--features=boolean,shortint,integer,internal-keycache,zk-pok,strings -p $(TFHE_SPEC) \
 		-- high_level_api::
 
 test_high_level_api_gpu: install_rs_build_toolchain install_cargo_nextest
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) nextest run --cargo-profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache,gpu -p $(TFHE_SPEC) \
+		--features=integer,internal-keycache,gpu -p $(TFHE_SPEC) \
 		-E "test(/high_level_api::.*gpu.*/)"
 
 .PHONY: test_strings # Run the tests for strings ci
 test_strings: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,strings -p $(TFHE_SPEC) \
+		--features=shortint,integer,strings -p $(TFHE_SPEC) \
 		-- strings::
 
 
 .PHONY: test_user_doc # Run tests from the .md documentation
 test_user_doc: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) --doc \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache,pbs-stats,zk-pok \
+		--features=boolean,shortint,integer,internal-keycache,pbs-stats,zk-pok,strings \
 		-p $(TFHE_SPEC) \
 		-- test_user_docs::
 
 .PHONY: test_user_doc_gpu # Run tests for GPU from the .md documentation
 test_user_doc_gpu: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) --doc \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache,gpu,zk-pok -p $(TFHE_SPEC) \
+		--features=boolean,shortint,integer,internal-keycache,gpu,zk-pok -p $(TFHE_SPEC) \
 		-- test_user_docs::
 
 
@@ -824,14 +845,12 @@ test_user_doc_gpu: install_rs_build_toolchain
 .PHONY: test_regex_engine # Run tests for regex_engine example
 test_regex_engine: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--example regex_engine \
-		--features=$(TARGET_ARCH_FEATURE),integer
+		--example regex_engine --features=integer
 
 .PHONY: test_sha256_bool # Run tests for sha256_bool example
 test_sha256_bool: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--example sha256_bool \
-		--features=$(TARGET_ARCH_FEATURE),boolean
+		--example sha256_bool --features=boolean
 
 .PHONY: test_examples # Run tests for examples
 test_examples: test_sha256_bool test_regex_engine
@@ -846,10 +865,10 @@ test_kreyvium: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
 		-p tfhe-trivium -- --test-threads=1 kreyvium::
 
-.PHONY: test_concrete_csprng # Run concrete-csprng tests
-test_concrete_csprng: install_rs_build_toolchain
+.PHONY: test_tfhe_csprng # Run tfhe-csprng tests
+test_tfhe_csprng: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE) -p concrete-csprng
+		-p tfhe-csprng
 
 .PHONY: test_zk_pok # Run tfhe-zk-pok tests
 test_zk_pok: install_rs_build_toolchain
@@ -867,23 +886,28 @@ test_zk_wasm_x86_compat_ci: check_nvm_installed
 test_zk_wasm_x86_compat: install_rs_build_toolchain build_node_js_api
 	cd tfhe/tests/zk_wasm_x86_test && npm install
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		-p tfhe --test zk_wasm_x86_test --features=$(TARGET_ARCH_FEATURE),integer,zk-pok
+		-p tfhe --test zk_wasm_x86_test --features=integer,zk-pok
 
 .PHONY: test_versionable # Run tests for tfhe-versionable subcrate
 test_versionable: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
 		--all-targets -p tfhe-versionable
 
+.PHONY: test_tfhe_lints # Run test on tfhe-lints
+test_tfhe_lints: install_cargo_dylint
+	cd utils/tfhe-lints && \
+	cargo test
+
 # The backward compat data repo holds historical binary data but also rust code to generate and load them.
 # Here we use the "patch" functionality of Cargo to make sure the repo used for the data is the same as the one used for the code.
 .PHONY: test_backward_compatibility_ci
 test_backward_compatibility_ci: install_rs_build_toolchain
 	TFHE_BACKWARD_COMPAT_DATA_DIR="$(BACKWARD_COMPAT_DATA_DIR)" RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--config "patch.'$(BACKWARD_COMPAT_DATA_URL)'.$(BACKWARD_COMPAT_DATA_PROJECT).path=\"tfhe/$(BACKWARD_COMPAT_DATA_DIR)\"" \
-		--features=$(TARGET_ARCH_FEATURE),shortint,integer,zk-pok -p $(TFHE_SPEC) test_backward_compatibility -- --nocapture
+		--config "patch.'$(BACKWARD_COMPAT_DATA_URL)'.$(BACKWARD_COMPAT_DATA_PROJECT).path=\"tests/$(BACKWARD_COMPAT_DATA_DIR)\"" \
+		--features=shortint,integer,zk-pok -p tests test_backward_compatibility -- --nocapture
 
 .PHONY: test_backward_compatibility # Same as test_backward_compatibility_ci but tries to clone the data repo first if needed
-test_backward_compatibility: tfhe/$(BACKWARD_COMPAT_DATA_DIR) test_backward_compatibility_ci
+test_backward_compatibility: tests/$(BACKWARD_COMPAT_DATA_DIR) test_backward_compatibility_ci
 
 .PHONY: backward_compat_branch # Prints the required backward compatibility branch
 backward_compat_branch:
@@ -895,7 +919,7 @@ doc: install_rs_check_toolchain
 	DOCS_RS=1 \
 	RUSTDOCFLAGS="--html-in-header katex-header.html" \
 	cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" doc \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,gpu,internal-keycache,experimental,zk-pok --no-deps -p $(TFHE_SPEC)
+		--features=boolean,shortint,integer,strings,gpu,internal-keycache,experimental,zk-pok --no-deps -p $(TFHE_SPEC)
 
 .PHONY: docs # Build rust doc alias for doc
 docs: doc
@@ -906,7 +930,7 @@ lint_doc: install_rs_check_toolchain
 	DOCS_RS=1 \
 	RUSTDOCFLAGS="--html-in-header katex-header.html -Dwarnings" \
 	cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" doc \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,gpu,internal-keycache,experimental,zk-pok -p $(TFHE_SPEC) --no-deps
+		--features=boolean,shortint,integer,strings,gpu,internal-keycache,experimental,zk-pok -p $(TFHE_SPEC) --no-deps
 
 .PHONY: lint_docs # Build rust doc with linting enabled alias for lint_doc
 lint_docs: lint_doc
@@ -935,7 +959,7 @@ check_md_links: install_mlc
 .PHONY: check_compile_tests # Build tests in debug without running them
 check_compile_tests: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --no-run \
-		--features=$(TARGET_ARCH_FEATURE),experimental,boolean,shortint,integer,internal-keycache \
+		--features=experimental,boolean,shortint,integer,internal-keycache \
 		-p $(TFHE_SPEC)
 
 	@if [[ "$(OS)" == "Linux" || "$(OS)" == "Darwin" ]]; then \
@@ -946,7 +970,7 @@ check_compile_tests: install_rs_build_toolchain
 .PHONY: check_compile_tests_benches_gpu # Build tests in debug without running them
 check_compile_tests_benches_gpu: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --no-run \
-		--features=$(TARGET_ARCH_FEATURE),experimental,boolean,shortint,integer,internal-keycache,gpu \
+		--features=experimental,boolean,shortint,integer,internal-keycache,gpu \
 		-p $(TFHE_SPEC)
 	mkdir -p "$(TFHECUDA_BUILD)" && \
 		cd "$(TFHECUDA_BUILD)" && \
@@ -1015,7 +1039,7 @@ no_dbg_log:
 	@./scripts/no_dbg_calls.sh
 
 .PHONY: dieharder_csprng # Run the dieharder test suite on our CSPRNG implementation
-dieharder_csprng: install_dieharder build_concrete_csprng
+dieharder_csprng: install_dieharder build_tfhe_csprng
 	./scripts/dieharder_test.sh
 
 #
@@ -1025,79 +1049,81 @@ dieharder_csprng: install_dieharder build_concrete_csprng
 .PHONY: print_doc_bench_parameters # Print parameters used in doc benchmarks
 print_doc_bench_parameters:
 	RUSTFLAGS="" cargo run --example print_doc_bench_parameters \
-	--features=$(TARGET_ARCH_FEATURE),shortint,internal-keycache -p tfhe
+	--features=shortint,internal-keycache -p tfhe
 
 .PHONY: bench_integer # Run benchmarks for unsigned integer
 bench_integer: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) __TFHE_RS_FAST_BENCH=$(FAST_BENCH) \
+	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) __TFHE_RS_FAST_BENCH=$(FAST_BENCH) __TFHE_RS_BENCH_TYPE=$(BENCH_TYPE) \
 	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench integer-bench \
-	--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
+	--features=integer,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
 
 .PHONY: bench_signed_integer # Run benchmarks for signed integer
 bench_signed_integer: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) __TFHE_RS_FAST_BENCH=$(FAST_BENCH) \
+	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) __TFHE_RS_FAST_BENCH=$(FAST_BENCH) __TFHE_RS_BENCH_TYPE=$(BENCH_TYPE) \
 	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench integer-signed-bench \
-	--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
+	--features=integer,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
 
 .PHONY: bench_integer_gpu # Run benchmarks for integer on GPU backend
 bench_integer_gpu: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) __TFHE_RS_FAST_BENCH=$(FAST_BENCH) \
+	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) __TFHE_RS_FAST_BENCH=$(FAST_BENCH) __TFHE_RS_BENCH_TYPE=$(BENCH_TYPE) \
 	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench integer-bench \
-	--features=$(TARGET_ARCH_FEATURE),integer,gpu,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
+	--features=integer,gpu,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
 
 .PHONY: bench_integer_compression # Run benchmarks for unsigned integer compression
 bench_integer_compression: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
+	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_TYPE=$(BENCH_TYPE) \
+	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench	glwe_packing_compression-integer-bench \
-	--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
+	--features=integer,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
 
 .PHONY: bench_integer_compression_gpu
 bench_integer_compression_gpu: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
+	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_TYPE=$(BENCH_TYPE) \
+	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench	glwe_packing_compression-integer-bench \
-	--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache,gpu -p $(TFHE_SPEC) --
+	--features=integer,internal-keycache,gpu -p $(TFHE_SPEC) --
 
 .PHONY: bench_integer_multi_bit # Run benchmarks for unsigned integer using multi-bit parameters
 bench_integer_multi_bit: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_TYPE=MULTI_BIT \
+	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_PARAM_TYPE=MULTI_BIT __TFHE_RS_BENCH_TYPE=$(BENCH_TYPE) \
 	__TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) __TFHE_RS_FAST_BENCH=$(FAST_BENCH) \
 	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench integer-bench \
-	--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
+	--features=integer,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
 
 .PHONY: bench_signed_integer_multi_bit # Run benchmarks for signed integer using multi-bit parameters
 bench_signed_integer_multi_bit: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_TYPE=MULTI_BIT \
+	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_PARAM_TYPE=MULTI_BIT __TFHE_RS_BENCH_TYPE=$(BENCH_TYPE) \
 	__TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) __TFHE_RS_FAST_BENCH=$(FAST_BENCH) \
 	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench integer-signed-bench \
-	--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
+	--features=integer,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
 
 .PHONY: bench_integer_multi_bit_gpu # Run benchmarks for integer on GPU backend using multi-bit parameters
 bench_integer_multi_bit_gpu: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_TYPE=MULTI_BIT \
-	__TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) __TFHE_RS_FAST_BENCH=$(FAST_BENCH) \
+	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_PARAM_TYPE=MULTI_BIT \
+	__TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) __TFHE_RS_FAST_BENCH=$(FAST_BENCH) __TFHE_RS_BENCH_TYPE=$(BENCH_TYPE) \
 	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench integer-bench \
-	--features=$(TARGET_ARCH_FEATURE),integer,gpu,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
+	--features=integer,gpu,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
 
 .PHONY: bench_unsigned_integer_multi_bit_gpu # Run benchmarks for unsigned integer on GPU backend using multi-bit parameters
 bench_unsigned_integer_multi_bit_gpu: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_TYPE=MULTI_BIT \
-	__TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) __TFHE_RS_FAST_BENCH=$(FAST_BENCH) \
+	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_PARAM_TYPE=MULTI_BIT \
+	__TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) __TFHE_RS_FAST_BENCH=$(FAST_BENCH) __TFHE_RS_BENCH_TYPE=$(BENCH_TYPE) \
 	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench integer-bench \
-	--features=$(TARGET_ARCH_FEATURE),integer,gpu,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) -- ::unsigned
+	--features=integer,gpu,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) -- ::unsigned
 
 .PHONY: bench_integer_zk # Run benchmarks for integer encryption with ZK proofs
 bench_integer_zk: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" \
+	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_TYPE=$(BENCH_TYPE) \
 	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench zk-pke-bench \
-	--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache,zk-pok,nightly-avx512 \
+	--features=integer,internal-keycache,zk-pok,nightly-avx512 \
 	-p $(TFHE_SPEC) --
 
 .PHONY: bench_shortint # Run benchmarks for shortint
@@ -1105,58 +1131,58 @@ bench_shortint: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) \
 	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench shortint-bench \
-	--features=$(TARGET_ARCH_FEATURE),shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
+	--features=shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
 
 .PHONY: bench_shortint_oprf # Run benchmarks for shortint
 bench_shortint_oprf: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" \
 	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench oprf-shortint-bench \
-	--features=$(TARGET_ARCH_FEATURE),shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
+	--features=shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
 
 .PHONY: bench_shortint_multi_bit # Run benchmarks for shortint using multi-bit parameters
 bench_shortint_multi_bit: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_TYPE=MULTI_BIT \
+	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_PARAM_TYPE=MULTI_BIT \
 	__TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) \
 	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench shortint-bench \
-	--features=$(TARGET_ARCH_FEATURE),shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
+	--features=shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
 
 .PHONY: bench_boolean # Run benchmarks for boolean
 bench_boolean: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench boolean-bench \
-	--features=$(TARGET_ARCH_FEATURE),boolean,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
+	--features=boolean,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
 
 .PHONY: bench_pbs # Run benchmarks for PBS
 bench_pbs: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench pbs-bench \
-	--features=$(TARGET_ARCH_FEATURE),boolean,shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
+	--features=boolean,shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
 
 .PHONY: bench_pbs128 # Run benchmarks for PBS using FFT 128 bits
 bench_pbs128: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench pbs128-bench \
-	--features=$(TARGET_ARCH_FEATURE),boolean,shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
+	--features=boolean,shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
 
 .PHONY: bench_pbs_gpu # Run benchmarks for PBS on GPU backend
 bench_pbs_gpu: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_FAST_BENCH=$(FAST_BENCH) cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench pbs-bench \
-	--features=$(TARGET_ARCH_FEATURE),boolean,shortint,gpu,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
+	--features=boolean,shortint,gpu,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
 
 .PHONY: bench_ks # Run benchmarks for keyswitch
 bench_ks: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench ks-bench \
-	--features=$(TARGET_ARCH_FEATURE),boolean,shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
+	--features=boolean,shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
 
 .PHONY: bench_ks_gpu # Run benchmarks for PBS on GPU backend
 bench_ks_gpu: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench ks-bench \
-	--features=$(TARGET_ARCH_FEATURE),boolean,shortint,gpu,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
+	--features=boolean,shortint,gpu,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
 
 bench_web_js_api_parallel_chrome: browser_path = "$(WEB_RUNNER_DIR)/chrome/chrome-linux64/chrome"
 bench_web_js_api_parallel_chrome: driver_path = "$(WEB_RUNNER_DIR)/chrome/chromedriver-linux64/chromedriver"
@@ -1192,13 +1218,13 @@ bench_web_js_api_parallel_firefox_ci: setup_venv
 bench_hlapi_erc20: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench hlapi-erc20 \
-	--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache,pbs-stats,nightly-avx512 -p $(TFHE_SPEC) --
+	--features=integer,internal-keycache,pbs-stats,nightly-avx512 -p $(TFHE_SPEC) --
 
 .PHONY: bench_hlapi_erc20_gpu # Run benchmarks for ECR20 operations on GPU
 bench_hlapi_erc20_gpu: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench hlapi-erc20 \
-	--features=$(TARGET_ARCH_FEATURE),integer,gpu,internal-keycache,pbs-stats,nightly-avx512 -p $(TFHE_SPEC) --
+	--features=integer,gpu,internal-keycache,pbs-stats,nightly-avx512 -p $(TFHE_SPEC) --
 
 .PHONY: bench_tfhe_zk_pok # Run benchmarks for the tfhe_zk_pok crate
 bench_tfhe_zk_pok: install_rs_check_toolchain
@@ -1213,32 +1239,32 @@ bench_tfhe_zk_pok: install_rs_check_toolchain
 gen_key_cache: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS) --cfg tarpaulin" cargo $(CARGO_RS_BUILD_TOOLCHAIN) run --profile $(CARGO_PROFILE) \
 		--example generates_test_keys \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,experimental,internal-keycache -p $(TFHE_SPEC) \
+		--features=boolean,shortint,experimental,internal-keycache -p $(TFHE_SPEC) \
 		-- $(MULTI_BIT_ONLY) $(COVERAGE_ONLY)
 
 .PHONY: gen_key_cache_core_crypto # Run function to generate keys and cache them for core_crypto tests
 gen_key_cache_core_crypto: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --tests --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),experimental,internal-keycache -p $(TFHE_SPEC) -- --nocapture \
+		--features=experimental,internal-keycache -p $(TFHE_SPEC) -- --nocapture \
 		core_crypto::keycache::generate_keys
 
 .PHONY: measure_hlapi_compact_pk_ct_sizes # Measure sizes of public keys and ciphertext for high-level API
 measure_hlapi_compact_pk_ct_sizes: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) run --profile $(CARGO_PROFILE) \
 	--example hlapi_compact_pk_ct_sizes \
-	--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache
+	--features=integer,internal-keycache
 
 .PHONY: measure_shortint_key_sizes # Measure sizes of bootstrapping and key switching keys for shortint
 measure_shortint_key_sizes: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) run --profile $(CARGO_PROFILE) \
 	--example shortint_key_sizes \
-	--features=$(TARGET_ARCH_FEATURE),shortint,internal-keycache
+	--features=shortint,internal-keycache
 
 .PHONY: measure_boolean_key_sizes # Measure sizes of bootstrapping and key switching keys for boolean
 measure_boolean_key_sizes: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) run --profile $(CARGO_PROFILE) \
 	--example boolean_key_sizes \
-	--features=$(TARGET_ARCH_FEATURE),boolean,internal-keycache
+	--features=boolean,internal-keycache
 
 .PHONY: parse_integer_benches # Run python parser to output a csv containing integer benches data
 parse_integer_benches:
@@ -1250,20 +1276,19 @@ parse_integer_benches:
 parse_wasm_benchmarks: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) run --profile $(CARGO_PROFILE) \
 	--example wasm_benchmarks_parser \
-	--features=$(TARGET_ARCH_FEATURE),shortint,internal-keycache \
+	--features=shortint,internal-keycache \
 	-- wasm_benchmark_results.json
 
 .PHONY: write_params_to_file # Gather all crypto parameters into a file with a Sage readable format.
 write_params_to_file: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) run --profile $(CARGO_PROFILE) \
-	--example write_params_to_file \
-	--features=$(TARGET_ARCH_FEATURE),boolean,shortint,internal-keycache
+	--example write_params_to_file --features=boolean,shortint,internal-keycache
 
 .PHONY: clone_backward_compat_data # Clone the data repo needed for backward compatibility tests
 clone_backward_compat_data:
-	./scripts/clone_backward_compat_data.sh $(BACKWARD_COMPAT_DATA_URL) $(BACKWARD_COMPAT_DATA_BRANCH) tfhe/$(BACKWARD_COMPAT_DATA_DIR)
+	./scripts/clone_backward_compat_data.sh $(BACKWARD_COMPAT_DATA_URL) $(BACKWARD_COMPAT_DATA_BRANCH) tests/$(BACKWARD_COMPAT_DATA_DIR)
 
-tfhe/$(BACKWARD_COMPAT_DATA_DIR): clone_backward_compat_data
+tests/$(BACKWARD_COMPAT_DATA_DIR): clone_backward_compat_data
 
 #
 # Real use case examples
@@ -1272,26 +1297,24 @@ tfhe/$(BACKWARD_COMPAT_DATA_DIR): clone_backward_compat_data
 .PHONY: regex_engine # Run regex_engine example
 regex_engine: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) run --profile $(CARGO_PROFILE) \
-	--example regex_engine \
-	--features=$(TARGET_ARCH_FEATURE),integer \
+	--example regex_engine --features=integer \
 	-- $(REGEX_STRING) $(REGEX_PATTERN)
 
 .PHONY: dark_market # Run dark market example
 dark_market: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) run --profile $(CARGO_PROFILE) \
 	--example dark_market \
-	--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache \
+	--features=integer,internal-keycache \
 	-- fhe-modified fhe-parallel plain fhe
 
 .PHONY: sha256_bool # Run sha256_bool example
 sha256_bool: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) run --profile $(CARGO_PROFILE) \
-	--example sha256_bool \
-	--features=$(TARGET_ARCH_FEATURE),boolean
+	--example sha256_bool --features=boolean
 
 .PHONY: pcc # pcc stands for pre commit checks (except GPU)
 pcc: no_tfhe_typo no_dbg_log check_fmt check_typos lint_doc check_md_docs_are_tested check_intra_md_links \
-clippy_all tfhe_lints check_compile_tests
+clippy_all check_compile_tests test_tfhe_lints tfhe_lints
 
 .PHONY: pcc_gpu # pcc stands for pre commit checks for GPU compilation
 pcc_gpu: clippy_gpu clippy_cuda_backend check_compile_tests_benches_gpu check_rust_bindings_did_not_change
@@ -1301,7 +1324,7 @@ fpcc: no_tfhe_typo no_dbg_log check_fmt check_typos lint_doc check_md_docs_are_t
 check_compile_tests
 
 .PHONY: conformance # Automatically fix problems that can be fixed
-conformance: fix_newline fmt
+conformance: fix_newline fmt fmt_js
 
 #=============================== FFT Section ==================================
 .PHONY: doc_fft # Build rust doc for tfhe-fft
@@ -1341,7 +1364,7 @@ build_fft: install_rs_build_toolchain
 		--features=fft128
 
 .PHONY: build_fft_no_std
-buildfft__no_std: install_rs_build_toolchain
+build_fft_no_std: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) build --release -p tfhe-fft \
 		--no-default-features
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) build --release -p tfhe-fft \
@@ -1373,7 +1396,7 @@ test_fft_nightly: install_rs_check_toolchain
 .PHONY: test_fft_no_std
 test_fft_no_std: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --release -p tfhe-fft \
-		--no-default-features 
+		--no-default-features
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --release -p tfhe-fft \
 		--no-default-features \
 		--features=fft128
@@ -1413,6 +1436,79 @@ bench_fft: install_rs_check_toolchain
 		--features=fft128
 #============================End FFT Section ==================================
 
+#=============================== NTT Section ==================================
+.PHONY: doc_ntt # Build rust doc for tfhe-ntt
+doc_ntt: install_rs_check_toolchain
+	@# Even though we are not in docs.rs, this allows to "just" build the doc
+	DOCS_RS=1 \
+	RUSTDOCFLAGS="--html-in-header katex-header.html" \
+	cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" doc \
+		--all-features --no-deps -p tfhe-ntt
+
+.PHONY: docs_ntt # Build rust doc tfhe-ntt, alias for doc
+docs_ntt: doc_ntt
+
+.PHONY: lint_doc_ntt # Build rust doc for tfhe-ntt with linting enabled
+lint_doc_ntt: install_rs_check_toolchain
+	@# Even though we are not in docs.rs, this allows to "just" build the doc
+	DOCS_RS=1 \
+	RUSTDOCFLAGS="--html-in-header katex-header.html -Dwarnings" \
+	cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" doc \
+		--all-features --no-deps -p tfhe-ntt
+
+.PHONY: lint_docs_ntt # Build rust doc for tfhe-ntt with linting enabled, alias for lint_doc
+lint_docs_ntt: lint_doc_ntt
+
+.PHONY: clippy_ntt # Run clippy lints on tfhe-ntt
+clippy_ntt: install_rs_check_toolchain
+	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy --all-targets \
+		--all-features -p tfhe-ntt -- --no-deps -D warnings
+
+.PHONY: pcc_ntt # pcc stands for pre commit checks
+pcc_ntt: check_fmt lint_doc_ntt clippy_ntt
+
+.PHONY: build_ntt
+build_ntt: install_rs_build_toolchain
+	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) build --release -p tfhe-ntt
+
+.PHONY: build_ntt_no_std
+build_ntt_no_std: install_rs_build_toolchain
+	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) build --release -p tfhe-ntt \
+		--no-default-features
+
+##### Tests #####
+
+.PHONY: test_ntt
+test_ntt: install_rs_build_toolchain
+	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --release -p tfhe-ntt
+
+.PHONY: test_ntt_nightly
+test_ntt_nightly: install_rs_check_toolchain
+	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) test --release -p tfhe-ntt \
+		--features=nightly
+
+.PHONY: test_ntt_no_std
+test_ntt_no_std: install_rs_build_toolchain
+	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --release -p tfhe-ntt \
+		--no-default-features
+
+.PHONY: test_ntt_no_std_nightly
+test_ntt_no_std_nightly: install_rs_check_toolchain
+	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) test --release -p tfhe-ntt \
+		--no-default-features \
+		--features=nightly
+
+.PHONY: test_ntt_all
+test_ntt_all: test_ntt test_ntt_no_std test_ntt_nightly test_ntt_no_std_nightly
+
+##### Bench #####
+
+.PHONY: bench_ntt # Run NTT benchmarks
+bench_ntt: install_rs_check_toolchain
+	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" bench --bench ntt -p tfhe-ntt \
+		--features=nightly
+#============================End NTT Section ==================================
+
 .PHONY: help # Generate list of targets with descriptions
 help:
 	@grep '^\.PHONY: .* #' Makefile | sed 's/\.PHONY: \(.*\) # \(.*\)/\1\t\2/' | expand -t30 | sort

README.md

@@ -70,22 +70,8 @@ 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", "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"] }
+tfhe = { version = "*", features = ["boolean", "shortint", "integer"] }
 ```
 
 > [!Note]

apps/trivium/Cargo.toml

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

apps/trivium/README.md

@@ -95,7 +95,7 @@ fn main() {
 			val >>= 1;
 		}
 	}
-	
+
 	let output_0_63    = "F4CD954A717F26A7D6930830C4E7CF0819F80E03F25F342C64ADC66ABA7F8A8E6EAA49F23632AE3CD41A7BD290A0132F81C6D4043B6E397D7388F3A03B5FE358".to_string();
 
 	let cipher_key = key.map(|x| FheBool::encrypt(x, &client_key));
@@ -129,24 +129,36 @@ Other sizes than 64 bit are expected to be available in the future.
 
 # FHE shortint Trivium implementation
 
-The same implementation is also available for generic Ciphertexts representing bits (meant to be used with parameters `PARAM_MESSAGE_1_CARRY_1_KS_PBS`). It uses a lower level API 
-of tfhe-rs, so the syntax is a little bit different. It also implements the `TransCiphering` trait. For optimization purposes, it does not internally run on the same 
-cryptographic parameters as the high level API of tfhe-rs. As such, it requires the usage of a casting key, to switch from one parameter space to another, which makes 
+The same implementation is also available for generic Ciphertexts representing bits (meant to be used with parameters `V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64`).
+It uses a lower level API of tfhe-rs, so the syntax is a little bit different. It also implements the `TransCiphering` trait. For optimization purposes, it does not internally run
+on the same cryptographic parameters as the high level API of tfhe-rs. As such, it requires the usage of a casting key, to switch from one parameter space to another, which makes
 its setup a little more intricate.
 
 Example code:
 ```rust
 use tfhe::shortint::prelude::*;
-use tfhe::shortint::CastingKey;
+use tfhe::shortint::parameters::{
+    V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64,
+    V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_GAUSSIAN_2M64,
+};
 use tfhe::{ConfigBuilder, generate_keys, FheUint64};
 use tfhe::prelude::*;
 use tfhe_trivium::TriviumStreamShortint;
 
 fn test_shortint() {
-	let config = ConfigBuilder::default().build();
+	let config = ConfigBuilder::default()
+        .use_custom_parameters(V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_GAUSSIAN_2M64)
+        .build();
 	let (hl_client_key, hl_server_key) = generate_keys(config);
-	let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
-	let ksk = CastingKey::new((&client_key, &server_key), (&hl_client_key, &hl_server_key));
+    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 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,
+    );
 
 	let key_string = "0053A6F94C9FF24598EB".to_string();
 	let mut key = [0; 80];

apps/trivium/benches/kreyvium_shortint.rs

@@ -1,23 +1,28 @@
 use criterion::Criterion;
 use tfhe::prelude::*;
-use tfhe::shortint::parameters::PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64;
+use tfhe::shortint::parameters::{
+    V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64,
+    V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_GAUSSIAN_2M64,
+};
 use tfhe::shortint::prelude::*;
 use tfhe::{generate_keys, ConfigBuilder, FheUint64};
 use tfhe_trivium::{KreyviumStreamShortint, TransCiphering};
 
 pub fn kreyvium_shortint_warmup(c: &mut Criterion) {
-    let config = ConfigBuilder::default().build();
+    let config = ConfigBuilder::default()
+        .use_custom_parameters(V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_GAUSSIAN_2M64)
+        .build();
     let (hl_client_key, hl_server_key) = generate_keys(config);
     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(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB000000000000".to_string();
@@ -57,18 +62,20 @@ pub fn kreyvium_shortint_warmup(c: &mut Criterion) {
 }
 
 pub fn kreyvium_shortint_gen(c: &mut Criterion) {
-    let config = ConfigBuilder::default().build();
+    let config = ConfigBuilder::default()
+        .use_custom_parameters(V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_GAUSSIAN_2M64)
+        .build();
     let (hl_client_key, hl_server_key) = generate_keys(config);
     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(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB000000000000".to_string();
@@ -103,18 +110,20 @@ pub fn kreyvium_shortint_gen(c: &mut Criterion) {
 }
 
 pub fn kreyvium_shortint_trans(c: &mut Criterion) {
-    let config = ConfigBuilder::default().build();
+    let config = ConfigBuilder::default()
+        .use_custom_parameters(V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_GAUSSIAN_2M64)
+        .build();
     let (hl_client_key, hl_server_key) = generate_keys(config);
     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(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB000000000000".to_string();

apps/trivium/benches/trivium_shortint.rs

@@ -1,23 +1,28 @@
 use criterion::Criterion;
 use tfhe::prelude::*;
-use tfhe::shortint::parameters::PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64;
+use tfhe::shortint::parameters::{
+    V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64,
+    V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_GAUSSIAN_2M64,
+};
 use tfhe::shortint::prelude::*;
 use tfhe::{generate_keys, ConfigBuilder, FheUint64};
 use tfhe_trivium::{TransCiphering, TriviumStreamShortint};
 
 pub fn trivium_shortint_warmup(c: &mut Criterion) {
-    let config = ConfigBuilder::default().build();
+    let config = ConfigBuilder::default()
+        .use_custom_parameters(V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_GAUSSIAN_2M64)
+        .build();
     let (hl_client_key, hl_server_key) = generate_keys(config);
     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(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB".to_string();
@@ -57,18 +62,20 @@ pub fn trivium_shortint_warmup(c: &mut Criterion) {
 }
 
 pub fn trivium_shortint_gen(c: &mut Criterion) {
-    let config = ConfigBuilder::default().build();
+    let config = ConfigBuilder::default()
+        .use_custom_parameters(V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_GAUSSIAN_2M64)
+        .build();
     let (hl_client_key, hl_server_key) = generate_keys(config);
     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(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB".to_string();
@@ -103,18 +110,20 @@ pub fn trivium_shortint_gen(c: &mut Criterion) {
 }
 
 pub fn trivium_shortint_trans(c: &mut Criterion) {
-    let config = ConfigBuilder::default().build();
+    let config = ConfigBuilder::default()
+        .use_custom_parameters(V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_GAUSSIAN_2M64)
+        .build();
     let (hl_client_key, hl_server_key) = generate_keys(config);
     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(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB".to_string();

apps/trivium/src/kreyvium/test.rs

@@ -1,6 +1,9 @@
 use crate::{KreyviumStream, KreyviumStreamByte, KreyviumStreamShortint, TransCiphering};
 use tfhe::prelude::*;
-use tfhe::shortint::parameters::PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64;
+use tfhe::shortint::parameters::{
+    V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64,
+    V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_GAUSSIAN_2M64,
+};
 use tfhe::{generate_keys, ConfigBuilder, FheBool, FheUint64, FheUint8};
 // Values for these tests come from the github repo renaud1239/Kreyvium,
 // commit fd6828f68711276c25f55e605935028f5e843f43
@@ -216,18 +219,20 @@ use tfhe::shortint::prelude::*;
 
 #[test]
 fn kreyvium_test_shortint_long() {
-    let config = ConfigBuilder::default().build();
+    let config = ConfigBuilder::default()
+        .use_custom_parameters(V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_GAUSSIAN_2M64)
+        .build();
     let (hl_client_key, hl_server_key) = generate_keys(config);
     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(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        V0_11_PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
 
     let key_string = "0053A6F94C9FF24598EB000000000000".to_string();

apps/trivium/src/trivium/test.rs

@@ -1,6 +1,9 @@
 use crate::{TransCiphering, TriviumStream, TriviumStreamByte, TriviumStreamShortint};
 use tfhe::prelude::*;
-use tfhe::shortint::parameters::PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64;
+use tfhe::shortint::parameters::{
+    V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64,
+    V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_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
@@ -352,18 +355,20 @@ use tfhe::shortint::prelude::*;
 
 #[test]
 fn trivium_test_shortint_long() {
-    let config = ConfigBuilder::default().build();
+    let config = ConfigBuilder::default()
+        .use_custom_parameters(V0_11_PARAM_MESSAGE_2_CARRY_2_PBS_KS_GAUSSIAN_2M64)
+        .build();
     let (hl_client_key, hl_server_key) = generate_keys(config);
     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(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
+        gen_keys(V0_11_PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
         (&underlying_ck, &underlying_sk),
-        PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
+        V0_11_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 = "0.6.0"
+version = "0.7.0"
 edition = "2021"
 authors = ["Zama team"]
 license = "BSD-3-Clause-Clear"
@@ -14,4 +14,4 @@ keywords = ["fully", "homomorphic", "encryption", "fhe", "cryptography"]
 [build-dependencies]
 cmake = { version = "0.1" }
 pkg-config = { version = "0.3" }
-bindgen = "0.70.1"
+bindgen = "0.71"

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

@@ -27,12 +27,25 @@ inline void cuda_error(cudaError_t code, const char *file, int line) {
     std::abort();                                                              \
   }
 
+void cuda_set_device(uint32_t gpu_index);
+
+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);

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

@@ -38,6 +38,7 @@ 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);
     }
@@ -64,7 +65,7 @@ template <typename Torus> struct int_decompression {
   Torus *tmp_extracted_lwe;
   uint32_t *tmp_indexes_array;
 
-  int_radix_lut<Torus> *carry_extract_lut;
+  int_radix_lut<Torus> *decompression_rescale_lut;
 
   int_decompression(cudaStream_t const *streams, uint32_t const *gpu_indexes,
                     uint32_t gpu_count, int_radix_params encryption_params,
@@ -83,7 +84,7 @@ template <typename Torus> struct int_decompression {
       Torus lwe_accumulator_size = (compression_params.glwe_dimension *
                                         compression_params.polynomial_size +
                                     1);
-      carry_extract_lut = new int_radix_lut<Torus>(
+      decompression_rescale_lut = new int_radix_lut<Torus>(
           streams, gpu_indexes, gpu_count, encryption_params, 1,
           num_radix_blocks, allocate_gpu_memory);
 
@@ -96,19 +97,30 @@ template <typename Torus> struct int_decompression {
           num_radix_blocks * lwe_accumulator_size * sizeof(Torus), streams[0],
           gpu_indexes[0]);
 
-      // Carry extract LUT
-      auto carry_extract_f = [encryption_params](Torus x) -> Torus {
-        return x / encryption_params.message_modulus;
+      // 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;
       };
 
-      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,
-          encryption_params.message_modulus, encryption_params.carry_modulus,
-          carry_extract_f);
+      auto effective_compression_message_modulus =
+          encryption_params.carry_modulus;
+      auto effective_compression_carry_modulus = 1;
 
-      carry_extract_lut->broadcast_lut(streams, gpu_indexes, gpu_indexes[0]);
+      generate_device_accumulator_with_encoding<Torus>(
+          streams[0], gpu_indexes[0], decompression_rescale_lut->get_lut(0, 0),
+          decompression_rescale_lut->get_degree(0),
+          decompression_rescale_lut->get_max_degree(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);
+
+      decompression_rescale_lut->broadcast_lut(streams, gpu_indexes, 0);
     }
   }
   void release(cudaStream_t const *streams, uint32_t const *gpu_indexes,
@@ -117,8 +129,8 @@ template <typename Torus> struct int_decompression {
     cuda_drop_async(tmp_extracted_lwe, streams[0], gpu_indexes[0]);
     cuda_drop_async(tmp_indexes_array, streams[0], gpu_indexes[0]);
 
-    carry_extract_lut->release(streams, gpu_indexes, gpu_count);
-    delete carry_extract_lut;
+    decompression_rescale_lut->release(streams, gpu_indexes, gpu_count);
+    delete decompression_rescale_lut;
   }
 };
 #endif

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

@@ -35,7 +35,18 @@ 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" {
+
+typedef struct {
+  void *ptr;
+  uint64_t *degrees;
+  uint64_t *noise_levels;
+  uint32_t num_radix_blocks;
+  uint32_t lwe_dimension;
+} CudaRadixCiphertextFFI;
+
 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,
@@ -43,14 +54,20 @@ void scratch_cuda_apply_univariate_lut_kb_64(
     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_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);
+    uint64_t lut_degree, 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, uint64_t lut_degree, bool allocate_gpu_memory);
+void cuda_apply_univariate_lut_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    CudaRadixCiphertextFFI *output_radix_lwe,
+    CudaRadixCiphertextFFI const *input_radix_lwe, int8_t *mem_ptr,
+    void *const *ksks, void *const *bsks);
 
 void cleanup_cuda_apply_univariate_lut_kb_64(void *const *streams,
                                              uint32_t const *gpu_indexes,
@@ -64,13 +81,15 @@ void scratch_cuda_apply_bivariate_lut_kb_64(
     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);
+    uint64_t lut_degree, 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);
+    CudaRadixCiphertextFFI *output_radix_lwe,
+    CudaRadixCiphertextFFI const *input_radix_lwe_1,
+    CudaRadixCiphertextFFI const *input_radix_lwe_2, int8_t *mem_ptr,
+    void *const *ksks, void *const *bsks, uint32_t num_radix_blocks,
+    uint32_t shift);
 
 void cleanup_cuda_apply_bivariate_lut_kb_64(void *const *streams,
                                             uint32_t const *gpu_indexes,
@@ -79,9 +98,10 @@ void cleanup_cuda_apply_bivariate_lut_kb_64(void *const *streams,
 
 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);
+    CudaRadixCiphertextFFI *output_radix_lwe,
+    CudaRadixCiphertextFFI const *input_radix_lwe, int8_t *mem_ptr,
+    void *const *ksks, void *const *bsks, 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,
@@ -103,17 +123,19 @@ void cleanup_cuda_full_propagation(void *const *streams,
 
 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, 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);
+    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,
-    void const *radix_lwe_right, void *const *bsks, void *const *ksks,
-    int8_t *mem_ptr, uint32_t polynomial_size, uint32_t num_blocks);
+    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,
@@ -121,15 +143,14 @@ void cleanup_cuda_integer_mult(void *const *streams,
 
 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,
+    CudaRadixCiphertextFFI *lwe_array_out,
+    CudaRadixCiphertextFFI const *lwe_array_in, 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);
+    CudaRadixCiphertextFFI *lwe_array, void const *scalar_input,
+    uint32_t num_scalars, 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,
@@ -142,8 +163,8 @@ void scratch_cuda_integer_radix_logical_scalar_shift_kb_64(
 
 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);
+    CudaRadixCiphertextFFI *lwe_array, uint32_t shift, int8_t *mem_ptr,
+    void *const *bsks, void *const *ksks);
 
 void scratch_cuda_integer_radix_arithmetic_scalar_shift_kb_64(
     void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
@@ -178,8 +199,8 @@ void scratch_cuda_integer_radix_shift_and_rotate_kb_64(
 
 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);
+    CudaRadixCiphertextFFI *lwe_array, CudaRadixCiphertextFFI const *lwe_shift,
+    int8_t *mem_ptr, void *const *bsks, void *const *ksks);
 
 void cleanup_cuda_integer_radix_shift_and_rotate(void *const *streams,
                                                  uint32_t const *gpu_indexes,
@@ -222,15 +243,17 @@ void scratch_cuda_integer_radix_bitop_kb_64(
 
 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);
+    CudaRadixCiphertextFFI *lwe_array_out,
+    CudaRadixCiphertextFFI const *lwe_array_1,
+    CudaRadixCiphertextFFI const *lwe_array_2, int8_t *mem_ptr,
+    void *const *bsks, void *const *ksks);
 
 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,
+    CudaRadixCiphertextFFI *lwe_array_out,
+    CudaRadixCiphertextFFI 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 *const *ksks);
 
 void cleanup_cuda_integer_bitop(void *const *streams,
                                 uint32_t const *gpu_indexes, uint32_t gpu_count,
@@ -247,9 +270,11 @@ void scratch_cuda_integer_radix_cmux_kb_64(
 
 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);
+    CudaRadixCiphertextFFI *lwe_array_out,
+    CudaRadixCiphertextFFI const *lwe_condition,
+    CudaRadixCiphertextFFI const *lwe_array_true,
+    CudaRadixCiphertextFFI const *lwe_array_false, int8_t *mem_ptr,
+    void *const *bsks, void *const *ksks);
 
 void cleanup_cuda_integer_radix_cmux(void *const *streams,
                                      uint32_t const *gpu_indexes,
@@ -266,8 +291,8 @@ void scratch_cuda_integer_radix_scalar_rotate_kb_64(
 
 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);
+    CudaRadixCiphertextFFI *lwe_array, uint32_t n, int8_t *mem_ptr,
+    void *const *bsks, void *const *ksks);
 
 void cleanup_cuda_integer_radix_scalar_rotate(void *const *streams,
                                               uint32_t const *gpu_indexes,
@@ -280,23 +305,62 @@ void scratch_cuda_propagate_single_carry_kb_64_inplace(
     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);
+    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, int8_t *mem_ptr, void *const *bsks,
-    void *const *ksks, uint32_t num_blocks);
+    CudaRadixCiphertextFFI *lwe_array, CudaRadixCiphertextFFI *carry_out,
+    const CudaRadixCiphertextFFI *carry_in, int8_t *mem_ptr, void *const *bsks,
+    void *const *ksks, uint32_t requested_flag, uint32_t uses_carry);
 
-void cuda_propagate_single_carry_get_input_carries_kb_64_inplace(
+void cuda_add_and_propagate_single_carry_kb_64_inplace(
     void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lwe_array, void *carry_out, void *input_carries, int8_t *mem_ptr,
-    void *const *bsks, void *const *ksks, uint32_t num_blocks);
+    CudaRadixCiphertextFFI *lhs_array, const CudaRadixCiphertextFFI *rhs_array,
+    CudaRadixCiphertextFFI *carry_out, const CudaRadixCiphertextFFI *carry_in,
+    int8_t *mem_ptr, void *const *bsks, void *const *ksks,
+    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,
@@ -316,25 +380,6 @@ 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_radix_overflowing_sub_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, bool allocate_gpu_memory);
-
-void cuda_integer_radix_overflowing_sub_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *radix_lwe_out, void *radix_lwe_overflowed, void const *radix_lwe_left,
-    void const *radix_lwe_right, int8_t *mem_ptr, void *const *bsks,
-    void *const *ksks, uint32_t num_blocks_in_radix);
-
-void cleanup_cuda_integer_radix_overflowing_sub(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,
@@ -357,42 +402,23 @@ void cleanup_cuda_integer_radix_scalar_mul(void *const *streams,
 
 void scratch_cuda_integer_div_rem_radix_ciphertext_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, bool allocate_gpu_memory);
+    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,
-    int8_t *mem_ptr, void *const *bsks, void *const *ksks,
+    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_signed_overflowing_add_or_sub_radix_ciphertext_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, int8_t signed_operation,
-    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
-    bool allocate_gpu_memory);
-
-void cuda_signed_overflowing_add_or_sub_radix_ciphertext_kb_64(
-    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
-    void *lhs, void const *rhs, void *overflowed, int8_t signed_operation,
-    int8_t *mem_ptr, void *const *bsks, void *const *ksks,
-    uint32_t num_blocks_in_radix);
-
-void cleanup_signed_overflowing_add_or_sub(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,
@@ -400,12 +426,13 @@ void scratch_cuda_integer_compute_prefix_sum_hillis_steele_64(
     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);
+    uint64_t lut_degree, 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);
+    CudaRadixCiphertextFFI *output_radix_lwe,
+    CudaRadixCiphertextFFI *generates_or_propagates, int8_t *mem_ptr,
+    void *const *ksks, void *const *bsks, uint32_t num_blocks);
 
 void cleanup_cuda_integer_compute_prefix_sum_hillis_steele_64(
     void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
@@ -413,9 +440,8 @@ void cleanup_cuda_integer_compute_prefix_sum_hillis_steele_64(
 
 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);
+                                            uint32_t gpu_count,
+                                            CudaRadixCiphertextFFI *lwe_array);
 
 void scratch_cuda_integer_abs_inplace_radix_ciphertext_kb_64(
     void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
@@ -428,13 +454,49 @@ void scratch_cuda_integer_abs_inplace_radix_ciphertext_kb_64(
 
 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);
+    CudaRadixCiphertextFFI *ct, int8_t *mem_ptr, bool is_signed,
+    void *const *bsks, void *const *ksks);
 
 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/integer/radix_ciphertext.h

@@ -0,0 +1,8 @@
+#ifndef CUDA_RADIX_CIPHERTEXT_H
+#define CUDA_RADIX_CIPHERTEXT_H
+
+void release_radix_ciphertext(cudaStream_t const stream,
+                              uint32_t const gpu_index,
+                              CudaRadixCiphertextFFI *data);
+
+#endif

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

@@ -21,8 +21,8 @@ void cuda_keyswitch_lwe_ciphertext_vector_64(
 
 void scratch_packing_keyswitch_lwe_list_to_glwe_64(
     void *stream, uint32_t gpu_index, int8_t **fp_ks_buffer,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t num_lwes,
-    bool allocate_gpu_memory);
+    uint32_t lwe_dimension, 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,

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

@@ -1,48 +1,52 @@
 #ifndef CUDA_LINALG_H_
 #define CUDA_LINALG_H_
 
+#include "integer/integer.h"
 #include <stdint.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,
-                                          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,
-                                          uint32_t input_lwe_dimension,
-                                          uint32_t input_lwe_ciphertext_count);
+void cuda_negate_lwe_ciphertext_vector_32(
+    void *stream, uint32_t gpu_index, void *lwe_array_out,
+    void const *lwe_array_in, const uint32_t input_lwe_dimension,
+    const 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, const uint32_t input_lwe_dimension,
+    const 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,
-                                       uint32_t input_lwe_dimension,
-                                       uint32_t input_lwe_ciphertext_count);
+                                       CudaRadixCiphertextFFI *output,
+                                       CudaRadixCiphertextFFI const *input_1,
+                                       CudaRadixCiphertextFFI const *input_2);
 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,
-                                       uint32_t input_lwe_dimension,
-                                       uint32_t input_lwe_ciphertext_count);
+                                       CudaRadixCiphertextFFI *output,
+                                       CudaRadixCiphertextFFI const *input_1,
+                                       CudaRadixCiphertextFFI const *input_2);
 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);
+    const uint32_t input_lwe_dimension,
+    const 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);
+    const uint32_t input_lwe_dimension,
+    const 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);
+    const uint32_t input_lwe_dimension,
+    const 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);
+    const uint32_t input_lwe_dimension,
+    const uint32_t input_lwe_ciphertext_count);
+void cuda_add_lwe_ciphertext_vector_plaintext_64(
+    void *stream, uint32_t gpu_index, void *lwe_array_out,
+    void const *lwe_array_in, const uint64_t plaintext_in,
+    const uint32_t input_lwe_dimension,
+    const uint32_t input_lwe_ciphertext_count);
 }
 
 #endif // CUDA_LINALG_H_

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

@@ -5,12 +5,12 @@
 
 template <typename Torus>
 bool supports_distributed_shared_memory_on_multibit_programmable_bootstrap(
-    uint32_t polynomial_size);
+    uint32_t polynomial_size, int max_shared_memory);
 
 template <typename Torus>
 bool has_support_to_cuda_programmable_bootstrap_tbc_multi_bit(
     uint32_t num_samples, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t level_count);
+    uint32_t level_count, int max_shared_memory);
 
 #if CUDA_ARCH >= 900
 template <typename Torus>
@@ -28,7 +28,7 @@ void cuda_tbc_multi_bit_programmable_bootstrap_lwe_ciphertext_vector(
     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 lut_count, uint32_t lut_stride);
+    uint32_t num_many_lut, uint32_t lut_stride);
 #endif
 
 template <typename Torus>
@@ -46,7 +46,7 @@ void cuda_cg_multi_bit_programmable_bootstrap_lwe_ciphertext_vector(
     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 lut_count, uint32_t lut_stride);
+    uint32_t num_many_lut, uint32_t lut_stride);
 
 template <typename Torus>
 void scratch_cuda_multi_bit_programmable_bootstrap(
@@ -63,7 +63,7 @@ void cuda_multi_bit_programmable_bootstrap_lwe_ciphertext_vector(
     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 lut_count, uint32_t lut_stride);
+    uint32_t num_many_lut, uint32_t lut_stride);
 
 template <typename Torus>
 uint64_t get_buffer_size_full_sm_multibit_programmable_bootstrap_keybundle(
@@ -114,6 +114,8 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::MULTI_BIT> {
              uint32_t polynomial_size, uint32_t level_count,
              uint32_t input_lwe_ciphertext_count, uint32_t lwe_chunk_size,
              PBS_VARIANT pbs_variant, bool allocate_gpu_memory) {
+    cuda_set_device(gpu_index);
+
     this->pbs_variant = pbs_variant;
     this->lwe_chunk_size = lwe_chunk_size;
     auto max_shared_memory = cuda_get_max_shared_memory(gpu_index);

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

@@ -61,7 +61,7 @@ get_buffer_size_partial_sm_programmable_bootstrap_cg(uint32_t polynomial_size) {
 
 template <typename Torus>
 bool supports_distributed_shared_memory_on_classic_programmable_bootstrap(
-    uint32_t polynomial_size);
+    uint32_t polynomial_size, int max_shared_memory);
 
 template <typename Torus, PBS_TYPE pbs_type> struct pbs_buffer;
 
@@ -77,10 +77,10 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
              uint32_t polynomial_size, uint32_t level_count,
              uint32_t input_lwe_ciphertext_count, PBS_VARIANT pbs_variant,
              bool allocate_gpu_memory) {
-
+    cuda_set_device(gpu_index);
     this->pbs_variant = pbs_variant;
 
-    auto max_shared_memory = cuda_get_max_shared_memory(0);
+    auto max_shared_memory = cuda_get_max_shared_memory(gpu_index);
 
     if (allocate_gpu_memory) {
       switch (pbs_variant) {
@@ -157,7 +157,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
 
         bool supports_dsm =
             supports_distributed_shared_memory_on_classic_programmable_bootstrap<
-                Torus>(polynomial_size);
+                Torus>(polynomial_size, max_shared_memory);
 
         uint64_t full_sm =
             get_buffer_size_full_sm_programmable_bootstrap_tbc<Torus>(
@@ -218,8 +218,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
 template <typename Torus>
 uint64_t get_buffer_size_programmable_bootstrap_cg(
     uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
-    uint32_t input_lwe_ciphertext_count) {
-  int max_shared_memory = cuda_get_max_shared_memory(0);
+    uint32_t input_lwe_ciphertext_count, uint32_t max_shared_memory) {
   uint64_t full_sm =
       get_buffer_size_full_sm_programmable_bootstrap_cg<Torus>(polynomial_size);
   uint64_t partial_sm =
@@ -245,7 +244,8 @@ template <typename Torus>
 bool has_support_to_cuda_programmable_bootstrap_cg(uint32_t glwe_dimension,
                                                    uint32_t polynomial_size,
                                                    uint32_t level_count,
-                                                   uint32_t num_samples);
+                                                   uint32_t num_samples,
+                                                   int max_shared_memory);
 
 template <typename Torus>
 void cuda_programmable_bootstrap_cg_lwe_ciphertext_vector(
@@ -255,7 +255,7 @@ void cuda_programmable_bootstrap_cg_lwe_ciphertext_vector(
     Torus const *lwe_input_indexes, double2 const *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 lut_count,
+    uint32_t level_count, uint32_t num_samples, uint32_t num_many_lut,
     uint32_t lut_stride);
 
 template <typename Torus>
@@ -266,7 +266,7 @@ void cuda_programmable_bootstrap_lwe_ciphertext_vector(
     Torus const *lwe_input_indexes, double2 const *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 lut_count,
+    uint32_t level_count, uint32_t num_samples, uint32_t num_many_lut,
     uint32_t lut_stride);
 
 #if (CUDA_ARCH >= 900)
@@ -278,7 +278,7 @@ void cuda_programmable_bootstrap_tbc_lwe_ciphertext_vector(
     Torus const *lwe_input_indexes, double2 const *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 lut_count,
+    uint32_t level_count, uint32_t num_samples, uint32_t num_many_lut,
     uint32_t lut_stride);
 
 template <typename Torus>

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

@@ -69,7 +69,7 @@ void cuda_programmable_bootstrap_lwe_ciphertext_vector_32(
     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 lut_count, uint32_t lut_stride);
+    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,
@@ -78,7 +78,7 @@ void cuda_programmable_bootstrap_lwe_ciphertext_vector_64(
     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 lut_count, uint32_t lut_stride);
+    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);

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

@@ -8,7 +8,7 @@ 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);
+    uint32_t num_samples, int max_shared_memory);
 
 void cuda_convert_lwe_multi_bit_programmable_bootstrap_key_64(
     void *stream, uint32_t gpu_index, void *dest, void const *src,
@@ -27,7 +27,7 @@ void cuda_multi_bit_programmable_bootstrap_lwe_ciphertext_vector_64(
     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 lut_count,
+    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,

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

@@ -11,7 +11,7 @@ void cuda_convert_lwe_ciphertext_vector_to_gpu(cudaStream_t stream,
                                                uint32_t gpu_index, T *dest,
                                                T *src, uint32_t number_of_cts,
                                                uint32_t lwe_dimension) {
-  cudaSetDevice(gpu_index);
+  cuda_set_device(gpu_index);
   uint64_t size = number_of_cts * (lwe_dimension + 1) * sizeof(T);
   cuda_memcpy_async_to_gpu(dest, src, size, stream, gpu_index);
 }
@@ -21,7 +21,7 @@ void cuda_convert_lwe_ciphertext_vector_to_cpu(cudaStream_t stream,
                                                uint32_t gpu_index, T *dest,
                                                T *src, uint32_t number_of_cts,
                                                uint32_t lwe_dimension) {
-  cudaSetDevice(gpu_index);
+  cuda_set_device(gpu_index);
   uint64_t size = number_of_cts * (lwe_dimension + 1) * sizeof(T);
   cuda_memcpy_async_to_cpu(dest, src, size, stream, gpu_index);
 }
@@ -55,7 +55,7 @@ __host__ void host_sample_extract(cudaStream_t stream, uint32_t gpu_index,
                                   Torus const *glwe_array_in,
                                   uint32_t const *nth_array, uint32_t num_nths,
                                   uint32_t glwe_dimension) {
-  cudaSetDevice(gpu_index);
+  cuda_set_device(gpu_index);
 
   dim3 grid(num_nths);
   dim3 thds(params::degree / params::opt);

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

@@ -0,0 +1,356 @@
+#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);
+}
+
+// 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;
+  auto write_state_idx =
+      num_lwe * lwe_dimension + 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);
+  synchronize_threads_in_block();
+  lwe_out[write_state_idx] = state;
+}
+
+// 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;
+  auto state_idx = num_lwe * lwe_dimension + val_idx;
+
+  Torus state = buffer_in[state_idx];
+  synchronize_threads_in_block();
+
+  Torus mod_b_mask = (1ll << base_log) - 1ll;
+
+  buffer_in[val_idx] = decompose_one<Torus>(state, mod_b_mask, base_log);
+  synchronize_threads_in_block();
+  buffer_in[state_idx] = state;
+}
+
+// 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.
+//
+// This code is adapted by generalizing the 1d block-tiling
+// kernel from https://github.com/siboehm/SGEMM_CUDA
+// to any matrix dimension
+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 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;
+    }
+    synchronize_threads_in_block();
+
+    // 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;
+      }
+    }
+    synchronize_threads_in_block();
+  }
+
+  // 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
+
+  cuda_set_device(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 times two (to keep both decomposition state and decomposed
+  // intermediate value)
+  int memory_unit = glwe_accumulator_size > lwe_dimension * 2
+                        ? glwe_accumulator_size
+                        : lwe_dimension * 2;
+
+  // 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 * level_count;
+
+  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());
+
+  auto ksk_block_size = glwe_accumulator_size;
+
+  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/ggsw.cuh

@@ -57,7 +57,7 @@ void batch_fft_ggsw_vector(cudaStream_t *streams, uint32_t *gpu_indexes,
   if (gpu_count != 1)
     PANIC("GPU error (batch_fft_ggsw_vector): multi-GPU execution is not "
           "supported yet.")
-  cudaSetDevice(gpu_indexes[0]);
+  cuda_set_device(gpu_indexes[0]);
 
   int shared_memory_size = sizeof(double) * polynomial_size;
 

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

@@ -1,6 +1,8 @@
+#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.
@@ -53,15 +55,17 @@ void cuda_keyswitch_lwe_ciphertext_vector_64(
 
 void scratch_packing_keyswitch_lwe_list_to_glwe_64(
     void *stream, uint32_t gpu_index, int8_t **fp_ks_buffer,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t num_lwes,
-    bool allocate_gpu_memory) {
+    uint32_t lwe_dimension, 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,
+      static_cast<cudaStream_t>(stream), gpu_index, fp_ks_buffer, lwe_dimension,
       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,
@@ -69,7 +73,7 @@ void cuda_packing_keyswitch_lwe_list_to_glwe_64(
     uint32_t output_polynomial_size, uint32_t base_log, uint32_t level_count,
     uint32_t num_lwes) {
 
-  host_packing_keyswitch_lwe_list_to_glwe<uint64_t>(
+  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),