pbs count

.github/actionlint.yaml

@@ -3,8 +3,6 @@ self-hosted-runner:
   labels:
     - m1mac
     - 4090-desktop
-    - large_windows_16_latest
-    - large_ubuntu_16
 # 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/workflows/approve_label.yml

@@ -1,5 +1,5 @@
-# Add labels in pull request
-name: PR label manager
+# Manage approved label in pull request
+name: PR approved label manager
 
 on:
   pull_request:
@@ -21,16 +21,14 @@ jobs:
         uses: actions-ecosystem/action-remove-labels@2ce5d41b4b6aa8503e285553f75ed56e0a40bae0
         with:
           # We use a PAT to have the same user (zama-bot) for label deletion as for creation.
-          github_token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          github_token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
           labels: approved
 
       # Add label only if the review is approved and if the label doesn't already exist
       - name: Add approved label
         uses: actions-ecosystem/action-add-labels@18f1af5e3544586314bbe15c0273249c770b2daf
-        if: ${{ github.event_name == 'pull_request_review' 
-          && github.event.review.state == 'approved'
-          && !contains(fromJSON(env.LABELS), 'approved') }}
+        if: ${{ github.event_name == 'pull_request_review' && github.event.review.state == 'approved' && !contains(fromJSON(env.LABELS), 'approved') }}
         with:
           # We need to use a PAT to be able to trigger `labeled` event for the other workflow.
-          github_token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          github_token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
           labels: approved

.github/workflows/aws_tfhe_backward_compat_tests.yml

@@ -1,120 +0,0 @@
-# Run backward compatibility tests
-name: Backward compatibility 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:
-  pull_request:
-
-jobs:
-  setup-instance:
-    name: Setup instance (backward-compat-tests)
-    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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: cpu-small
-
-  backward-compat-tests:
-    name: Backward compatibility tests
-    needs: [ setup-instance ]
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: true
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
-        with:
-          toolchain: stable
-
-      - name: Install git-lfs
-        run: |
-          sudo apt update && sudo apt -y install git-lfs
-
-      - name: Use specific data branch
-        if: ${{ contains(github.event.pull_request.labels.*.name, 'data_PR') }}
-        env:
-          PR_BRANCH: ${{ github.head_ref || github.ref_name }}
-        run: |
-          echo "BACKWARD_COMPAT_DATA_BRANCH=${PR_BRANCH}" >> "${GITHUB_ENV}"
-
-      - name: Get backward compat branch
-        id: backward_compat_branch
-        run: |
-          BRANCH="$(make backward_compat_branch)"
-          echo "branch=${BRANCH}" >> "${GITHUB_OUTPUT}"
-
-      - name: Clone test data
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
-          repository: zama-ai/tfhe-backward-compat-data
-          path: tfhe/tfhe-backward-compat-data
-          lfs: 'true'
-          ref: ${{ steps.backward_compat_branch.outputs.branch }}
-
-      - name: Run backward compatibility tests
-        run: |
-          make test_backward_compatibility_ci
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Backward compatibility tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (backward-compat-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, backward-compat-tests ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (backward-compat-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_fast_tests.yml

@@ -1,4 +1,4 @@
-# Run a small subset of tests to ensure quick feedback.
+# Run a small subset of shortint and integer tests to ensure quick feedback.
 name: Fast AWS Tests on CPU
 
 env:
@@ -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:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
@@ -19,192 +18,85 @@ on:
   pull_request:
 
 jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    permissions:
-      pull-requests: write
-    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 }}
-      core_crypto_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.core_crypto_any_changed ||
-        steps.changed-files.outputs.dependencies_any_changed }}
-      boolean_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.boolean_any_changed ||
-        steps.changed-files.outputs.dependencies_any_changed }}
-      shortint_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.shortint_any_changed ||
-        steps.changed-files.outputs.dependencies_any_changed }}
-      integer_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.integer_any_changed ||
-        steps.changed-files.outputs.dependencies_any_changed }}
-      wasm_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.wasm_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 }}
-      user_docs_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.user_docs_any_changed ||
-        steps.changed-files.outputs.dependencies_any_changed }}
-      any_file_changed: ${{ env.IS_PULL_REQUEST == 'false' || steps.aggregated-changes.outputs.any_changed }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@c65cd883420fd2eb864698a825fc4162dd94482c
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            dependencies:
-              - tfhe/Cargo.toml
-              - concrete-csprng/**
-              - tfhe-zk-pok/**
-            csprng:
-              - concrete-csprng/**
-            zk_pok:
-              - tfhe-zk-pok/**
-            core_crypto:
-              - tfhe/src/core_crypto/**
-            boolean:
-              - tfhe/src/core_crypto/**
-              - tfhe/src/boolean/**
-            shortint:
-              - tfhe/src/core_crypto/**
-              - tfhe/src/shortint/**
-            integer:
-              - tfhe/src/core_crypto/**
-              - tfhe/src/shortint/**
-              - tfhe/src/integer/**
-            wasm:
-              - tfhe/src/**
-              - tfhe/js_on_wasm_tests/**
-              - tfhe/web_wasm_parallel_tests/**
-              - '!tfhe/src/c_api/**'
-              - '!tfhe/src/boolean/**'
-            high_level_api:
-              - tfhe/src/**
-              - '!tfhe/src/c_api/**'
-              - '!tfhe/src/boolean/**'
-              - '!tfhe/src/c_api/**'
-              - '!tfhe/src/js_on_wasm_api/**'
-            user_docs:
-              - tfhe/src/**
-              - '!tfhe/src/c_api/**'
-              - 'tfhe/docs/**.md'
-              - README.md
-
-      - name: Aggregate file changes
-        id: aggregated-changes
-        if: ( steps.changed-files.outputs.dependencies_any_changed == 'true' ||
-          steps.changed-files.outputs.csprng_any_changed == 'true' ||
-          steps.changed-files.outputs.zk_pok_any_changed == 'true' ||
-          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.integer_any_changed == 'true' ||
-          steps.changed-files.outputs.wasm_any_changed == 'true' ||
-          steps.changed-files.outputs.high_level_api_any_changed == 'true' ||
-          steps.changed-files.outputs.user_docs_any_changed == 'true')
-        run: |
-          echo "any_changed=true" >> "$GITHUB_OUTPUT"
-
-  setup-instance:
-    name: Setup instance (fast-tests)
-    if: github.event_name != 'pull_request' ||
-      needs.should-run.outputs.any_file_changed == 'true'
-    needs: should-run
+  setup-ec2:
+    name: Setup EC2 instance (fast-tests)
     runs-on: ubuntu-latest
     outputs:
       runner-name: ${{ steps.start-instance.outputs.label }}
+      instance-id: ${{ steps.start-instance.outputs.ec2-instance-id }}
+      aws-region: ${{ steps.start-instance.outputs.aws-region }}
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         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
 
   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 ]
+    needs: setup-ec2
     concurrency:
       group: ${{ github.workflow }}_${{ github.ref }}
       cancel-in-progress: true
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    runs-on: ${{ needs.setup-ec2.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: stable
 
       - name: Run concrete-csprng tests
-        if: needs.should-run.outputs.csprng_test == 'true'
         run: |
           make test_concrete_csprng
 
-      - name: Run tfhe-zk-pok tests
-        if: needs.should-run.outputs.zk_pok_test == 'true'
-        run: |
-          make test_zk_pok
-
       - name: Run core tests
-        if: needs.should-run.outputs.core_crypto_test == 'true'
         run: |
           AVX512_SUPPORT=ON make test_core_crypto
 
       - name: Run boolean tests
-        if: needs.should-run.outputs.boolean_test == 'true'
         run: |
           make test_boolean
 
       - name: Run user docs tests
-        if: needs.should-run.outputs.user_docs_test == 'true'
         run: |
           make test_user_doc
 
       - name: Run js on wasm API tests
-        if: needs.should-run.outputs.wasm_test == 'true'
         run: |
           make test_nodejs_wasm_api_in_docker
 
       - name: Gen Keys if required
-        if: needs.should-run.outputs.shortint_test == 'true' ||
-          needs.should-run.outputs.integer_test == 'true'
         run: |
           make gen_key_cache
 
       - name: Run shortint tests
-        if: needs.should-run.outputs.shortint_test == 'true'
         run: |
           BIG_TESTS_INSTANCE=TRUE FAST_TESTS=TRUE make test_shortint_ci
 
       - name: Run integer tests
-        if: needs.should-run.outputs.integer_test == 'true'
         run: |
           BIG_TESTS_INSTANCE=TRUE FAST_TESTS=TRUE make test_integer_ci
 
+      - name: Run shortint multi-bit tests
+        run: |
+          BIG_TESTS_INSTANCE=TRUE FAST_TESTS=TRUE make test_shortint_multi_bit_ci
+
+      - name: Run integer multi-bit tests
+        run: |
+          BIG_TESTS_INSTANCE=TRUE FAST_TESTS=TRUE make test_integer_multi_bit_ci
+
       - name: Run high-level API tests
-        if: needs.should-run.outputs.high_level_api_test == 'true'
         run: |
           make test_high_level_api
 
@@ -215,31 +107,32 @@ jobs:
       - name: Slack Notification
         if: ${{ always() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Fast AWS tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
 
-  teardown-instance:
-    name: Teardown instance (fast-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, fast-tests ]
+  teardown-ec2:
+    name: Teardown EC2 instance (fast-tests)
+    if: ${{ always() && needs.setup-ec2.result != 'skipped' }}
+    needs: [ setup-ec2, fast-tests ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         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 }}
+          region: ${{ needs.setup-ec2.outputs.aws-region }}
+          label: ${{ needs.setup-ec2.outputs.runner-name }}
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (fast-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "EC2 teardown (fast-tests) failed. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_gpu_4090_tests.yml

@@ -16,17 +16,12 @@ on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
   pull_request:
-    types: [ labeled ]
-  schedule:
-    # Nightly tests @ 1AM after each work day
-    - cron: "0 1 * * MON-FRI"
+    types: [labeled]
 
 jobs:
   cuda-tests-linux:
     name: CUDA tests (RTX 4090)
-    if: github.event_name == 'workflow_dispatch' ||
-      contains(github.event.label.name, '4090_test') ||
-      (github.event_name == 'schedule' &&  github.repository == 'zama-ai/tfhe-rs')
+    if: ${{ github.event_name == 'workflow_dispatch' || contains(github.event.label.name, '4090_test') }}
     concurrency:
       group: ${{ github.workflow }}_${{ github.ref }}
       cancel-in-progress: true
@@ -34,12 +29,10 @@ jobs:
 
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: stable
 
@@ -68,7 +61,7 @@ jobs:
           make test_high_level_api_gpu
 
       - uses: actions-ecosystem/action-remove-labels@2ce5d41b4b6aa8503e285553f75ed56e0a40bae0
-        if: ${{ always() && github.event_name == 'pull_request' }}
+        if: ${{ github.event_name == 'pull_request' }}
         with:
           labels: 4090_test
           github_token: ${{ secrets.GITHUB_TOKEN }}
@@ -76,7 +69,7 @@ jobs:
       - name: Slack Notification
         if: ${{ always() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "CUDA RTX 4090 tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_gpu_tests.yml

@@ -1,5 +1,5 @@
-# Perfom tfhe-cuda-backend post-commit checks on an AWS instance
-name: TFHE Cuda Backend - Post-commit Checks
+# Compile and test tfhe-cuda-backend on an AWS instance
+name: TFHE Cuda Backend - Full tests
 
 env:
   CARGO_TERM_COLOR: always
@@ -13,33 +13,36 @@ env:
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
 
 on:
+  # Allows you to run this workflow manually from the Actions tab as an alternative.
+  workflow_dispatch:
   pull_request:
 
 jobs:
-  setup-instance:
-    name: Setup instance (cuda-pcc)
+  setup-ec2:
+    name: Setup EC2 instance (cuda-tests)
     runs-on: ubuntu-latest
     outputs:
       runner-name: ${{ steps.start-instance.outputs.label }}
+      instance-id: ${{ steps.start-instance.outputs.ec2-instance-id }}
+      aws-region: ${{ steps.start-instance.outputs.aws-region }}
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
           slab-url: ${{ secrets.SLAB_BASE_URL }}
           job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: gpu-build
+          profile: gpu-test
 
-  cuda-pcc:
-    name: CUDA post-commit checks
-    needs: setup-instance
+  cuda-tests-linux:
+    name: CUDA tests
+    needs: setup-ec2
     concurrency:
       group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: true
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
+    runs-on: ${{ needs.setup-ec2.outputs.runner-name }}
     strategy:
       fail-fast: false
       # explicit include-based build matrix, of known valid options
@@ -53,16 +56,14 @@ jobs:
 
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: stable
 
@@ -93,34 +94,51 @@ jobs:
         run: |
           make pcc_gpu
 
+      - name: Run core crypto, integer and internal CUDA backend tests
+        run: |
+          make test_gpu
+
+      - name: Run user docs tests
+        run: |
+          make test_user_doc_gpu
+
+      - name: Test C API
+        run: |
+          make test_c_api_gpu
+
+      - name: Run High Level API Tests
+        run: |
+          make test_high_level_api_gpu
+
       - name: Slack Notification
         if: ${{ always() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         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 tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
 
-  teardown-instance:
-    name: Teardown instance (cuda-pcc)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-pcc ]
+  teardown-ec2:
+    name: Teardown EC2 instance (cuda-tests)
+    if: ${{ always() && needs.setup-ec2.result != 'skipped' }}
+    needs: [ setup-ec2, cuda-tests-linux ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         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 }}
+          region: ${{ needs.setup-ec2.outputs.aws-region }}
+          label: ${{ needs.setup-ec2.outputs.runner-name }}
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-pcc) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "EC2 teardown (cuda-tests) failed. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_integer_tests.yml

@@ -10,72 +10,53 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-  # 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
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
   pull_request:
     types: [ labeled ]
-  push:
-    branches:
-      - main
-  schedule:
-    # Nightly tests @ 3AM after each work day
-    - cron: "0 3 * * MON-FRI"
 
 jobs:
-  setup-instance:
-    name: Setup instance (unsigned-integer-tests)
-    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 == 'workflow_dispatch'
+  setup-ec2:
+    name: Setup EC2 instance (unsigned-integer-tests)
+    if: ${{ github.event_name == 'workflow_dispatch' || contains(github.event.label.name, 'approved') }}
     runs-on: ubuntu-latest
     outputs:
       runner-name: ${{ steps.start-instance.outputs.label }}
+      instance-id: ${{ steps.start-instance.outputs.ec2-instance-id }}
+      aws-region: ${{ steps.start-instance.outputs.aws-region }}
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         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
 
   unsigned-integer-tests:
     name: Unsigned integer tests
-    needs: setup-instance
+    needs: setup-ec2
     concurrency:
-      group: ${{ github.workflow }}_${{ 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 }}
+      group: ${{ github.workflow }}_${{ github.ref }}
+      cancel-in-progress: true
+    runs-on: ${{ needs.setup-ec2.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: stable
 
-      - name: Should skip big parameters set
-        if: github.event_name == 'pull_request'
-        run: |
-          echo "NO_BIG_PARAMS=TRUE" >> "${GITHUB_ENV}"
-
       - name: Gen Keys if required
         run: |
           make GEN_KEY_CACHE_MULTI_BIT_ONLY=TRUE gen_key_cache
@@ -90,36 +71,37 @@ jobs:
 
       - name: Run unsigned integer tests
         run: |
-          AVX512_SUPPORT=ON NO_BIG_PARAMS=${{ env.NO_BIG_PARAMS }} BIG_TESTS_INSTANCE=TRUE make test_unsigned_integer_ci
+          AVX512_SUPPORT=ON BIG_TESTS_INSTANCE=TRUE make test_unsigned_integer_ci
 
       - name: Slack Notification
         if: ${{ always() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Unsigned Integer tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
 
-  teardown-instance:
-    name: Teardown instance (unsigned-integer-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, unsigned-integer-tests ]
+  teardown-ec2:
+    name: Teardown EC2 instance (unsigned-integer-tests)
+    if: ${{ always() && needs.setup-ec2.result != 'skipped' }}
+    needs: [ setup-ec2, unsigned-integer-tests ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         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 }}
+          region: ${{ needs.setup-ec2.outputs.aws-region }}
+          label: ${{ needs.setup-ec2.outputs.runner-name }}
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (unsigned-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "EC2 teardown (unsigned-integer-tests) failed. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_signed_integer_tests.yml

@@ -10,72 +10,53 @@ env:
   SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
   SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-  # 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
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
   pull_request:
     types: [ labeled ]
-  push:
-    branches:
-      - main
-  schedule:
-    # Nightly tests @ 3AM after each work day
-    - cron: "0 3 * * MON-FRI"
 
 jobs:
-  setup-instance:
-    name: Setup instance (signed-integer-tests)
-    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 == 'workflow_dispatch'
+  setup-ec2:
+    name: Setup EC2 instance (signed-integer-tests)
+    if: ${{ github.event_name == 'workflow_dispatch' || contains(github.event.label.name, 'approved') }}
     runs-on: ubuntu-latest
     outputs:
       runner-name: ${{ steps.start-instance.outputs.label }}
+      instance-id: ${{ steps.start-instance.outputs.ec2-instance-id }}
+      aws-region: ${{ steps.start-instance.outputs.aws-region }}
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         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
 
   signed-integer-tests:
     name: Signed integer tests
-    needs: setup-instance
+    needs: setup-ec2
     concurrency:
-      group: ${{ github.workflow }}_${{ 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 }}
+      group: ${{ github.workflow }}_${{ github.ref }}
+      cancel-in-progress: true
+    runs-on: ${{ needs.setup-ec2.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: stable
 
-      - name: Should skip big parameters set
-        if: github.event_name == 'pull_request'
-        run: |
-          echo "NO_BIG_PARAMS=TRUE" >> "${GITHUB_ENV}"
-
       - name: Gen Keys if required
         run: |
           make GEN_KEY_CACHE_MULTI_BIT_ONLY=TRUE gen_key_cache
@@ -94,36 +75,37 @@ jobs:
 
       - name: Run signed integer tests
         run: |
-          AVX512_SUPPORT=ON NO_BIG_PARAMS=${{ env.NO_BIG_PARAMS }} BIG_TESTS_INSTANCE=TRUE make test_signed_integer_ci
+          AVX512_SUPPORT=ON BIG_TESTS_INSTANCE=TRUE make test_signed_integer_ci
 
       - name: Slack Notification
         if: ${{ always() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Signed Integer tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
 
-  teardown-instance:
-    name: Teardown instance (signed-integer-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, signed-integer-tests ]
+  teardown-ec2:
+    name: Teardown EC2 instance (signed-integer-tests)
+    if: ${{ always() && needs.setup-ec2.result != 'skipped' }}
+    needs: [ setup-ec2, signed-integer-tests ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         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 }}
+          region: ${{ needs.setup-ec2.outputs.aws-region }}
+          label: ${{ needs.setup-ec2.outputs.runner-name }}
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (signed-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "EC2 teardown (signed-integer-tests) failed. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_tests.yml

@@ -10,213 +10,91 @@ 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:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
   pull_request:
     types: [ labeled ]
-  schedule:
-    # Nightly tests @ 1AM after each work day
-    - cron: "0 1 * * MON-FRI"
 
 jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    if: github.event_name != 'schedule' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
-    permissions:
-      pull-requests: write
-    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 }}
-      core_crypto_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.core_crypto_any_changed ||
-        steps.changed-files.outputs.dependencies_any_changed }}
-      boolean_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.boolean_any_changed ||
-        steps.changed-files.outputs.dependencies_any_changed }}
-      shortint_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.shortint_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 }}
-      c_api_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.c_api_any_changed ||
-        steps.changed-files.outputs.dependencies_any_changed }}
-      examples_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.examples_any_changed ||
-        steps.changed-files.outputs.dependencies_any_changed }}
-      apps_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.apps_any_changed || steps.changed-files.outputs.dependencies_any_changed }}
-      user_docs_test: ${{ env.IS_PULL_REQUEST == 'false' ||
-        steps.changed-files.outputs.user_docs_any_changed ||
-        steps.changed-files.outputs.dependencies_any_changed }}
-      any_file_changed: ${{ env.IS_PULL_REQUEST == 'false' || steps.aggregated-changes.outputs.any_changed }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@c65cd883420fd2eb864698a825fc4162dd94482c
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            dependencies:
-              - tfhe/Cargo.toml
-              - concrete-csprng/**
-              - tfhe-zk-pok/**
-            csprng:
-              - concrete-csprng/**
-            zk_pok:
-              - tfhe-zk-pok/**
-            core_crypto:
-              - tfhe/src/core_crypto/**
-            boolean:
-              - tfhe/src/core_crypto/**
-              - tfhe/src/boolean/**
-            shortint:
-              - tfhe/src/core_crypto/**
-              - tfhe/src/shortint/**
-            high_level_api:
-              - tfhe/src/**
-              - '!tfhe/src/c_api/**'
-              - '!tfhe/src/boolean/**'
-              - '!tfhe/src/js_on_wasm_api/**'
-            c_api:
-              - tfhe/src/**
-            examples:
-              - tfhe/src/**
-              - '!tfhe/src/c_api/**'
-              - tfhe/examples/**
-            apps:
-              - tfhe/src/**
-              - '!tfhe/src/c_api/**'
-              - apps/trivium/src/**
-            user_docs:
-              - tfhe/src/**
-              - '!tfhe/src/c_api/**'
-              - 'tfhe/docs/**.md'
-              - README.md
-
-      - name: Aggregate file changes
-        id: aggregated-changes
-        if: ( steps.changed-files.outputs.dependencies_any_changed == 'true' ||
-          steps.changed-files.outputs.csprng_any_changed == 'true' ||
-          steps.changed-files.outputs.zk_pok_any_changed == 'true' ||
-          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.high_level_api_any_changed == 'true' ||
-          steps.changed-files.outputs.c_api_any_changed == 'true' ||
-          steps.changed-files.outputs.examples_any_changed == 'true' ||
-          steps.changed-files.outputs.apps_any_changed == 'true' ||
-          steps.changed-files.outputs.user_docs_any_changed == 'true')
-        run: |
-          echo "any_changed=true" >> "$GITHUB_OUTPUT"
-
-  setup-instance:
-    name: Setup instance (cpu-tests)
-    if: github.event_name != 'pull_request' ||
-      (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.any_file_changed == 'true')
-    needs: should-run
+  setup-ec2:
+    name: Setup EC2 instance (cpu-tests)
+    if: ${{ github.event_name == 'workflow_dispatch' || contains(github.event.label.name, 'approved') }}
     runs-on: ubuntu-latest
     outputs:
       runner-name: ${{ steps.start-instance.outputs.label }}
+      instance-id: ${{ steps.start-instance.outputs.ec2-instance-id }}
+      aws-region: ${{ steps.start-instance.outputs.aws-region }}
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         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: CPU tests
-    if: github.event_name != 'pull_request' ||
-      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
-    needs: [ should-run, setup-instance ]
+    needs: setup-ec2
     concurrency:
-      group: ${{ github.workflow }}_${{github.event_name}}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.ref }}
       cancel-in-progress: true
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    runs-on: ${{ needs.setup-ec2.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: stable
 
       - name: Run concrete-csprng tests
-        if: needs.should-run.outputs.csprng_test == 'true'
         run: |
           make test_concrete_csprng
 
-      - name: Run tfhe-zk-pok tests
-        if: needs.should-run.outputs.zk_pok_test == 'true'
-        run: |
-          make test_zk_pok
-
       - name: Run core tests
-        if: needs.should-run.outputs.core_crypto_test == 'true'
         run: |
           AVX512_SUPPORT=ON make test_core_crypto
 
       - name: Run boolean tests
-        if: needs.should-run.outputs.boolean_test == 'true'
         run: |
           make test_boolean
 
       - name: Run C API tests
-        if: needs.should-run.outputs.c_api_test == 'true'
         run: |
           make test_c_api
 
       - name: Run user docs tests
-        if: needs.should-run.outputs.user_docs_test == 'true'
         run: |
           make test_user_doc
 
       - name: Gen Keys if required
-        if: needs.should-run.outputs.shortint_test == 'true'
         run: |
           make gen_key_cache
 
       - name: Run shortint tests
-        if: needs.should-run.outputs.shortint_test == 'true'
         run: |
           BIG_TESTS_INSTANCE=TRUE make test_shortint_ci
 
       - name: Run high-level API tests
-        if: needs.should-run.outputs.high_level_api_test == 'true'
         run: |
           BIG_TESTS_INSTANCE=TRUE make test_high_level_api
 
       - name: Run example tests
-        if: needs.should-run.outputs.examples_test == 'true'
         run: |
           make test_examples
           make dark_market
 
       - name: Run apps tests
-        if: needs.should-run.outputs.apps_test == 'true'
         run: |
           make test_trivium
           make test_kreyvium
@@ -224,31 +102,32 @@ jobs:
       - name: Slack Notification
         if: ${{ always() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "CPU tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
 
-  teardown-instance:
-    name: Teardown instance (cpu-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cpu-tests ]
+  teardown-ec2:
+    name: Teardown EC2 instance (cpu-tests)
+    if: ${{ always() && needs.setup-ec2.result != 'skipped' }}
+    needs: [ setup-ec2, cpu-tests ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         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 }}
+          region: ${{ needs.setup-ec2.outputs.aws-region }}
+          label: ${{ needs.setup-ec2.outputs.runner-name }}
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cpu-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "EC2 teardown (cpu-tests) failed. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_wasm_tests.yml

@@ -18,90 +18,83 @@ on:
     types: [ labeled ]
 
 jobs:
-  setup-instance:
-    name: Setup instance (wasm-tests)
+  setup-ec2:
+    name: Setup EC2 instance (wasm-tests)
     if: ${{ github.event_name == 'workflow_dispatch' || contains(github.event.label.name, 'approved') }}
     runs-on: ubuntu-latest
     outputs:
       runner-name: ${{ steps.start-instance.outputs.label }}
+      instance-id: ${{ steps.start-instance.outputs.ec2-instance-id }}
+      aws-region: ${{ steps.start-instance.outputs.aws-region }}
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
           slab-url: ${{ secrets.SLAB_BASE_URL }}
           job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
           profile: cpu-small
 
   wasm-tests:
     name: WASM tests
-    needs: setup-instance
+    needs: setup-ec2
     concurrency:
       group: ${{ github.workflow }}_${{ github.ref }}
       cancel-in-progress: true
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    runs-on: ${{ needs.setup-ec2.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: stable
 
-      - name: Install Node
-        run: |
-          make install_node
-
-      - name: Run fmt checks
-        run: |
-          make check_fmt_js
-
       - name: Run js on wasm API tests
         run: |
           make test_nodejs_wasm_api_in_docker
 
       - name: Run parallel wasm tests
         run: |
-          make test_web_js_api_parallel_ci
+          make install_node
+          make ci_test_web_js_api_parallel
 
       - name: Slack Notification
         if: ${{ always() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "WASM tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
 
-  teardown-instance:
-    name: Teardown instance (wasm-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, wasm-tests ]
+  teardown-ec2:
+    name: Teardown EC2 instance (wasm-tests)
+    if: ${{ always() && needs.setup-ec2.result != 'skipped' }}
+    needs: [ setup-ec2, wasm-tests ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         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 }}
+          region: ${{ needs.setup-ec2.outputs.aws-region }}
+          label: ${{ needs.setup-ec2.outputs.runner-name }}
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (wasm-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "EC2 teardown (wasm-tests) failed. (${{ env.ACTION_RUN_URL }})"

.github/workflows/boolean_benchmark.yml

@@ -3,9 +3,30 @@ name: Boolean benchmarks
 
 on:
   workflow_dispatch:
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 1a.m.
-    - cron: '0 1 * * 6'
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
+      # This input is not used in this workflow but still mandatory since a calling workflow could
+      # use it. If a triggering command include a user_inputs field, then the triggered workflow
+      # must include this very input, otherwise the workflow won't be called.
+      # See start_full_benchmarks.yml as example.
+      user_inputs:
+        description: "Type of benchmarks to run"
+        type: string
+        default: "weekly_benchmarks"
 
 env:
   CARGO_TERM_COLOR: always
@@ -13,60 +34,36 @@ env:
   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 (boolean-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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: bench
-
-  boolean-benchmarks:
+  run-boolean-benchmarks:
     name: Execute boolean benchmarks in EC2
-    needs: setup-instance
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    continue-on-error: true
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ !cancelled() }}
     steps:
+      - name: Instance configuration used
+        run: |
+          echo "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Get benchmark date
+        run: |
+          echo "BENCH_DATE=$(date --iso-8601=seconds)" >> "${GITHUB_ENV}"
+
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         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)";
-          } >> "${GITHUB_ENV}"
-
       - name: Set up home
         # "Install rust" step require root user to have a HOME directory which is not set.
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: nightly
 
@@ -76,12 +73,14 @@ jobs:
 
       - name: Parse results
         run: |
+          COMMIT_DATE="$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})"
+          COMMIT_HASH="$(git describe --tags --dirty)"
           python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
           --database tfhe_rs \
-          --hardware "hpc7a.96xlarge" \
-          --project-version "${{ env.COMMIT_HASH }}" \
+          --hardware ${{ inputs.instance_type }} \
+          --project-version "${COMMIT_HASH}" \
           --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
+          --commit-date "${COMMIT_DATE}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
           --name-suffix avx512 \
@@ -98,17 +97,17 @@ jobs:
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
         with:
           name: ${{ github.sha }}_boolean
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -127,31 +126,11 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Boolean benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (boolean-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, boolean-benchmarks ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (boolean-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "Boolean benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/cargo_build.yml

@@ -19,11 +19,11 @@ jobs:
 
     strategy:
       matrix:
-        os: [large_ubuntu_16, macos-latest-large, large_windows_16_latest]
+        os: [ubuntu-latest, macos-latest-large, windows-latest]
       fail-fast: false
 
     steps:
-      - uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+      - uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
 
       - name: Install and run newline linter checks
         if: matrix.os == 'ubuntu-latest'

.github/workflows/check_commit.yml

@@ -10,7 +10,7 @@ jobs:
       - name: Check first line
         uses: gsactions/commit-message-checker@16fa2d5de096ae0d35626443bcd24f1e756cafee
         with:
-          pattern: '^((feat|fix|chore|refactor|style|test|docs|doc)(\([\w\-_]+\))?\!?\:) .+$'
+          pattern: '^((feat|fix|chore|refactor|style|test|docs|doc)(\(\w+\))?\:) .+$'
           flags: "gs"
           error: 'Your first line has to contain a commit type and scope like "feat(my_feature): msg".'
           excludeDescription: "true" # optional: this excludes the description body of a pull request

.github/workflows/ci_lint.yml

@@ -13,7 +13,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
 
       - name: Get actionlint
         run: |

.github/workflows/code_coverage.yml

@@ -6,58 +6,70 @@ env:
   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:
-  # Code coverage workflow is only run via workflow_dispatch event since execution duration is not stabilized yet.
+    # All the inputs are provided by Slab
+    inputs:
+      instance_id:
+        description: "AWS instance ID"
+        type: string
+      instance_image_id:
+        description: "AWS instance AMI ID"
+        type: string
+      instance_type:
+        description: "AWS instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: 'Slab request ID'
+        type: string
+      fork_repo:
+        description: 'Name of forked repo as user/repo'
+        type: string
+      fork_git_sha:
+        description: 'Git SHA to checkout from fork'
+        type: string
 
 jobs:
-  setup-instance:
-    name: Setup instance (code-coverage)
-    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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: cpu-small
-
   code-coverage:
-    name: Code coverage tests
-    needs: setup-instance
     concurrency:
-      group: ${{ github.workflow }}_${{ github.event_name }}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{ github.ref }}_${{ inputs.instance_image_id }}_${{ inputs.instance_type }}
       cancel-in-progress: true
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    timeout-minutes: 5760 # 4 days
+    runs-on: ${{ inputs.runner_name }}
+    timeout-minutes: 11520 # 8 days
     steps:
+      # Step used for log purpose.
+      - name: Instance configuration used
+        run: |
+          echo "ID: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+          echo "Fork repo: ${{ inputs.fork_repo }}"
+          echo "Fork git sha: ${{ inputs.fork_git_sha }}"
+
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          repository: ${{ inputs.fork_repo }}
+          ref: ${{ inputs.fork_git_sha }}
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: stable
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@c65cd883420fd2eb864698a825fc4162dd94482c
+        uses: tj-actions/changed-files@aa08304bd477b800d468db44fe10f6c61f7f7b11
         with:
           files_yaml: |
             tfhe:
@@ -87,7 +99,7 @@ jobs:
           make test_shortint_cov
 
       - name: Upload tfhe coverage to Codecov
-        uses: codecov/codecov-action@e28ff129e5465c2c0dcc6f003fc735cb6ae0c673
+        uses: codecov/codecov-action@54bcd8715eee62d40e33596ef5e8f0f48dbbccab
         if: steps.changed-files.outputs.tfhe_any_changed == 'true'
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
@@ -101,7 +113,7 @@ jobs:
           make test_integer_cov
 
       - name: Upload tfhe coverage to Codecov
-        uses: codecov/codecov-action@e28ff129e5465c2c0dcc6f003fc735cb6ae0c673
+        uses: codecov/codecov-action@54bcd8715eee62d40e33596ef5e8f0f48dbbccab
         if: steps.changed-files.outputs.tfhe_any_changed == 'true'
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
@@ -112,31 +124,11 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
           SLACK_MESSAGE: "Code coverage finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (code-coverage)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, code-coverage ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (code-coverage) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/core_crypto_benchmark.yml

@@ -3,6 +3,30 @@ name: Core crypto benchmarks
 
 on:
   workflow_dispatch:
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
+      # This input is not used in this workflow but still mandatory since a calling workflow could
+      # use it. If a triggering command include a user_inputs field, then the triggered workflow
+      # must include this very input, otherwise the workflow won't be called.
+      # See start_full_benchmarks.yml as example.
+      user_inputs:
+        description: "Type of benchmarks to run"
+        type: string
+        default: "weekly_benchmarks"
 
 env:
   CARGO_TERM_COLOR: always
@@ -10,93 +34,71 @@ env:
   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 (core-crypto-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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: bench
-
-  core-crypto-benchmarks:
+  run-core-crypto-benchmarks:
     name: Execute core crypto benchmarks in EC2
-    needs: setup-instance
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ !cancelled() }}
     steps:
+      - name: Instance configuration used
+        run: |
+          echo "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Get benchmark date
+        run: |
+          echo "BENCH_DATE=$(date --iso-8601=seconds)" >> "${GITHUB_ENV}"
+
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         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)";
-          } >> "${GITHUB_ENV}"
-
       - name: Set up home
         # "Install rust" step require root user to have a HOME directory which is not set.
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: nightly
 
       - name: Run benchmarks with AVX512
         run: |
           make bench_pbs
-          make bench_pbs128
           make bench_ks
 
       - name: Parse results
         run: |
+          COMMIT_DATE="$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})"
+          COMMIT_HASH="$(git describe --tags --dirty)"
           python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
           --database tfhe_rs \
-          --hardware "hpc7a.96xlarge" \
-          --project-version "${{ env.COMMIT_HASH }}" \
+          --hardware ${{ inputs.instance_type }} \
+          --project-version "${COMMIT_HASH}" \
           --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
+          --commit-date "${COMMIT_DATE}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --name-suffix avx512 \
           --walk-subdirs \
           --throughput
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
         with:
           name: ${{ github.sha }}_core_crypto
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -115,31 +117,11 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "PBS benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (core-crypto-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, core-crypto-benchmarks ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (core-crypto-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "PBS benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/core_crypto_gpu_benchmark.yml

@@ -1,45 +1,43 @@
-# Run core crypto benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
+# Run core crypto benchmarks on an AWS instance with CUDA and return parsed results to Slab CI bot.
 name: Core crypto GPU benchmarks
 
 on:
   workflow_dispatch:
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 1a.m.
-    - cron: '0 1 * * 6'
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
+      # This input is not used in this workflow but still mandatory since a calling workflow could
+      # use it. If a triggering command include a user_inputs field, then the triggered workflow
+      # must include this very input, otherwise the workflow won't be called.
+      # See start_full_benchmarks.yml as example.
+      user_inputs:
+        description: "Type of benchmarks to run"
+        type: string
+        default: "weekly_benchmarks"
 
 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 }}
-  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-core-crypto-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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        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-core-crypto-benchmarks:
-    name: Execute GPU core crypto benchmarks
-    needs: setup-instance
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+  run-core-crypto-benchmarks:
+    name: Execute GPU core crypto benchmarks in EC2
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ !cancelled() }}
     strategy:
       fail-fast: false
       # explicit include-based build matrix, of known valid options
@@ -47,43 +45,33 @@ jobs:
         include:
           - os: ubuntu-22.04
             cuda: "12.2"
-            gcc: 11
+            gcc: 9
     env:
       CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
     steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
+      - name: Instance configuration used
         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 "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Get benchmark date
+        run: |
+          echo "BENCH_DATE=$(date --iso-8601=seconds)" >> "${GITHUB_ENV}"
 
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         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)";
-          } >> "${GITHUB_ENV}"
-
       - name: Set up home
         # "Install rust" step require root user to have a HOME directory which is not set.
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: nightly
 
@@ -115,30 +103,32 @@ jobs:
 
       - name: Parse results
         run: |
+          COMMIT_DATE="$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})"
+          COMMIT_HASH="$(git describe --tags --dirty)"
           python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
           --database tfhe_rs \
-          --hardware "n3-H100x1" \
+          --hardware ${{ inputs.instance_type }} \
           --backend gpu \
-          --project-version "${{ env.COMMIT_HASH }}" \
+          --project-version "${COMMIT_HASH}" \
           --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
+          --commit-date "${COMMIT_DATE}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --name-suffix avx512 \
           --walk-subdirs \
           --throughput
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
         with:
           name: ${{ github.sha }}_core_crypto
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -154,39 +144,14 @@ jobs:
           -d @${{ env.RESULTS_FILENAME }} \
           ${{ secrets.SLAB_URL }}
 
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-core-crypto-benchmarks ]
-    runs-on: ubuntu-latest
-    if: ${{ !success() && !cancelled() }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ needs.cuda-core-crypto-benchmarks.result }}
-          SLACK_MESSAGE: "PBS GPU benchmarks finished with status: ${{ needs.cuda-core-crypto-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-integer-full-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-core-crypto-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "PBS GPU benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/csprng_randomness_tests.yml

@@ -17,44 +17,44 @@ on:
   pull_request:
     types: [ labeled ]
 
+
 jobs:
-  setup-instance:
-    name: Setup instance (csprng-randomness-tests)
+  setup-ec2:
+    name: Setup EC2 instance (csprng-randomness-tests)
     if: ${{ github.event_name == 'workflow_dispatch' || contains(github.event.label.name, 'approved') }}
     runs-on: ubuntu-latest
     outputs:
       runner-name: ${{ steps.start-instance.outputs.label }}
+      instance-id: ${{ steps.start-instance.outputs.ec2-instance-id }}
+      aws-region: ${{ steps.start-instance.outputs.aws-region }}
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
           slab-url: ${{ secrets.SLAB_BASE_URL }}
           job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
           profile: cpu-small
 
   csprng-randomness-tests:
     name: CSPRNG randomness tests
-    needs: setup-instance
+    needs: setup-ec2
     concurrency:
       group: ${{ github.workflow }}_${{ github.ref }}
       cancel-in-progress: true
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    runs-on: ${{ needs.setup-ec2.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: stable
 
@@ -65,31 +65,32 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "concrete-csprng randomness check finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
 
-  teardown-instance:
-    name: Teardown instance (csprng-randomness-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, csprng-randomness-tests ]
+  teardown-ec2:
+    name: Teardown EC2 instance (csprng-randomness-tests)
+    if: ${{ always() && needs.setup-ec2.result != 'skipped' }}
+    needs: [ setup-ec2, csprng-randomness-tests ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@8562abbdc96b3619bd5debe1fb934db298f9a044
         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 }}
+          region: ${{ needs.setup-ec2.outputs.aws-region }}
+          label: ${{ needs.setup-ec2.outputs.runner-name }}
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (csprng-randomness-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "EC2 teardown (csprng-randomness-tests) failed. (${{ env.ACTION_RUN_URL }})"

.github/workflows/data_pr_close.yml

@@ -1,123 +0,0 @@
-name: Close or Merge corresponding PR on the data repo
-
-# When a PR with the data_PR tag is closed or merged, this will close the corresponding PR in the data repo.
-
-env:
-  TARGET_REPO_API_URL: ${{ github.api_url }}/repos/zama-ai/tfhe-backward-compat-data
-  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 }}
-  PR_BRANCH: ${{ github.head_ref || github.ref_name }}
-  CLOSE_TYPE: ${{ github.event.pull_request.merged && 'merge' || 'close' }}
-
-# only trigger on pull request closed events
-on:
-  pull_request:
-    types: [ closed ]
-
-# The same pattern is used for jobs that use the github api:
-# - save the result of the API call in the env var "GH_API_RES". Since the var is multiline
-# we use this trick: https://docs.github.com/en/actions/using-workflows/workflow-commands-for-github-actions#example-of-a-multiline-string
-# - "set +e" will make sure we reach the last "echo EOF" even in case of error
-# - "set -o" pipefail makes one line piped command return the error of the first failure
-# - 'RES="$?"' and 'exit $RES' are used to return the error code if a command failed. Without it, with "set +e"
-# the script will always return 0 because of the "echo EOF".
-
-
-jobs:
-  auto_close_job:
-    if: ${{ contains(github.event.pull_request.labels.*.name, 'data_PR') }}
-    runs-on: ubuntu-latest
-    steps:
-    - name: Find corresponding Pull Request in the data repo
-      run: |
-        {
-          set +e
-          set -o pipefail
-          echo 'TARGET_REPO_PR<<EOF'
-          curl --fail-with-body --no-progress-meter -L -X GET \
-          -H "Accept: application/vnd.github+json" \
-          -H "X-GitHub-Api-Version: 2022-11-28"  \
-          ${{ env.TARGET_REPO_API_URL }}/pulls\?head=${{ github.repository_owner }}:${{ env.PR_BRANCH }} | jq -e '.[0]' | sed 's/null/{ "message": "corresponding PR not found" }/'
-          RES="$?"
-          echo EOF
-        } >> "${GITHUB_ENV}"
-        exit $RES
-
-    - name: Comment on the PR to indicate the reason of the close
-      run: |
-        {
-          set +e
-          set -o pipefail
-          echo 'GH_API_RES<<EOF'
-          curl --fail-with-body --no-progress-meter -L -X POST \
-          -H "Accept: application/vnd.github+json" \
-          -H "Authorization: Bearer ${{ secrets.FHE_ACTIONS_TOKEN }}" \
-          -H "X-GitHub-Api-Version: 2022-11-28" \
-          ${{ fromJson(env.TARGET_REPO_PR).comments_url }} \
-          -d '{ "body": "PR ${{ env.CLOSE_TYPE }}d because the corresponding PR in main repo was ${{ env.CLOSE_TYPE }}d: ${{ github.repository }}#${{ github.event.number  }}" }'
-          RES="$?"
-          echo EOF
-        } >> "${GITHUB_ENV}"
-        exit $RES
-
-    - name: Merge the Pull Request in the data repo
-      if: ${{ github.event.pull_request.merged }}
-      run: |
-        {
-          set +e
-          set -o pipefail
-          echo 'GH_API_RES<<EOF'
-          curl --fail-with-body --no-progress-meter -L -X PUT \
-          -H "Accept: application/vnd.github+json" \
-          -H "Authorization: Bearer ${{ secrets.FHE_ACTIONS_TOKEN }}" \
-          -H "X-GitHub-Api-Version: 2022-11-28" \
-          ${{ fromJson(env.TARGET_REPO_PR).url }}/merge \
-          -d '{ "merge_method": "rebase" }'
-          RES="$?"
-          echo EOF
-        } >> "${GITHUB_ENV}"
-        exit $RES
-
-    - name: Close the Pull Request in the data repo
-      if: ${{ !github.event.pull_request.merged }}
-      run: |
-        {
-          set +e
-          set -o pipefail
-          echo 'GH_API_RES<<EOF'
-          curl --fail-with-body --no-progress-meter -L -X PATCH \
-          -H "Accept: application/vnd.github+json" \
-          -H "Authorization: Bearer ${{ secrets.FHE_ACTIONS_TOKEN }}" \
-          -H "X-GitHub-Api-Version: 2022-11-28" \
-          ${{ fromJson(env.TARGET_REPO_PR).url }} \
-          -d '{ "state": "closed" }'
-          RES="$?"
-          echo EOF
-        } >> "${GITHUB_ENV}"
-        exit $RES
-
-    - name: Delete the associated branch in the data repo
-      run: |
-        {
-          set +e
-          set -o pipefail
-          echo 'GH_API_RES<<EOF'
-          curl --fail-with-body --no-progress-meter -L -X DELETE \
-          -H "Accept: application/vnd.github+json" \
-          -H "Authorization: Bearer ${{ secrets.FHE_ACTIONS_TOKEN }}" \
-          -H "X-GitHub-Api-Version: 2022-11-28" \
-          ${{ env.TARGET_REPO_API_URL }}/git/refs/heads/${{ env.PR_BRANCH }}
-          RES="$?"
-          echo EOF
-        } >> "${GITHUB_ENV}"
-        exit $RES
-
-    - name: Slack Notification
-      if: ${{ always() && job.status == 'failure' }}
-      continue-on-error: true
-      uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-      env:
-        SLACK_COLOR: ${{ job.status }}
-        SLACK_MESSAGE: "Failed to auto-${{ env.CLOSE_TYPE }} PR on data repo: ${{ fromJson(env.GH_API_RES || env.TARGET_REPO_PR).message }}"

.github/workflows/gpu_4090_full_benchmark.yml

@@ -1,5 +1,5 @@
-# Run benchmarks on an RTX 4090 machine and return parsed results to Slab CI bot.
-name: TFHE Cuda Backend - 4090 benchmarks
+# Run all benchmarks on an RTX 4090 machine and return parsed results to Slab CI bot.
+name: TFHE Cuda Backend - 4090 full benchmarks
 
 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 }}
-  FAST_BENCH: TRUE
 
 on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
@@ -24,22 +23,23 @@ on:
 
 jobs:
   cuda-integer-benchmarks:
-    name: Cuda integer benchmarks (RTX 4090)
-    if: ${{ github.event_name == 'workflow_dispatch' ||
-      github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs' ||
-      contains(github.event.label.name, '4090_bench') }}
+    name: Cuda integer benchmarks for all operations flavor  (RTX 4090)
+    if: ${{ github.event_name == 'workflow_dispatch' || github.event_name == 'schedule' || contains(github.event.label.name, '4090_bench') }}
     concurrency:
       group: ${{ github.workflow }}_${{ github.ref }}_cuda_integer_bench
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
+      cancel-in-progress: true
     runs-on: ["self-hosted", "4090-desktop"]
     timeout-minutes: 1440 # 24 hours
     strategy:
       fail-fast: false
       max-parallel: 1
+      matrix:
+        command: [integer, integer_multi_bit]
+        op_flavor: [default, unchecked]
 
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           fetch-depth: 0
 
@@ -50,23 +50,22 @@ jobs:
             echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
-          echo "FAST_BENCH=TRUE" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
 
       - name: Run integer benchmarks
         run: |
-          make BENCH_OP_FLAVOR=default bench_integer_multi_bit_gpu
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_${{ matrix.command }}_gpu
 
       - name: Parse results
         run: |
@@ -82,9 +81,9 @@ jobs:
           --throughput
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
         with:
-          name: ${{ github.sha }}_integer_multi_bit_gpu_default
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Send data to Slab
@@ -104,10 +103,10 @@ jobs:
       - name: Slack Notification
         if: ${{ always() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         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 failed. (${{ env.ACTION_RUN_URL }})"
 
   cuda-core-crypto-benchmarks:
     name: Cuda core crypto benchmarks  (RTX 4090)
@@ -115,13 +114,13 @@ jobs:
     needs: cuda-integer-benchmarks
     concurrency:
       group: ${{ github.workflow }}_${{ github.ref }}_cuda_core_crypto_bench
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
+      cancel-in-progress: true
     runs-on: ["self-hosted", "4090-desktop"]
     timeout-minutes: 1440 # 24 hours
 
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           fetch-depth: 0
 
@@ -134,18 +133,18 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
 
-      - name: Run core crypto benchmarks
+      - name: Run integer benchmarks
         run: |
           make bench_pbs_gpu
           make bench_ks_gpu
@@ -164,7 +163,7 @@ jobs:
           --throughput
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
         with:
           name: ${{ github.sha }}_core_crypto
           path: ${{ env.RESULTS_FILENAME }}
@@ -186,16 +185,16 @@ jobs:
       - name: Slack Notification
         if: ${{ !success() && !cancelled() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         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 failed. (${{ env.ACTION_RUN_URL }})"
 
   remove_github_label:
     name: Remove 4090 bench label
-    if: ${{ always() && github.event_name == 'pull_request' }}
+    if: ${{ github.event_name == 'pull_request' }}
     needs: [cuda-integer-benchmarks, cuda-core-crypto-benchmarks]
-    runs-on: ubuntu-latest
+    runs-on: ["self-hosted", "4090-desktop"]
     steps:
       - uses: actions-ecosystem/action-remove-labels@2ce5d41b4b6aa8503e285553f75ed56e0a40bae0
         with:

.github/workflows/gpu_fast_h100_tests.yml

@@ -1,199 +0,0 @@
-# Compile and test tfhe-cuda-backend on an H100 VM on hyperstack
-name: TFHE Cuda Backend - Fast tests on H100
-
-env:
-  CARGO_TERM_COLOR: always
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUSTFLAGS: "-C target-cpu=native"
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' }}
-
-on:
-  # Allows you to run this workflow manually from the Actions tab as an alternative.
-  workflow_dispatch:
-  pull_request:
-      types: [ labeled ]
-
-jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    permissions:
-      pull-requests: write
-    outputs:
-      gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@c65cd883420fd2eb864698a825fc4162dd94482c
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            gpu:
-              - tfhe/Cargo.toml
-              - tfhe/build.rs
-              - backends/tfhe-cuda-backend/**
-              - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/gpu/**
-              - tfhe/shortint/parameters/**
-              - tfhe/src/high_level_api/**
-              - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
-              - Makefile
-              - '.github/workflows/gpu_fast_h100_tests.yml'
-              - scripts/**
-              - ci/**
-
-  setup-instance:
-    name: Setup instance (cuda-h100-tests)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
-      (github.event.action != 'labeled' && needs.should-run.outputs.gpu_test == 'true') ||
-      (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.gpu_test == 'true')
-    runs-on: ubuntu-latest
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        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-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')
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    strategy:
-      fail-fast: false
-      # explicit include-based build matrix, of known valid options
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
-    steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
-        run: |
-          sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
-          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          cd cmake-${{ env.CMAKE_VERSION }}
-          ./bootstrap
-          make -j"$(nproc)"
-          sudo make install
-
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
-        with:
-          toolchain: stable
-
-      - name: Export CUDA variables
-        if: ${{ !cancelled() }}
-        run: |
-          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
-          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
-          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
-          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
-
-      # Specify the correct host compilers
-      - name: Export gcc and g++ variables
-        if: ${{ !cancelled() }}
-        run: |
-          {
-            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
-            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "HOME=/home/ubuntu";
-          } >> "${GITHUB_ENV}"
-
-      - name: Check device is detected
-        if: ${{ !cancelled() }}
-        run: nvidia-smi
-
-      - name: Run core crypto and internal CUDA backend tests
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_core_crypto_gpu
-          BIG_TESTS_INSTANCE=TRUE make test_cuda_backend
-
-      - name: Run user docs tests
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_user_doc_gpu
-
-      - name: Test C API
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_c_api_gpu
-
-      - name: Run High Level API Tests
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_high_level_api_gpu
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-tests-linux.result != 'skipped' }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
-          SLACK_MESSAGE: "Fast H100 tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-h100-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_fast_tests.yml

@@ -1,197 +0,0 @@
-# Compile and test tfhe-cuda-backend on an AWS instance
-name: TFHE Cuda Backend - Fast tests
-
-env:
-  CARGO_TERM_COLOR: always
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUSTFLAGS: "-C target-cpu=native"
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' }}
-
-on:
-  # Allows you to run this workflow manually from the Actions tab as an alternative.
-  workflow_dispatch:
-  pull_request:
-
-jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    permissions:
-      pull-requests: write
-    outputs:
-      gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@c65cd883420fd2eb864698a825fc4162dd94482c
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            gpu:
-              - tfhe/Cargo.toml
-              - tfhe/build.rs
-              - backends/tfhe-cuda-backend/**
-              - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/gpu/**
-              - tfhe/shortint/parameters/**
-              - tfhe/src/high_level_api/**
-              - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
-              - '.github/workflows/gpu_fast_tests.yml'
-              - Makefile
-              - scripts/**
-              - ci/**
-
-  setup-instance:
-    name: Setup instance (cuda-tests)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
-      needs.should-run.outputs.gpu_test == 'true'
-    runs-on: ubuntu-latest
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: hyperstack
-          profile: gpu-test
-
-  cuda-tests-linux:
-    name: CUDA tests
-    needs: [ should-run, setup-instance ]
-    if: github.event_name != 'pull_request' ||
-      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    strategy:
-      fail-fast: false
-      # explicit include-based build matrix, of known valid options
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
-    steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
-        run: |
-          sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
-          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          cd cmake-${{ env.CMAKE_VERSION }}
-          ./bootstrap
-          make -j"$(nproc)"
-          sudo make install
-
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
-        with:
-          toolchain: stable
-
-      - name: Export CUDA variables
-        if: ${{ !cancelled() }}
-        run: |
-          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
-          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
-          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
-          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
-
-      # Specify the correct host compilers
-      - name: Export gcc and g++ variables
-        if: ${{ !cancelled() }}
-        run: |
-          {
-            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
-            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "HOME=/home/ubuntu";
-          } >> "${GITHUB_ENV}"
-
-      - name: Check device is detected
-        if: ${{ !cancelled() }}
-        run: nvidia-smi
-
-      - name: Run core crypto and internal CUDA backend tests
-        run: |
-          make test_core_crypto_gpu
-          make test_cuda_backend
-
-      - name: Run user docs tests
-        run: |
-          make test_user_doc_gpu
-
-      - name: Test C API
-        run: |
-          make test_c_api_gpu
-
-      - name: Run High Level API Tests
-        run: |
-          make test_high_level_api_gpu
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-tests-linux.result != 'skipped' }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
-          SLACK_MESSAGE: "Base GPU tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_full_multi_gpu_tests.yml

@@ -1,199 +0,0 @@
-# Compile and test tfhe-cuda-backend on an AWS instance
-name: TFHE Cuda Backend - Full tests multi-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 }}
-  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' }}
-
-on:
-  # Allows you to run this workflow manually from the Actions tab as an alternative.
-  workflow_dispatch:
-  pull_request:
-    types: [ labeled ]
-
-jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    permissions:
-      pull-requests: write
-    outputs:
-      gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@c65cd883420fd2eb864698a825fc4162dd94482c
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            gpu:
-              - tfhe/Cargo.toml
-              - tfhe/build.rs
-              - backends/tfhe-cuda-backend/**
-              - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/gpu/**
-              - tfhe/shortint/parameters/**
-              - tfhe/src/high_level_api/**
-              - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
-              - Makefile
-              - '.github/workflows/**_multi_gpu_tests.yml'
-              - scripts/**
-              - ci/**
-
-  setup-instance:
-    name: Setup instance (cuda-tests-multi-gpu)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
-      (github.event.action != 'labeled' && needs.should-run.outputs.gpu_test == 'true') ||
-      (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.gpu_test == 'true')
-    runs-on: ubuntu-latest
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        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-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')
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    strategy:
-      fail-fast: false
-      # explicit include-based build matrix, of known valid options
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
-    steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
-        run: |
-          sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
-          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          cd cmake-${{ env.CMAKE_VERSION }}
-          ./bootstrap
-          make -j"$(nproc)"
-          sudo make install
-
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
-        with:
-          toolchain: stable
-
-      - name: Export CUDA variables
-        if: ${{ !cancelled() }}
-        run: |
-          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
-          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
-          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
-          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
-
-      # Specify the correct host compilers
-      - name: Export gcc and g++ variables
-        if: ${{ !cancelled() }}
-        run: |
-          {
-            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
-            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "HOME=/home/ubuntu";
-          } >> "${GITHUB_ENV}"
-
-      - name: Check device is detected
-        if: ${{ !cancelled() }}
-        run: nvidia-smi
-
-      # No need to test core_crypto and classic PBS in integer since it's already tested on single GPU.
-      - name: Run multi-bit CUDA integer tests
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_integer_multi_bit_gpu_ci
-
-      - name: Run user docs tests
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_user_doc_gpu
-
-      - name: Test C API
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_c_api_gpu
-
-      - name: Run High Level API Tests
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_high_level_api_gpu
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-tests-linux.result != 'skipped' }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
-          SLACK_MESSAGE: "Multi-GPU tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-tests-multi-gpu)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-tests-multi-gpu) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_signed_integer_h100_tests.yml

@@ -1,188 +0,0 @@
-# Signed integer GPU tests on an H100 VM on hyperstack
-name: TFHE Cuda Backend - Signed integer tests on H100
-
-env:
-  CARGO_TERM_COLOR: always
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUSTFLAGS: "-C target-cpu=native"
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' }}
-
-on:
-  # Allows you to run this workflow manually from the Actions tab as an alternative.
-  workflow_dispatch:
-  pull_request:
-      types: [ labeled ]
-
-jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    permissions:
-      pull-requests: write
-    outputs:
-      gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@c65cd883420fd2eb864698a825fc4162dd94482c
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            gpu:
-              - tfhe/Cargo.toml
-              - tfhe/build.rs
-              - backends/tfhe-cuda-backend/**
-              - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/gpu/**
-              - tfhe/shortint/parameters/**
-              - tfhe/src/high_level_api/**
-              - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
-              - Makefile
-              - '.github/workflows/gpu_signed_integer_h100_tests.yml'
-              - scripts/**
-              - ci/**
-
-  setup-instance:
-    name: Setup instance (cuda-h100-tests)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
-      (github.event.action != 'labeled' && needs.should-run.outputs.gpu_test == 'true') ||
-      (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.gpu_test == 'true')
-    runs-on: ubuntu-latest
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        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-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')
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    strategy:
-      fail-fast: false
-      # explicit include-based build matrix, of known valid options
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
-    steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
-        run: |
-          sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
-          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          cd cmake-${{ env.CMAKE_VERSION }}
-          ./bootstrap
-          make -j"$(nproc)"
-          sudo make install
-
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
-        with:
-          toolchain: stable
-
-      - name: Export CUDA variables
-        if: ${{ !cancelled() }}
-        run: |
-          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
-          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
-          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
-          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
-
-      # Specify the correct host compilers
-      - name: Export gcc and g++ variables
-        if: ${{ !cancelled() }}
-        run: |
-          {
-            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
-            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "HOME=/home/ubuntu";
-          } >> "${GITHUB_ENV}"
-
-      - name: Check device is detected
-        if: ${{ !cancelled() }}
-        run: nvidia-smi
-
-      - name: Run signed integer tests
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_signed_integer_gpu_ci
-
-      - name: Run signed integer multi-bit tests
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_signed_integer_multi_bit_gpu_ci
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-tests-linux.result != 'skipped' }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
-          SLACK_MESSAGE: "Integer GPU H100 tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-h100-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_signed_integer_tests.yml

@@ -1,202 +0,0 @@
-# Compile and test tfhe-cuda-backend signed integer on an AWS instance
-name: TFHE Cuda Backend - Signed integer tests
-
-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 }}
-  FAST_TESTS: TRUE
-  NIGHTLY_TESTS: FALSE
-  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' }}
-
-on:
-  # Allows you to run this workflow manually from the Actions tab as an alternative.
-  workflow_dispatch:
-  pull_request:
-    types:
-      - opened
-      - synchronize
-      - labeled
-  schedule:
-    # Nightly tests @ 1AM after each work day
-    - cron: "0 1 * * MON-FRI"
-
-jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    permissions:
-      pull-requests: write
-    outputs:
-      gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@c65cd883420fd2eb864698a825fc4162dd94482c
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            gpu:
-              - tfhe/Cargo.toml
-              - tfhe/build.rs
-              - backends/tfhe-cuda-backend/**
-              - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/gpu/**
-              - tfhe/shortint/parameters/**
-              - tfhe/src/high_level_api/**
-              - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
-              - '.github/workflows/gpu_signed_integer_tests.yml'
-              - Makefile
-              - scripts/**
-              - ci/**
-
-  setup-instance:
-    name: Setup instance (cuda-signed-integer-tests)
-    runs-on: ubuntu-latest
-    needs: should-run
-    if: (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
-      github.event_name == 'workflow_dispatch' ||
-      (github.event.action != 'labeled' && needs.should-run.outputs.gpu_test == 'true')
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: hyperstack
-          profile: gpu-test
-
-  cuda-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')
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    strategy:
-      fail-fast: false
-      # explicit include-based build matrix, of known valid options
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
-    steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
-        run: |
-          sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
-          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          cd cmake-${{ env.CMAKE_VERSION }}
-          ./bootstrap
-          make -j"$(nproc)"
-          sudo make install
-
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
-        with:
-          toolchain: stable
-
-      - name: Export CUDA variables
-        if: ${{ !cancelled() }}
-        run: |
-          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
-          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
-          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
-          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
-
-      # Specify the correct host compilers
-      - name: Export gcc and g++ variables
-        if: ${{ !cancelled() }}
-        run: |
-          {
-            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
-            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "HOME=/home/ubuntu";
-          } >> "${GITHUB_ENV}"
-
-      - name: Should run nightly tests
-        if: github.event_name == 'schedule'
-        run: |
-          {
-            echo "FAST_TESTS=FALSE";
-            echo "NIGHTLY_TESTS=TRUE";
-          } >> "${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
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-signed-integer-tests ]
-    runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-signed-integer-tests.result != 'skipped' }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ needs.cuda-signed-integer-tests.result }}
-          SLACK_MESSAGE: "Base GPU tests finished with status: ${{ needs.cuda-signed-integer-tests.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-signed-integer-tests ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-signed-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_unsigned_integer_h100_tests.yml

@@ -1,188 +0,0 @@
-# Test unsigned integers on an H100 VM on hyperstack
-name: TFHE Cuda Backend - Unsigned integer tests on H100
-
-env:
-  CARGO_TERM_COLOR: always
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUSTFLAGS: "-C target-cpu=native"
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' }}
-
-on:
-  # Allows you to run this workflow manually from the Actions tab as an alternative.
-  workflow_dispatch:
-  pull_request:
-      types: [ labeled ]
-
-jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    permissions:
-      pull-requests: write
-    outputs:
-      gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@c65cd883420fd2eb864698a825fc4162dd94482c
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            gpu:
-              - tfhe/Cargo.toml
-              - tfhe/build.rs
-              - backends/tfhe-cuda-backend/**
-              - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/gpu/**
-              - tfhe/shortint/parameters/**
-              - tfhe/src/high_level_api/**
-              - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
-              - Makefile
-              - '.github/workflows/gpu_unsigned_integer_tests.yml'
-              - scripts/**
-              - ci/**
-
-  setup-instance:
-    name: Setup instance (cuda-h100-tests)
-    needs: should-run
-    if: github.event_name != 'pull_request' ||
-      (github.event.action != 'labeled' && needs.should-run.outputs.gpu_test == 'true') ||
-      (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.gpu_test == 'true')
-    runs-on: ubuntu-latest
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        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-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')
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    strategy:
-      fail-fast: false
-      # explicit include-based build matrix, of known valid options
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11 
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
-    steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
-        run: |
-          sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
-          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          cd cmake-${{ env.CMAKE_VERSION }}
-          ./bootstrap
-          make -j"$(nproc)"
-          sudo make install
-
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
-        with:
-          toolchain: stable
-
-      - name: Export CUDA variables
-        if: ${{ !cancelled() }}
-        run: |
-          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
-          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
-          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
-          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
-
-      # Specify the correct host compilers
-      - name: Export gcc and g++ variables
-        if: ${{ !cancelled() }}
-        run: |
-          {
-            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
-            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "HOME=/home/ubuntu";
-          } >> "${GITHUB_ENV}"
-
-      - name: Check device is detected
-        if: ${{ !cancelled() }}
-        run: nvidia-smi
-
-      - name: Run unsigned integer tests
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_unsigned_integer_gpu_ci
-
-      - name: Run unsigned integer multi-bit tests
-        run: |
-          BIG_TESTS_INSTANCE=TRUE make test_unsigned_integer_multi_bit_gpu_ci
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-tests-linux.result != 'skipped' }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
-          SLACK_MESSAGE: "Unsigned integer GPU H100 tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-h100-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-tests-linux ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_unsigned_integer_tests.yml

@@ -1,199 +0,0 @@
-# Compile and test tfhe-cuda-backend unsigned integer on an AWS instance
-name: TFHE Cuda Backend - Unsigned integer tests
-
-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 }}
-  FAST_TESTS: TRUE
-  NIGHTLY_TESTS: FALSE
-
-on:
-  # Allows you to run this workflow manually from the Actions tab as an alternative.
-  workflow_dispatch:
-  pull_request:
-    types:
-      - opened
-      - synchronize
-      - labeled
-  schedule:
-    # Nightly tests @ 1AM after each work day
-    - cron: "0 1 * * MON-FRI"
-
-jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    permissions:
-      pull-requests: write
-    outputs:
-      gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@c65cd883420fd2eb864698a825fc4162dd94482c
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            gpu:
-              - tfhe/Cargo.toml
-              - tfhe/build.rs
-              - backends/tfhe-cuda-backend/**
-              - tfhe/src/core_crypto/gpu/**
-              - tfhe/src/integer/gpu/**
-              - tfhe/shortint/parameters/**
-              - tfhe/src/high_level_api/**
-              - tfhe/src/c_api/**
-              - 'tfhe/docs/**.md'
-              - '.github/workflows/gpu_unsigned_integer_tests.yml'
-              - Makefile
-              - scripts/**
-              - ci/**
-
-  setup-instance:
-    name: Setup instance (cuda-unsigned-integer-tests)
-    needs: should-run
-    if: (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
-      github.event_name == 'workflow_dispatch' ||
-      (github.event.action != 'labeled' && needs.should-run.outputs.gpu_test == 'true')
-    runs-on: ubuntu-latest
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: hyperstack
-          profile: gpu-test
-
-  cuda-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')
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    strategy:
-      fail-fast: false
-      # explicit include-based build matrix, of known valid options
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
-    steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
-        run: |
-          sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
-          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          cd cmake-${{ env.CMAKE_VERSION }}
-          ./bootstrap
-          make -j"$(nproc)"
-          sudo make install
-
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
-        with:
-          toolchain: stable
-
-      - name: Export CUDA variables
-        if: ${{ !cancelled() }}
-        run: |
-          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
-          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
-          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
-          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
-
-      # Specify the correct host compilers
-      - name: Export gcc and g++ variables
-        if: ${{ !cancelled() }}
-        run: |
-          {
-            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
-            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "HOME=/home/ubuntu";
-          } >> "${GITHUB_ENV}"
-
-      - name: Should run nightly tests
-        if: github.event_name == 'schedule'
-        run: |
-          {
-            echo "FAST_TESTS=FALSE";
-            echo "NIGHTLY_TESTS=TRUE";
-          } >> "${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
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-unsigned-integer-tests ]
-    runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-unsigned-integer-tests.result != 'skipped' }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ needs.cuda-unsigned-integer-tests.result }}
-          SLACK_MESSAGE: "Unsigned integer GPU tests finished with status: ${{ needs.cuda-unsigned-integer-tests.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-tests)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-unsigned-integer-tests ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-unsigned-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/integer_benchmark.yml

@@ -0,0 +1,130 @@
+# Run integer benchmarks on an AWS instance and return parsed results to Slab CI bot.
+name: Integer benchmarks
+
+on:
+  workflow_dispatch:
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
+
+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"
+
+jobs:
+  run-integer-benchmarks:
+    name: Execute integer benchmarks in EC2
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ !cancelled() }}
+    steps:
+      - name: Instance configuration used
+        run: |
+          echo "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Get benchmark date
+        run: |
+          echo "BENCH_DATE=$(date --iso-8601=seconds)" >> "${GITHUB_ENV}"
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          fetch-depth: 0
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
+        with:
+          toolchain: nightly
+
+      - name: Run benchmarks with AVX512
+        run: |
+          make FAST_BENCH=TRUE bench_integer
+
+      - 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@5d5d22a31266ced268874388b861e4b58bb5c2f3
+        with:
+          name: ${{ github.sha }}_csv_integer
+          path: ${{ env.PARSE_INTEGER_BENCH_CSV_FILE }}
+
+      - name: Parse results
+        run: |
+          COMMIT_DATE="$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})"
+          COMMIT_HASH="$(git describe --tags --dirty)"
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware ${{ inputs.instance_type }} \
+          --project-version "${COMMIT_HASH}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${COMMIT_DATE}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --throughput
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
+        with:
+          name: ${{ github.sha }}_integer
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          echo "Computing HMac on results file"
+          SIGNATURE="$(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@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "Integer benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/integer_cpu_benchmark.yml

@@ -1,191 +0,0 @@
-# Run all integer benchmarks on an AWS instance and return parsed results to Slab CI bot.
-name: Integer benchmarks
-
-on:
-  workflow_dispatch:
-    inputs:
-      all_precisions:
-        description: "Run all precisions"
-        type: boolean
-        default: false
-
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 1a.m.
-    - cron: '0 1 * * 6'
-    # Quarterly benchmarks will be triggered right before end of quarter, the 25th of the current month at 4a.m.
-    # These benchmarks are far longer to execute hence the reason to run them only four time a year.
-    - cron: '0 4 25 MAR,JUN,SEP,DEC *'
-
-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
-    if: github.event_name != 'schedule' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
-    outputs:
-      op_flavor: ${{ steps.set_op_flavor.outputs.op_flavor }}
-    steps:
-      - name: Weekly benchmarks
-        if: github.event_name == 'workflow_dispatch' ||
-          github.event.schedule == '0 1 * * 6'
-        run: |
-          echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
-
-      - name: Quarterly benchmarks
-        if: github.event.schedule == '0 4 25 MAR,JUN,SEP,DEC *'
-        run: |
-          echo "OP_FLAVOR=[\"default\", \"smart\", \"unchecked\", \"misc\"]" >> "${GITHUB_ENV}"
-
-      -  name: Set operation flavor output
-         id: set_op_flavor
-         run: |
-          echo "op_flavor=${{ toJSON(env.OP_FLAVOR) }}" >> "${GITHUB_OUTPUT}"
-
-  setup-instance:
-    name: Setup instance (integer-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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: bench
-
-  integer-benchmarks:
-    name: Execute integer benchmarks for all operations flavor
-    needs: [ prepare-matrix, setup-instance ]
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    continue-on-error: true
-    timeout-minutes: 1440  # 24 hours
-    strategy:
-      max-parallel: 1
-      matrix:
-        command: [ integer, integer_multi_bit]
-        op_flavor: ${{ fromJson(needs.prepare-matrix.outputs.op_flavor) }}
-    steps:
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Get benchmark details
-        run: |
-          {
-            echo "BENCH_DATE=$(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)";
-          } >> "${GITHUB_ENV}"
-
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
-        with:
-          toolchain: nightly
-
-      - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        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: Run benchmarks with AVX512
-        run: |
-          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_${{ matrix.command }}
-
-      - name: Parse results
-        run: |
-          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
-          --database tfhe_rs \
-          --hardware "hpc7a.96xlarge" \
-          --project-version "${{ env.COMMIT_HASH }}" \
-          --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
-          --bench-date "${{ env.BENCH_DATE }}" \
-          --walk-subdirs \
-          --name-suffix avx512 \
-          --throughput
-
-      - name: Upload parsed results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
-        with:
-          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
-          path: ${{ env.RESULTS_FILENAME }}
-
-      - name: Send data to Slab
-        shell: bash
-        run: |
-          echo "Computing HMac on results file"
-          SIGNATURE="$(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@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Integer full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (integer-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, integer-benchmarks ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (integer-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/integer_full_benchmark.yml

@@ -0,0 +1,158 @@
+# Run all integer benchmarks on an AWS instance and return parsed results to Slab CI bot.
+name: Integer full benchmarks
+
+on:
+  workflow_dispatch:
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
+      user_inputs:
+        description: "Type of benchmarks to run"
+        type: string
+        default: "weekly_benchmarks"
+
+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"
+
+jobs:
+  prepare-matrix:
+    name: Prepare operations matrix
+    runs-on: ubuntu-latest
+    outputs:
+      op_flavor: ${{ steps.set_op_flavor.outputs.op_flavor }}
+    steps:
+      - name: Weekly benchmarks
+        if: ${{ github.event.inputs.user_inputs == 'weekly_benchmarks' }}
+        run: |
+          echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
+
+      - name: Quarterly benchmarks
+        if: ${{ github.event.inputs.user_inputs == 'quarterly_benchmarks' }}
+        run: |
+          echo "OP_FLAVOR=[\"default\", \"smart\", \"unchecked\", \"misc\"]" >> "${GITHUB_ENV}"
+
+      -  name: Set operation flavor output
+         id: set_op_flavor
+         run: |
+          echo "op_flavor=${{ toJSON(env.OP_FLAVOR) }}" >> "${GITHUB_OUTPUT}"
+
+  integer-benchmarks:
+    name: Execute integer benchmarks for all operations flavor
+    needs: prepare-matrix
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ !cancelled() }}
+    continue-on-error: true
+    timeout-minutes: 1440  # 24 hours
+    strategy:
+      max-parallel: 1
+      matrix:
+        command: [ integer, integer_multi_bit]
+        op_flavor: ${{ fromJson(needs.prepare-matrix.outputs.op_flavor) }}
+    steps:
+      - name: Instance configuration used
+        run: |
+          echo "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          fetch-depth: 0
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(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)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
+        with:
+          toolchain: nightly
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
+
+      - name: Run benchmarks with AVX512
+        run: |
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_${{ matrix.command }}
+
+      - name: Parse results
+        run: |
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware ${{ inputs.instance_type }} \
+          --project-version "${{ 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@5d5d22a31266ced268874388b861e4b58bb5c2f3
+        with:
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          echo "Computing HMac on results file"
+          SIGNATURE="$(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 }}
+
+  slack-notification:
+    name: Slack Notification
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ failure() }}
+    needs: integer-benchmarks
+    steps:
+      - name: Notify
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "Integer full benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/integer_gpu_benchmark.yml

@@ -1,11 +1,24 @@
-# Run integer benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
+# Run integer benchmarks on an AWS instance with CUDA and return parsed results to Slab CI bot.
 name: Integer GPU benchmarks
 
 on:
   workflow_dispatch:
-  push:
-    branches:
-      - main
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
 
 env:
   CARGO_TERM_COLOR: always
@@ -14,35 +27,12 @@ env:
   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-benchmarks)
-    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 }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        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 }}
+  run-integer-benchmarks:
+    name: Execute integer benchmarks in EC2
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ !cancelled() }}
     strategy:
       fail-fast: false
       # explicit include-based build matrix, of known valid options
@@ -50,43 +40,33 @@ jobs:
         include:
           - os: ubuntu-22.04
             cuda: "12.2"
-            gcc: 11
+            gcc: 9
     env:
       CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
     steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
+      - name: Instance configuration used
         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 "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Get benchmark date
+        run: |
+          echo "BENCH_DATE=$(date --iso-8601=seconds)" >> "${GITHUB_ENV}"
 
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         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)";
-          } >> "${GITHUB_ENV}"
-
       - name: Set up home
         # "Install rust" step require root user to have a HOME directory which is not set.
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: nightly
 
@@ -110,10 +90,6 @@ jobs:
             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
@@ -124,37 +100,39 @@ jobs:
             parse_integer_benches
 
       - name: Upload csv results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
         with:
           name: ${{ github.sha }}_csv_integer
           path: ${{ env.PARSE_INTEGER_BENCH_CSV_FILE }}
 
       - name: Parse results
         run: |
+          COMMIT_DATE="$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})"
+          COMMIT_HASH="$(git describe --tags --dirty)"
           python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
           --database tfhe_rs \
-          --hardware "n3-H100x1" \
+          --hardware ${{ inputs.instance_type }} \
           --backend gpu \
-          --project-version "${{ env.COMMIT_HASH }}" \
+          --project-version "${COMMIT_HASH}" \
           --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
+          --commit-date "${COMMIT_DATE}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
           --name-suffix avx512 \
           --throughput
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
         with:
           name: ${{ github.sha }}_integer
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -170,39 +148,14 @@ jobs:
           -d @${{ env.RESULTS_FILENAME }} \
           ${{ secrets.SLAB_URL }}
 
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-integer-benchmarks ]
-    runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-integer-benchmarks.result != 'skipped' && failure() }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ needs.cuda-integer-benchmarks.result }}
-          SLACK_MESSAGE: "Integer GPU benchmarks finished with status: ${{ needs.cuda-integer-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-integer-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-integer-benchmarks, slack-notify ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        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() }}
+        if: ${{ !success() && !cancelled() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-integer-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "Integer GPU benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/integer_gpu_full_benchmark.yml

@@ -1,11 +1,32 @@
-# Run all integer benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
+# Run all integer benchmarks on an AWS 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'
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
+      # This input is not used in this workflow but still mandatory since a calling workflow could
+      # use it. If a triggering command include a user_inputs field, then the triggered workflow
+      # must include this very input, otherwise the workflow won't be called.
+      # See start_full_benchmarks.yml as example.
+      user_inputs:
+        description: "Type of benchmarks to run"
+        type: string
+        default: "weekly_benchmarks"
 
 env:
   CARGO_TERM_COLOR: always
@@ -13,36 +34,13 @@ env:
   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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        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 }}
+  integer-benchmarks:
+    name: Execute integer benchmarks for all operations flavor
+    runs-on: ${{ github.event.inputs.runner_name }}
     timeout-minutes: 1440 # 24 hours
+    if: ${{ !cancelled() }}
     continue-on-error: true
     strategy:
       fail-fast: false
@@ -54,25 +52,19 @@ jobs:
         include:
           - os: ubuntu-22.04
             cuda: "12.2"
-            gcc: 11
+            gcc: 9
     env:
       CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
     steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
+      - name: Instance configuration used
         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 "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
 
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           fetch-depth: 0
 
@@ -90,7 +82,7 @@ jobs:
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: nightly
 
@@ -115,15 +107,11 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Check device is detected
-        if: ${{ !cancelled() }}
-        run: nvidia-smi
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
 
       - name: Run benchmarks with AVX512
         run: |
@@ -133,7 +121,7 @@ jobs:
         run: |
           python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
           --database tfhe_rs \
-          --hardware "n3-H100x1" \
+          --hardware ${{ inputs.instance_type }} \
           --backend gpu \
           --project-version "${{ env.COMMIT_HASH }}" \
           --branch ${{ github.ref_name }} \
@@ -144,7 +132,7 @@ jobs:
           --throughput
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
         with:
           name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
           path: ${{ env.RESULTS_FILENAME }}
@@ -163,39 +151,19 @@ jobs:
           -d @${{ env.RESULTS_FILENAME }} \
           ${{ secrets.SLAB_URL }}
 
-  slack-notify:
+  slack-notification:
     name: Slack Notification
-    needs: [ setup-instance, cuda-integer-full-benchmarks ]
-    runs-on: ubuntu-latest
+    runs-on: ${{ github.event.inputs.runner_name }}
     if: ${{ !success() && !cancelled() }}
-    continue-on-error: true
+    needs: integer-benchmarks
     steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ needs.cuda-integer-full-benchmarks.result }}
-          SLACK_MESSAGE: "Integer GPU full benchmarks finished with status: ${{ needs.cuda-integer-full-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-integer-full-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-integer-full-benchmarks, slack-notify ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        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() }}
+      - name: Notify
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-integer-full-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "Integer GPU full benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/integer_multi_bit_benchmark.yml

@@ -0,0 +1,130 @@
+# Run integer benchmarks with multi-bit cryptographic parameters on an AWS instance and return parsed results to Slab CI bot.
+name: Integer Multi-bit benchmarks
+
+on:
+  workflow_dispatch:
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
+
+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"
+
+jobs:
+  run-integer-benchmarks:
+    name: Execute integer multi-bit benchmarks in EC2
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ !cancelled() }}
+    steps:
+      - name: Instance configuration used
+        run: |
+          echo "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Get benchmark date
+        run: |
+          echo "BENCH_DATE=$(date --iso-8601=seconds)" >> "${GITHUB_ENV}"
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          fetch-depth: 0
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
+        with:
+          toolchain: nightly
+
+      - name: Run multi-bit benchmarks with AVX512
+        run: |
+          make FAST_BENCH=TRUE bench_integer_multi_bit
+
+      - 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@5d5d22a31266ced268874388b861e4b58bb5c2f3
+        with:
+          name: ${{ github.sha }}_csv_integer
+          path: ${{ env.PARSE_INTEGER_BENCH_CSV_FILE }}
+
+      - name: Parse results
+        run: |
+          COMMIT_DATE="$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})"
+          COMMIT_HASH="$(git describe --tags --dirty)"
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware ${{ inputs.instance_type }} \
+          --project-version "${COMMIT_HASH}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${COMMIT_DATE}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --throughput
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
+        with:
+          name: ${{ github.sha }}_integer
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          echo "Computing HMac on results file"
+          SIGNATURE="$(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@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "Integer benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/integer_multi_bit_gpu_benchmark.yml

@@ -1,21 +1,24 @@
-# Run integer benchmarks with multi-bit cryptographic parameters on an instance and return parsed results to Slab CI bot.
+# Run integer benchmarks with multi-bit cryptographic parameters on an AWS 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'
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
 
 env:
   CARGO_TERM_COLOR: always
@@ -24,38 +27,13 @@ env:
   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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        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 }}
+  cuda-integer-benchmarks:
+    name: Execute integer multi-bit benchmarks in EC2
+    runs-on: ${{ github.event.inputs.runner_name }}
     timeout-minutes: 1440 # 24 hours
+    if: ${{ !cancelled() }}
     strategy:
       fail-fast: false
       # explicit include-based build matrix, of known valid options
@@ -63,43 +41,33 @@ jobs:
         include:
           - os: ubuntu-22.04
             cuda: "12.2"
-            gcc: 11
+            gcc: 9
     env:
       CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
     steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
+      - name: Instance configuration used
         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 "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Get benchmark date
+        run: |
+          echo "BENCH_DATE=$(date --iso-8601=seconds)" >> "${GITHUB_ENV}"
 
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         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)";
-          } >> "${GITHUB_ENV}"
-
       - name: Set up home
         # "Install rust" step require root user to have a HOME directory which is not set.
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: nightly
 
@@ -123,23 +91,9 @@ jobs:
             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
+          make FAST_BENCH=TRUE BENCH_OP_FLAVOR=default bench_integer_multi_bit_gpu
 
       - name: Parse benchmarks to csv
         run: |
@@ -147,37 +101,39 @@ jobs:
             parse_integer_benches
 
       - name: Upload csv results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
         with:
           name: ${{ github.sha }}_csv_integer
           path: ${{ env.PARSE_INTEGER_BENCH_CSV_FILE }}
 
       - name: Parse results
         run: |
+          COMMIT_DATE="$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})"
+          COMMIT_HASH="$(git describe --tags --dirty)"
           python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
           --database tfhe_rs \
-          --hardware "n3-H100x1" \
+          --hardware ${{ inputs.instance_type }} \
           --backend gpu \
-          --project-version "${{ env.COMMIT_HASH }}" \
+          --project-version "${COMMIT_HASH}" \
           --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
+          --commit-date "${COMMIT_DATE}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
           --name-suffix avx512 \
           --throughput
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
         with:
           name: ${{ github.sha }}_integer
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -193,40 +149,14 @@ jobs:
           -d @${{ env.RESULTS_FILENAME }} \
           ${{ secrets.SLAB_URL }}
 
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-integer-multi-bit-benchmarks ]
-    runs-on: ubuntu-latest
-    if: ${{ !success() && !cancelled() }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ needs.cuda-integer-multi-bit-benchmarks.result }}
-          SLACK_MESSAGE: "Integer GPU multi-bit benchmarks finished with status: ${{ needs.cuda-integer-multi-bit-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-integer-full-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-integer-multi-bit-benchmarks, slack-notify ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        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() }}
+        if: ${{ !success() && !cancelled() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-integer-multi-bit-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "Integer GPU benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/integer_multi_bit_multi_gpu_benchmark.yml

@@ -1,221 +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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: hyperstack
-          profile: multi-h100
-
-  cuda-integer-multi-bit-multi-gpu-benchmarks:
-    name: Execute multi GPU integer multi-bit benchmarks
-    needs: setup-instance
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    timeout-minutes: 1440 # 24 hours
-    continue-on-error: true
-    strategy:
-      fail-fast: false
-      max-parallel: 1
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
-    steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
-        run: |
-          sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev
-          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          cd cmake-${{ env.CMAKE_VERSION }}
-          ./bootstrap
-          make -j"$(nproc)"
-          sudo make install
-
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Get benchmark details
-        run: |
-          {
-            echo "BENCH_DATE=$(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)";
-          } >> "${GITHUB_ENV}"
-
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${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@692973e3d937129bcbf40652eb9f2f61becf3332
-        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@834a144ee995460fba8ed112a2fc961b36a5ec5a
-        with:
-          name: ${{ github.sha }}_integer
-          path: ${{ env.RESULTS_FILENAME }}
-
-      - name: Send data to Slab
-        shell: bash
-        run: |
-          echo "Computing HMac on results file"
-          SIGNATURE="$(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 }}
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-integer-multi-bit-multi-gpu-benchmarks ]
-    runs-on: ubuntu-latest
-    if: ${{ !success() && !cancelled() }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ needs.cuda-integer-multi-bit-multi-gpu-benchmarks.result }}
-          SLACK_MESSAGE: "Integer multi GPU multi-bit benchmarks finished with status: ${{ needs.cuda-integer-multi-bit-multi-gpu-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-integer-multi-bit-multi-gpu-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-integer-multi-bit-multi-gpu-benchmarks ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-integer-multi-bit-multi-gpu-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/integer_multi_gpu_full_benchmark.yml

@@ -1,201 +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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        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 for all operations flavor
-    needs: setup-instance
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    timeout-minutes: 1440 # 24 hours
-    continue-on-error: true
-    strategy:
-      fail-fast: false
-      max-parallel: 1
-      matrix:
-        command: [integer, integer_multi_bit]
-        op_flavor: [default, unchecked]
-        # explicit include-based build matrix, of known valid options
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
-    steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
-        run: |
-          sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev
-          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          cd cmake-${{ env.CMAKE_VERSION }}
-          ./bootstrap
-          make -j"$(nproc)"
-          sudo make install
-
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Get benchmark details
-        run: |
-          {
-            echo "BENCH_DATE=$(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)";
-          } >> "${GITHUB_ENV}"
-
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${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@692973e3d937129bcbf40652eb9f2f61becf3332
-        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@834a144ee995460fba8ed112a2fc961b36a5ec5a
-        with:
-          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
-          path: ${{ env.RESULTS_FILENAME }}
-
-      - name: Send data to Slab
-        shell: bash
-        run: |
-          echo "Computing HMac on results file"
-          SIGNATURE="$(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 }}
-
-  slack-notify:
-    name: Slack Notification
-    needs: [ setup-instance, cuda-integer-full-multi-gpu-benchmarks ]
-    runs-on: ubuntu-latest
-    if: ${{ !success() && !cancelled() }}
-    continue-on-error: true
-    steps:
-      - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ needs.cuda-integer-full-multi-gpu-benchmarks.result }}
-          SLACK_MESSAGE: "Integer GPU full benchmarks finished with status: ${{ needs.cuda-integer-full-multi-gpu-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (cuda-integer-full-multi-gpu-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-integer-full-multi-gpu-benchmarks ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-integer-full-multi-gpu-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/m1_tests.yml

@@ -3,7 +3,7 @@ name: Tests on M1 CPU
 on:
   workflow_dispatch:
   pull_request:
-    types: [ labeled ]
+    types: [labeled]
   # Have a nightly build for M1 tests
   schedule:
     # * is a special character in YAML so you have to quote this string
@@ -18,9 +18,6 @@ env:
   RUST_MIN_STACK: "8388608"
   ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
   FAST_TESTS: "TRUE"
-  # We clear the cache to reduce memory pressure because of the numerous processes of cargo
-  # nextest
-  TFHE_RS_CLEAR_IN_MEMORY_KEY_CACHE: "1"
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.head_ref }}
@@ -34,12 +31,10 @@ jobs:
     timeout-minutes: 720
 
     steps:
-      - uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          persist-credentials: 'false'
+      - uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: stable
 
@@ -79,10 +74,6 @@ jobs:
         run: |
           make test_concrete_csprng
 
-      - name: Run tfhe-zk-pok tests
-        run: |
-          make test_zk_pok
-
       - name: Run core tests
         run: |
           make test_core_crypto
@@ -91,13 +82,6 @@ jobs:
         run: |
           make test_boolean
 
-      # Because we do "illegal" things with the build system which Cargo does not seem to like much
-      # we need to clear the cache to make sure the C API is built properly and does not use a stale
-      # cached version
-      - name: Clear build cache
-        run: |
-          cargo clean
-
       - name: Run C API tests
         run: |
           make test_c_api
@@ -149,7 +133,7 @@ jobs:
       - name: Slack Notification
         if: ${{ needs.cargo-builds.result != 'skipped' }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ needs.cargo-builds.result }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/make_release.yml

@@ -20,72 +20,20 @@ on:
         description: "Push node js package"
         type: boolean
         default: true
-      npm_latest_tag:
-        description: "Set NPM tag as latest"
-        type: boolean
-        default: false
 
 env:
   ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  NPM_TAG: ""
 
 jobs:
-  package:
-    runs-on: ubuntu-latest
-    outputs:
-      hash: ${{ steps.hash.outputs.hash }}
-    steps:
-      - name: Checkout
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-      - name: Prepare package
-        run: |
-          cargo package -p tfhe
-      - uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a # v4.3.6
-        with:
-          name: crate
-          path: target/package/*.crate
-      - name: generate hash
-        id: hash
-        run: cd target/package && echo "hash=$(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 Release
-    needs: [package] # for comparing hashes
     runs-on: ubuntu-latest
-    permissions:
-      contents: read
-      id-token: write
     steps:
       - name: Checkout
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           fetch-depth: 0
-      - name: Create NPM version tag
-        if: ${{ inputs.npm_latest_tag }}
-        run: |
-          echo "NPM_TAG=latest" >> "${GITHUB_ENV}"
-      - name: Download artifact
-        uses: actions/download-artifact@fa0a91b85d4f404e444e00e005971372dc801d16 # v4.1.8
-        with:
-          name: crate
-          path: target/package
+
       - name: Publish crate.io package
         if: ${{ inputs.push_to_crates }}
         env:
@@ -94,36 +42,18 @@ jobs:
         run: |
           cargo publish -p tfhe --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@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: failure
-          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-          SLACK_MESSAGE: "SLSA tfhe crate - hash comparison failure: (${{ env.ACTION_RUN_URL }})"
-          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-
       - name: Build web package
         if: ${{ inputs.push_web_package }}
         run: |
-          make build_web_js_api_parallel
+          make build_web_js_api
 
       - name: Publish web package
         if: ${{ inputs.push_web_package }}
-        uses: JS-DevTools/npm-publish@19c28f1ef146469e409470805ea4279d47c3d35c
+        uses: JS-DevTools/npm-publish@4b07b26a2f6e0a51846e1870223e545bae91c552
         with:
           token: ${{ secrets.NPM_TOKEN }}
           package: tfhe/pkg/package.json
           dry-run: ${{ inputs.dry_run }}
-          tag: ${{ env.NPM_TAG }}
-          provenance: true
 
       - name: Build Node package
         if: ${{ inputs.push_node_package }}
@@ -135,18 +65,16 @@ jobs:
 
       - name: Publish Node package
         if: ${{ inputs.push_node_package }}
-        uses: JS-DevTools/npm-publish@19c28f1ef146469e409470805ea4279d47c3d35c
+        uses: JS-DevTools/npm-publish@4b07b26a2f6e0a51846e1870223e545bae91c552
         with:
           token: ${{ secrets.NPM_TOKEN }}
           package: tfhe/pkg/package.json
           dry-run: ${{ inputs.dry_run }}
-          tag: ${{ env.NPM_TAG }}
-          provenance: true
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/make_release_concrete_csprng.yml

@@ -18,7 +18,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           fetch-depth: 0
 
@@ -32,7 +32,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/make_release_cuda.yml

@@ -1,129 +0,0 @@
-# Publish new release of tfhe-cuda-backend on crates.io.
-name: Publish CUDA release
-
-on:
-  workflow_dispatch:
-    inputs:
-      dry_run:
-        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 }}
-  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 (publish-cuda-release)
-    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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: gpu-test
-
-  publish-cuda-release:
-    name: Publish CUDA Release
-    needs: setup-instance
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    strategy:
-      fail-fast: false
-      # explicit include-based build matrix, of known valid options
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 9
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-    steps:
-      - name: Checkout
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install latest stable
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
-        with:
-          toolchain: stable
-
-      - name: Export CUDA variables
-        if: ${{ !cancelled() }}
-        run: |
-          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
-          {
-            echo "CUDA_PATH=$CUDA_PATH";
-            echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH";
-            echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc";
-          } >> "${GITHUB_ENV}"
-
-      # 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: 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: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "tfhe-cuda-backend release finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (publish-release)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, publish-cuda-release ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (publish-cuda-release) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/make_release_zk_pok.yml

@@ -1,42 +0,0 @@
-# Publish new release of tfhe-zk-pok on crates.io.
-name: Publish tfhe-zk-pok 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:
-  publish_release:
-    name: Publish tfhe-zk-pok Release
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        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-zk-pok --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-          SLACK_MESSAGE: "tfhe-zk-pok release failed: (${{ env.ACTION_RUN_URL }})"
-          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/parameters_check.yml

@@ -14,17 +14,17 @@ on:
 
 jobs:
   params-curves-security-check:
-    runs-on: large_ubuntu_16
+    runs-on: ubuntu-latest
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
 
       - name: Checkout lattice-estimator
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           repository: malb/lattice-estimator
           path: lattice_estimator
-          ref: 'e80ec6bbbba212428b0e92d0467c18629cf9ed67'
+          ref: '53508253629d3b5d31a2ad110e85dc69391ccb95'
 
       - name: Install Sage
         run: |
@@ -42,7 +42,7 @@ jobs:
       - name: Slack Notification
         if: ${{ always() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/shortint_benchmark.yml

@@ -0,0 +1,128 @@
+# Run shortint benchmarks on an AWS instance and return parsed results to Slab CI bot.
+name: Shortint benchmarks
+
+on:
+  workflow_dispatch:
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
+
+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"
+
+jobs:
+  run-shortint-benchmarks:
+    name: Execute shortint benchmarks in EC2
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ !cancelled() }}
+    steps:
+      - name: Instance configuration used
+        run: |
+          echo "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Get benchmark date
+        run: |
+          echo "BENCH_DATE=$(date --iso-8601=seconds)" >> "${GITHUB_ENV}"
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          fetch-depth: 0
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
+        with:
+          toolchain: nightly
+
+      - name: Run benchmarks with AVX512
+        run: |
+          make bench_shortint
+
+      - name: Parse results
+        run: |
+          COMMIT_DATE="$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})"
+          COMMIT_HASH="$(git describe --tags --dirty)"
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware ${{ inputs.instance_type }} \
+          --project-version "${COMMIT_HASH}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${COMMIT_DATE}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --throughput
+
+      - name: Measure key sizes
+        run: |
+          make measure_shortint_key_sizes
+
+      - name: Parse key sizes results
+        run: |
+          python3 ./ci/benchmark_parser.py tfhe/shortint_key_sizes.csv ${{ env.RESULTS_FILENAME }} \
+          --key-sizes \
+          --append-results
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
+        with:
+          name: ${{ github.sha }}_shortint
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          echo "Computing HMac on results file"
+          SIGNATURE="$(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@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "Shortint benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/shortint_cpu_benchmark.yml

@@ -1,193 +0,0 @@
-# Run all shortint benchmarks on an AWS instance and return parsed results to Slab CI bot.
-name: Shortint full benchmarks
-
-on:
-  workflow_dispatch:
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 1a.m.
-    - cron: '0 1 * * 6'
-    # Quarterly benchmarks will be triggered right before end of quarter, the 25th of the current month at 4a.m.
-    # These benchmarks are far longer to execute hence the reason to run them only four time a year.
-    - cron: '0 4 25 MAR,JUN,SEP,DEC *'
-
-
-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:
-  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:
-      op_flavor: ${{ steps.set_op_flavor.outputs.op_flavor }}
-    steps:
-      - name: Weekly benchmarks
-        if: github.event_name == 'workflow_dispatch' ||
-          github.event.schedule == '0 1 * * 6'
-        run: |
-          echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
-
-      - name: Quarterly benchmarks
-        if: github.event.schedule == '0 4 25 MAR,JUN,SEP,DEC *'
-        run: |
-          echo "OP_FLAVOR=[\"default\", \"smart\", \"unchecked\"]" >> "${GITHUB_ENV}"
-
-      - name: Set operation flavor output
-        id: set_op_flavor
-        run: |
-          echo "op_flavor=${{ toJSON(env.OP_FLAVOR) }}" >> "${GITHUB_OUTPUT}"
-
-  setup-instance:
-    name: Setup instance (shortint-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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: bench
-
-  shortint-benchmarks:
-    name: Execute shortint benchmarks for all operations flavor
-    needs: [ prepare-matrix, setup-instance ]
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    continue-on-error: true
-    strategy:
-      max-parallel: 1
-      matrix:
-        op_flavor: ${{ fromJson(needs.prepare-matrix.outputs.op_flavor) }}
-    steps:
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Get benchmark details
-        run: |
-          {
-            echo "BENCH_DATE=$(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)";
-          } >> "${GITHUB_ENV}"
-
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
-        with:
-          toolchain: nightly
-
-      - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          repository: zama-ai/slab
-          path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Run benchmarks with AVX512
-        run: |
-          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_shortint
-
-      - name: Parse results
-        run: |
-          COMMIT_DATE="$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})"
-          COMMIT_HASH="$(git describe --tags --dirty)"
-          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
-          --database tfhe_rs \
-          --hardware "hpc7a.96xlarge" \
-          --project-version "${COMMIT_HASH}" \
-          --branch ${{ github.ref_name }} \
-          --commit-date "${COMMIT_DATE}" \
-          --bench-date "${{ env.BENCH_DATE }}" \
-          --walk-subdirs \
-          --name-suffix avx512 \
-          --throughput
-
-      # This small benchmark needs to be executed only once.
-      - name: Measure key sizes
-        if: matrix.op_flavor == 'default'
-        run: |
-          make measure_shortint_key_sizes
-
-      - name: Parse key sizes results
-        if: matrix.op_flavor == 'default'
-        run: |
-          python3 ./ci/benchmark_parser.py tfhe/shortint_key_sizes.csv ${{ env.RESULTS_FILENAME }} \
-          --key-sizes \
-          --append-results
-
-      - name: Upload parsed results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
-        with:
-          name: ${{ github.sha }}_shortint_${{ matrix.op_flavor }}
-          path: ${{ env.RESULTS_FILENAME }}
-
-      - name: Send data to Slab
-        shell: bash
-        run: |
-          echo "Computing HMac on results file"
-          SIGNATURE="$(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@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Shortint full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (shortint-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, shortint-benchmarks ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (shortint-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/shortint_full_benchmark.yml

@@ -0,0 +1,152 @@
+# Run all shortint benchmarks on an AWS instance and return parsed results to Slab CI bot.
+name: Shortint full benchmarks
+
+on:
+  workflow_dispatch:
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
+      # This input is not used in this workflow but still mandatory since a calling workflow could
+      # use it. If a triggering command include a user_inputs field, then the triggered workflow
+      # must include this very input, otherwise the workflow won't be called.
+      # See start_full_benchmarks.yml as example.
+      user_inputs:
+        description: "Type of benchmarks to run"
+        type: string
+        default: "weekly_benchmarks"
+
+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"
+
+jobs:
+  shortint-benchmarks:
+    name: Execute shortint benchmarks for all operations flavor
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ !cancelled() }}
+    strategy:
+      max-parallel: 1
+      matrix:
+        op_flavor: [ default, smart, unchecked ]
+    steps:
+      - name: Instance configuration used
+        run: |
+          echo "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          fetch-depth: 0
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(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)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
+        with:
+          toolchain: nightly
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
+
+      - name: Run benchmarks with AVX512
+        run: |
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_shortint
+
+      - name: Parse results
+        run: |
+          COMMIT_DATE="$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})"
+          COMMIT_HASH="$(git describe --tags --dirty)"
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware ${{ inputs.instance_type }} \
+          --project-version "${COMMIT_HASH}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${COMMIT_DATE}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --throughput
+
+      # This small benchmark needs to be executed only once.
+      - name: Measure key sizes
+        if: matrix.op_flavor == 'default'
+        run: |
+          make measure_shortint_key_sizes
+
+      - name: Parse key sizes results
+        if: matrix.op_flavor == 'default'
+        run: |
+          python3 ./ci/benchmark_parser.py tfhe/shortint_key_sizes.csv ${{ env.RESULTS_FILENAME }} \
+          --key-sizes \
+          --append-results
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
+        with:
+          name: ${{ github.sha }}_shortint_${{ matrix.op_flavor }}
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          echo "Computing HMac on results file"
+          SIGNATURE="$(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 }}
+
+  slack-notification:
+    name: Slack Notification
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ failure() }}
+    needs: shortint-benchmarks
+    steps:
+      - name: Notify
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "Shortint full benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/signed_integer_benchmark.yml

@@ -0,0 +1,130 @@
+# Run signed integer benchmarks on an AWS instance and return parsed results to Slab CI bot.
+name: Signed Integer benchmarks
+
+on:
+  workflow_dispatch:
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
+
+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"
+
+jobs:
+  run-integer-benchmarks:
+    name: Execute signed integer benchmarks in EC2
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ !cancelled() }}
+    steps:
+      - name: Instance configuration used
+        run: |
+          echo "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Get benchmark date
+        run: |
+          echo "BENCH_DATE=$(date --iso-8601=seconds)" >> "${GITHUB_ENV}"
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          fetch-depth: 0
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
+        with:
+          toolchain: nightly
+
+      - name: Run benchmarks with AVX512
+        run: |
+          make FAST_BENCH=TRUE bench_signed_integer
+
+      - 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@5d5d22a31266ced268874388b861e4b58bb5c2f3
+        with:
+          name: ${{ github.sha }}_csv_integer
+          path: ${{ env.PARSE_INTEGER_BENCH_CSV_FILE }}
+
+      - name: Parse results
+        run: |
+          COMMIT_DATE="$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})"
+          COMMIT_HASH="$(git describe --tags --dirty)"
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware ${{ inputs.instance_type }} \
+          --project-version "${COMMIT_HASH}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${COMMIT_DATE}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --throughput
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
+        with:
+          name: ${{ github.sha }}_integer
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          echo "Computing HMac on results file"
+          SIGNATURE="$(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@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "Signed integer benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/signed_integer_cpu_benchmark.yml

@@ -1,191 +0,0 @@
-# Run all signed integer benchmarks on an AWS instance and return parsed results to Slab CI bot.
-name: Signed Integer full benchmarks
-
-on:
-  workflow_dispatch:
-    inputs:
-      all_precisions:
-        description: "Run all precisions"
-        type: boolean
-        default: false
-
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 1a.m.
-    - cron: '0 1 * * 6'
-    # Quarterly benchmarks will be triggered right before end of quarter, the 25th of the current month at 4a.m.
-    # These benchmarks are far longer to execute hence the reason to run them only four time a year.
-    - cron: '0 4 25 MAR,JUN,SEP,DEC *'
-
-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
-    if: github.event_name != 'schedule' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
-    outputs:
-      op_flavor: ${{ steps.set_op_flavor.outputs.op_flavor }}
-    steps:
-      - name: Weekly benchmarks
-        if: github.event_name == 'workflow_dispatch' ||
-          github.event.schedule == '0 1 * * 6'
-        run: |
-          echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
-
-      - name: Quarterly benchmarks
-        if: github.event.schedule == '0 4 25 MAR,JUN,SEP,DEC *'
-        run: |
-          echo "OP_FLAVOR=[\"default\", \"unchecked\"]" >> "${GITHUB_ENV}"
-
-      - name: Set operation flavor output
-        id: set_op_flavor
-        run: |
-          echo "op_flavor=${{ toJSON(env.OP_FLAVOR) }}" >> "${GITHUB_OUTPUT}"
-
-  setup-instance:
-    name: Setup instance (signed-integer-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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: bench
-
-  signed-integer-benchmarks:
-    name: Execute signed integer benchmarks for all operations flavor
-    needs: [ prepare-matrix, setup-instance ]
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    concurrency:
-      group: ${{ github.workflow }}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    continue-on-error: true
-    timeout-minutes: 1440  # 24 hours
-    strategy:
-      max-parallel: 1
-      matrix:
-        command: [ integer, integer_multi_bit ]
-        op_flavor: [ default, unchecked ]
-    steps:
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Get benchmark details
-        run: |
-          {
-            echo "BENCH_DATE=$(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)";
-          } >> "${GITHUB_ENV}"
-
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
-        with:
-          toolchain: nightly
-
-      - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        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: Run benchmarks with AVX512
-        run: |
-          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_signed_${{ matrix.command }}
-
-      - name: Parse results
-        run: |
-          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
-          --database tfhe_rs \
-          --hardware "hpc7a.96xlarge" \
-          --project-version "${{ env.COMMIT_HASH }}" \
-          --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
-          --bench-date "${{ env.BENCH_DATE }}" \
-          --walk-subdirs \
-          --name-suffix avx512 \
-          --throughput
-
-      - name: Upload parsed results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
-        with:
-          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
-          path: ${{ env.RESULTS_FILENAME }}
-
-      - name: Send data to Slab
-        shell: bash
-        run: |
-          echo "Computing HMac on results file"
-          SIGNATURE="$(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@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Signed integer full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (integer-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, signed-integer-benchmarks ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (signed-integer-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/signed_integer_full_benchmark.yml

@@ -0,0 +1,136 @@
+# Run all signed integer benchmarks on an AWS instance and return parsed results to Slab CI bot.
+name: Signed Integer full benchmarks
+
+on:
+  workflow_dispatch:
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
+      user_inputs:
+        description: "Type of benchmarks to run"
+        type: string
+        default: "weekly_benchmarks"
+
+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"
+
+jobs:
+  integer-benchmarks:
+    name: Execute signed integer benchmarks for all operations flavor
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ !cancelled() }}
+    continue-on-error: true
+    timeout-minutes: 1440  # 24 hours
+    strategy:
+      max-parallel: 1
+      matrix:
+        command: [ integer, integer_multi_bit ]
+        op_flavor: [ default, unchecked ]
+    steps:
+      - name: Instance configuration used
+        run: |
+          echo "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          fetch-depth: 0
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(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)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
+        with:
+          toolchain: nightly
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
+
+      - name: Run benchmarks with AVX512
+        run: |
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_signed_${{ matrix.command }}
+
+      - name: Parse results
+        run: |
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware ${{ inputs.instance_type }} \
+          --project-version "${{ 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@5d5d22a31266ced268874388b861e4b58bb5c2f3
+        with:
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          echo "Computing HMac on results file"
+          SIGNATURE="$(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 }}
+
+  slack-notification:
+    name: Slack Notification
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ failure() }}
+    needs: integer-benchmarks
+    steps:
+      - name: Notify
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "Signed integer full benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/signed_integer_multi_bit_benchmark.yml

@@ -0,0 +1,130 @@
+# Run signed integer benchmarks with multi-bit cryptographic parameters on an AWS instance and return parsed results to Slab CI bot.
+name: Signed Integer Multi-bit benchmarks
+
+on:
+  workflow_dispatch:
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
+
+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"
+
+jobs:
+  run-integer-benchmarks:
+    name: Execute signed integer multi-bit benchmarks in EC2
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ !cancelled() }}
+    steps:
+      - name: Instance configuration used
+        run: |
+          echo "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Get benchmark date
+        run: |
+          echo "BENCH_DATE=$(date --iso-8601=seconds)" >> "${GITHUB_ENV}"
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          fetch-depth: 0
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
+        with:
+          toolchain: nightly
+
+      - name: Run multi-bit benchmarks with AVX512
+        run: |
+          make FAST_BENCH=TRUE bench_signed_integer_multi_bit
+
+      - 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@5d5d22a31266ced268874388b861e4b58bb5c2f3
+        with:
+          name: ${{ github.sha }}_csv_integer
+          path: ${{ env.PARSE_INTEGER_BENCH_CSV_FILE }}
+
+      - name: Parse results
+        run: |
+          COMMIT_DATE="$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})"
+          COMMIT_HASH="$(git describe --tags --dirty)"
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware ${{ inputs.instance_type }} \
+          --project-version "${COMMIT_HASH}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${COMMIT_DATE}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --throughput
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
+        with:
+          name: ${{ github.sha }}_integer
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          echo "Computing HMac on results file"
+          SIGNATURE="$(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@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "Signed integer benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/start_benchmarks.yml

@@ -0,0 +1,135 @@
+# Start all benchmark jobs on Slab CI bot.
+name: Start all benchmarks
+
+on:
+  push:
+    branches:
+      - "main"
+  workflow_dispatch:
+    inputs:
+      # The input name must be the name of the slab command to launch
+      boolean_bench:
+        description: "Run Boolean benches"
+        type: boolean
+        default: true
+      shortint_bench:
+        description: "Run shortint benches"
+        type: boolean
+        default: true
+      integer_bench:
+        description: "Run integer benches"
+        type: boolean
+        default: true
+      signed_integer_bench:
+        description: "Run signed integer benches"
+        type: boolean
+        default: true
+      integer_multi_bit_bench:
+        description: "Run integer multi bit benches"
+        type: boolean
+        default: true
+      signed_integer_multi_bit_bench:
+        description: "Run signed integer multi bit benches"
+        type: boolean
+        default: true
+      core_crypto_bench:
+        description: "Run core crypto benches"
+        type: boolean
+        default: true
+      core_crypto_gpu_bench:
+        description: "Run core crypto benches on GPU"
+        type: boolean
+        default: true
+      wasm_client_bench:
+        description: "Run WASM client benches"
+        type: boolean
+        default: true
+
+jobs:
+  start-benchmarks:
+    if: ${{ (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs') || github.event_name == 'workflow_dispatch' }}
+    strategy:
+      matrix:
+        command: [ boolean_bench, shortint_bench,
+                   integer_bench, integer_multi_bit_bench,
+                   signed_integer_bench, signed_integer_multi_bit_bench,
+                   integer_gpu_bench, integer_multi_bit_gpu_bench,
+                   core_crypto_bench, core_crypto_gpu_bench, wasm_client_bench ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          fetch-depth: 0
+
+      - name: Check for file changes
+        id: changed-files
+        uses: tj-actions/changed-files@aa08304bd477b800d468db44fe10f6c61f7f7b11
+        with:
+          files_yaml: |
+            common_benches:
+              - toolchain.txt
+              - Makefile
+              - ci/slab.toml
+              - tfhe/Cargo.toml
+              - tfhe/src/core_crypto/**
+              - .github/workflows/start_benchmarks.yml
+            boolean_bench:
+              - tfhe/src/boolean/**
+              - tfhe/benches/boolean/**
+              - .github/workflows/boolean_benchmark.yml
+            shortint_bench:
+              - tfhe/src/shortint/**
+              - tfhe/benches/shortint/**
+              - .github/workflows/shortint_benchmark.yml
+            integer_bench:
+              - tfhe/src/shortint/**
+              - tfhe/src/integer/**
+              - tfhe/benches/integer/bench.rs
+              - .github/workflows/integer_benchmark.yml
+            integer_multi_bit_bench:
+              - tfhe/src/shortint/**
+              - tfhe/src/integer/**
+              - tfhe/benches/integer/bench.rs
+              - .github/workflows/integer_multi_bit_benchmark.yml
+            signed_integer_bench:
+              - tfhe/src/shortint/**
+              - tfhe/src/integer/**
+              - tfhe/benches/integer/signed_bench.rs
+              - .github/workflows/signed_integer_benchmark.yml
+            signed_integer_multi_bit_bench:
+              - tfhe/src/shortint/**
+              - tfhe/src/integer/**
+              - tfhe/benches/integer/signed_bench.rs
+              - .github/workflows/signed_integer_multi_bit_benchmark.yml
+            core_crypto_bench:
+              - tfhe/src/core_crypto/**
+              - tfhe/benches/core_crypto/**
+              - .github/workflows/core_crypto_benchmark.yml
+            wasm_client_bench:
+              - tfhe/web_wasm_parallel_tests/**
+              - .github/workflows/wasm_client_benchmark.yml
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
+
+      - name: Start AWS job in Slab
+        # If manually triggered check that the current bench has been requested
+        # Otherwise if it's on push check that files relevant to benchmarks have changed
+        if: (github.event_name == 'workflow_dispatch' && github.event.inputs[matrix.command] == 'true') || (github.event_name == 'push' && (steps.changed-files.outputs.common_benches_any_changed == 'true' || steps.changed-files.outputs[format('{0}_any_changed', matrix.command)] == 'true'))
+        shell: bash
+        run: |
+          echo -n '{"command": "${{ matrix.command }}", "git_ref": "${{ github.ref }}", "sha": "${{ github.sha }}"}' > command.json
+          SIGNATURE="$(slab/scripts/hmac_calculator.sh command.json '${{ secrets.JOB_SECRET }}')"
+          curl -v -k \
+          --fail-with-body \
+          -H "Content-Type: application/json" \
+          -H "X-Slab-Repository: ${{ github.repository }}" \
+          -H "X-Slab-Command: start_aws" \
+          -H "X-Hub-Signature-256: sha256=${SIGNATURE}" \
+          -d @command.json \
+          ${{ secrets.SLAB_URL }}

.github/workflows/start_full_benchmarks.yml

@@ -0,0 +1,66 @@
+# Start all benchmark jobs, including full shortint and integer, on Slab CI bot.
+name: Start full suite benchmarks
+
+on:
+  schedule:
+    # Weekly benchmarks will be triggered each Saturday at 1a.m.
+    - cron: '0 1 * * 6'
+    # Quarterly benchmarks will be triggered right before end of quarter, the 25th of the current month at 4a.m.
+    # These benchmarks are far longer to execute hence the reason to run them only four time a year.
+    - cron: '0 4 25 MAR,JUN,SEP,DEC *'
+  workflow_dispatch:
+    inputs:
+      benchmark_type:
+        description: 'Benchmark type'
+        required: true
+        default: 'weekly'
+        type: choice
+        options:
+          - weekly
+          - quarterly
+
+jobs:
+  start-benchmarks:
+    if: ${{ (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') || github.event_name == 'workflow_dispatch' }}
+    strategy:
+      matrix:
+        command: [ boolean_bench, shortint_full_bench,
+                   integer_full_bench, signed_integer_full_bench, integer_gpu_full_bench,
+                   core_crypto_bench, core_crypto_gpu_bench, wasm_client_bench ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          fetch-depth: 0
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
+
+      - name: Set benchmarks type as weekly
+        if: (github.event_name == 'workflow_dispatch' && inputs.benchmark_type == 'weekly') || github.event.schedule == '0 1 * * 6'
+        run: |
+          echo "BENCH_TYPE=weekly_benchmarks" >> "${GITHUB_ENV}"
+
+      - name: Set benchmarks type as quarterly
+        if: (github.event_name == 'workflow_dispatch' && inputs.benchmark_type == 'quarterly') || github.event.schedule == '0 4 25 MAR,JUN,SEP,DEC *'
+        run: |
+          echo "BENCH_TYPE=quarterly_benchmarks" >> "${GITHUB_ENV}"
+
+      - name: Start AWS job in Slab
+        shell: bash
+        run: |
+          echo -n '{"command": "${{ matrix.command }}", "git_ref": "${{ github.ref }}", "sha": "${{ github.sha }}", "user_inputs": "${{ env.BENCH_TYPE }}"}' > command.json
+          SIGNATURE="$(slab/scripts/hmac_calculator.sh command.json '${{ secrets.JOB_SECRET }}')"
+          curl -v -k \
+          --fail-with-body \
+          -H "Content-Type: application/json" \
+          -H "X-Slab-Repository: ${{ github.repository }}" \
+          -H "X-Slab-Command: start_aws" \
+          -H "X-Hub-Signature-256: sha256=${SIGNATURE}" \
+          -d @command.json \
+          ${{ secrets.SLAB_URL }}

.github/workflows/sync_on_push.yml

@@ -13,20 +13,25 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           fetch-depth: 0
+      - name: Save repo
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
+        with:
+          name: repo-archive
+          path: '.'
       - name: git-sync
         uses: wei/git-sync@55c6b63b4f21607da0e9877ca9b4d11a29fc6d83
         with:
           source_repo: "zama-ai/tfhe-rs"
           source_branch: "main"
-          destination_repo: "https://${{ secrets.BOT_USERNAME }}:${{ secrets.FHE_ACTIONS_TOKEN }}@github.com/${{ secrets.SYNC_DEST_REPO }}"
+          destination_repo: "https://${{ secrets.BOT_USERNAME }}:${{ secrets.CONCRETE_ACTIONS_TOKEN }}@github.com/${{ secrets.SYNC_DEST_REPO }}"
           destination_branch: "main"
       - name: git-sync tags
         uses: wei/git-sync@55c6b63b4f21607da0e9877ca9b4d11a29fc6d83
         with:
           source_repo: "zama-ai/tfhe-rs"
           source_branch: "refs/tags/*"
-          destination_repo: "https://${{ secrets.BOT_USERNAME }}:${{ secrets.FHE_ACTIONS_TOKEN }}@github.com/${{ secrets.SYNC_DEST_REPO }}"
+          destination_repo: "https://${{ secrets.BOT_USERNAME }}:${{ secrets.CONCRETE_ACTIONS_TOKEN }}@github.com/${{ secrets.SYNC_DEST_REPO }}"
           destination_branch: "refs/tags/*"

.github/workflows/wasm_client_benchmark.yml

@@ -1,14 +1,32 @@
-# Run WASM client benchmarks on an instance and return parsed results to Slab CI bot.
+# Run WASM client benchmarks on an AWS instance and return parsed results to Slab CI bot.
 name: WASM client benchmarks
 
 on:
   workflow_dispatch:
-  push:
-    branches:
-      - main
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 1a.m.
-    - cron: '0 1 * * 6'
+    inputs:
+      instance_id:
+        description: "Instance ID"
+        type: string
+      instance_image_id:
+        description: "Instance AMI ID"
+        type: string
+      instance_type:
+        description: "Instance product type"
+        type: string
+      runner_name:
+        description: "Action runner name"
+        type: string
+      request_id:
+        description: "Slab request ID"
+        type: string
+      # This input is not used in this workflow but still mandatory since a calling workflow could
+      # use it. If a triggering command include a user_inputs field, then the triggered workflow
+      # must include this very input, otherwise the workflow won't be called.
+      # See start_full_benchmarks.yml as example.
+      user_inputs:
+        description: "Type of benchmarks to run"
+        type: string
+        default: "weekly_benchmarks"
 
 env:
   CARGO_TERM_COLOR: always
@@ -16,81 +34,28 @@ env:
   ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
   RUST_BACKTRACE: "full"
   RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
 
 jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    if: github.event_name == 'workflow_dispatch' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
-      (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs')
-    permissions:
-      pull-requests: write
-    outputs:
-      wasm_bench: ${{ steps.changed-files.outputs.wasm_bench_any_changed }}
+  run-wasm-client-benchmarks:
+    name: Execute WASM client benchmarks in EC2
+    runs-on: ${{ github.event.inputs.runner_name }}
+    if: ${{ !cancelled() }}
     steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@c65cd883420fd2eb864698a825fc4162dd94482c
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            wasm_bench:
-              - tfhe/Cargo.toml
-              - concrete-csprng/**
-              - tfhe-zk-pok/**
-              - tfhe/src/**
-              - '!tfhe/src/c_api/**'
-              - tfhe/web_wasm_parallel_tests/**
-              - .github/workflows/wasm_client_benchmark.yml
-
-  setup-instance:
-    name: Setup instance (wasm-client-benchmarks)
-    if: github.event_name == 'workflow_dispatch' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
-      (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs' && needs.should-run.outputs.wasm_bench)
-    needs: should-run
-    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@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: cpu-small
-
-  wasm-client-benchmarks:
-    name: Execute WASM client benchmarks
-    needs: setup-instance
-    if: needs.setup-instance.result != 'skipped'
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    steps:
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Get benchmark details
+      - name: Instance configuration used
         run: |
-          {
-            echo "BENCH_DATE=$(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 "IDs: ${{ inputs.instance_id }}"
+          echo "AMI: ${{ inputs.instance_image_id }}"
+          echo "Type: ${{ inputs.instance_type }}"
+          echo "Request ID: ${{ inputs.request_id }}"
+
+      - name: Get benchmark date
+        run: |
+          echo "BENCH_DATE=$(date --iso-8601=seconds)" >> "${GITHUB_ENV}"
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
+        with:
+          fetch-depth: 0
 
       - name: Set up home
         # "Install rust" step require root user to have a HOME directory which is not set.
@@ -98,24 +63,27 @@ jobs:
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
+        uses: dtolnay/rust-toolchain@be73d7920c329f220ce78e0234b8f96b7ae60248
         with:
           toolchain: nightly
 
       - name: Run benchmarks
         run: |
           make install_node
-          make bench_web_js_api_parallel_ci
+          make ci_bench_web_js_api_parallel
 
       - name: Parse results
         run: |
           make parse_wasm_benchmarks
+
+          COMMIT_DATE="$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})"
+          COMMIT_HASH="$(git describe --tags --dirty)"
           python3 ./ci/benchmark_parser.py tfhe/wasm_pk_gen.csv ${{ env.RESULTS_FILENAME }} \
           --database tfhe_rs \
-          --hardware "m6i.4xlarge" \
-          --project-version "${{ env.COMMIT_HASH }}" \
+          --hardware ${{ inputs.instance_type }} \
+          --project-version "${COMMIT_HASH}" \
           --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
+          --commit-date "${COMMIT_DATE}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --key-gen
 
@@ -130,17 +98,17 @@ jobs:
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
+        uses: actions/upload-artifact@5d5d22a31266ced268874388b861e4b58bb5c2f3
         with:
           name: ${{ github.sha }}_wasm
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11
         with:
           repository: zama-ai/slab
           path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          token: ${{ secrets.CONCRETE_ACTIONS_TOKEN }}
 
       - name: Send data to Slab
         shell: bash
@@ -159,31 +127,11 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@b24d75fe0e728a4bf9fc42ee217caa686d141ee8
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "WASM benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (wasm-client-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, wasm-client-benchmarks ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (wasm-client-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "WASM benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/zk_pke_benchmark.yml

@@ -1,197 +0,0 @@
-# Run PKE Zero-Knowledge benchmarks on an instance and return parsed results to Slab CI bot.
-name: PKE ZK benchmarks
-
-on:
-  workflow_dispatch:
-  push:
-    branches:
-      - main
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 3a.m.
-    - cron: '0 3 * * 6'
-env:
-  CARGO_TERM_COLOR: always
-  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
-  PARSE_INTEGER_BENCH_CSV_FILE: tfhe_rs_integer_benches_${{ github.sha }}.csv
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-
-jobs:
-  should-run:
-    runs-on: ubuntu-latest
-    if: github.event_name == 'workflow_dispatch' ||
-      ((github.event_name == 'push' || github.event_name == 'schedule') && github.repository == 'zama-ai/tfhe-rs')
-    outputs:
-      zk_pok_changed: ${{ steps.changed-files.outputs.zk_pok_any_changed }}
-    steps:
-      - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Check for file changes
-        id: changed-files
-        uses: tj-actions/changed-files@c65cd883420fd2eb864698a825fc4162dd94482c
-        with:
-          since_last_remote_commit: true
-          files_yaml: |
-            zk_pok:
-              - tfhe/Cargo.toml
-              - concrete-csprng/**
-              - tfhe-zk-pok/**
-              - tfhe/src/core_crypto/**
-              - tfhe/src/shortint/**
-              - tfhe/src/integer/**
-              - tfhe/src/zk.rs
-              - tfhe/benches/integer/zk_pke.rs
-              - .github/workflows/zk_pke_benchmark.yml
-
-  setup-instance:
-    name: Setup instance (pke-zk-benchmarks)
-    runs-on: ubuntu-latest
-    needs: should-run
-    if: github.event_name == 'workflow_dispatch' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
-      (github.event_name == 'push' &&
-      github.repository == 'zama-ai/tfhe-rs' &&
-      needs.should-run.outputs.zk_pok_changed == 'true')
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: start
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
-          profile: bench
-
-  pke-zk-benchmarks:
-    name: Execute PKE ZK benchmarks
-    if: needs.setup-instance.result != 'skipped'
-    needs: setup-instance
-    concurrency:
-      group: ${{ github.workflow }}_${{github.event_name}}_${{ github.ref }}
-      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
-    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
-    steps:
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Get benchmark details
-        run: |
-          {
-            echo "BENCH_DATE=$(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)";
-          } >> "${GITHUB_ENV}"
-
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
-      - name: Install rust
-        uses: dtolnay/rust-toolchain@7b1c307e0dcbda6122208f10795a713336a9b35a
-        with:
-          toolchain: nightly
-
-      - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          repository: zama-ai/slab
-          path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Run benchmarks with AVX512
-        run: |
-          make bench_integer_zk
-
-      - name: Parse results
-        run: |
-          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
-          --database tfhe_rs \
-          --hardware "hpc7a.96xlarge" \
-          --backend cpu \
-          --project-version "${{ env.COMMIT_HASH }}" \
-          --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
-          --bench-date "${{ env.BENCH_DATE }}" \
-          --walk-subdirs \
-          --name-suffix avx512 \
-          --throughput
-
-      - name: Parse CRS sizes results
-        run: |
-          python3 ./ci/benchmark_parser.py tfhe/pke_zk_crs_sizes.csv ${{ env.RESULTS_FILENAME }} \
-          --key-sizes \
-          --append-results
-
-      - name: Upload parsed results artifact
-        uses: actions/upload-artifact@834a144ee995460fba8ed112a2fc961b36a5ec5a
-        with:
-          name: ${{ github.sha }}_integer_zk
-          path: ${{ env.RESULTS_FILENAME }}
-
-      - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          repository: zama-ai/slab
-          path: slab
-          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
-
-      - name: Send data to Slab
-        shell: bash
-        run: |
-          echo "Computing HMac on results file"
-          SIGNATURE="$(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: ${{ !success() && !cancelled() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "PKE ZK benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
-
-  teardown-instance:
-    name: Teardown instance (pke-zk-benchmarks)
-    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, pke-zk-benchmarks ]
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stop instance
-        id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
-        with:
-          mode: stop
-          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
-          slab-url: ${{ secrets.SLAB_BASE_URL }}
-          job-secret: ${{ secrets.JOB_SECRET }}
-          label: ${{ needs.setup-instance.outputs.runner-name }}
-
-      - name: Slack Notification
-        if: ${{ failure() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (pke-zk-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.gitignore

@@ -7,7 +7,6 @@ target/
 # In case of symlinked keys
 /keys
 
-**/*.rmeta
 **/Cargo.lock
 **/*.bin
 
@@ -23,9 +22,3 @@ dieharder_run.log
 
 # Cuda local build
 backends/tfhe-cuda-backend/cuda/cmake-build-debug/
-
-# WASM tests
-tfhe/web_wasm_parallel_tests/server.PID
-
-# Dir used for backward compatibility test data
-tfhe/tfhe-backward-compat-data/

Cargo.toml

@@ -1,21 +1,6 @@
 [workspace]
 resolver = "2"
-members = [
-    "tfhe",
-    "tfhe-zk-pok",
-    "tasks",
-    "apps/trivium",
-    "concrete-csprng",
-    "backends/tfhe-cuda-backend",
-    "utils/tfhe-versionable",
-    "utils/tfhe-versionable-derive",
-]
-
-exclude = [
-    "tfhe/backward_compatibility_tests",
-    "utils/cargo-tfhe-lints-inner",
-    "utils/cargo-tfhe-lints"
-]
+members = ["tfhe", "tasks", "apps/trivium", "concrete-csprng", "backends/tfhe-cuda-backend"]
 
 [profile.bench]
 lto = "fat"
@@ -32,4 +17,3 @@ lto = "off"
 inherits = "dev"
 opt-level = 3
 lto = "off"
-debug-assertions = false

Makefile

@@ -3,7 +3,6 @@ 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)
 CARGO_PROFILE?=release
@@ -16,15 +15,19 @@ GEN_KEY_CACHE_COVERAGE_ONLY?=FALSE
 PARSE_INTEGER_BENCH_CSV_FILE?=tfhe_rs_integer_benches.csv
 FAST_TESTS?=FALSE
 FAST_BENCH?=FALSE
-NIGHTLY_TESTS?=FALSE
 BENCH_OP_FLAVOR?=DEFAULT
-NODE_VERSION=22.4
+NODE_VERSION=20
 FORWARD_COMPAT?=OFF
-BACKWARD_COMPAT_DATA_URL=https://github.com/zama-ai/tfhe-backward-compat-data.git
-BACKWARD_COMPAT_DATA_BRANCH?=v0.1
-BACKWARD_COMPAT_DATA_PROJECT=tfhe-backward-compat-data
-BACKWARD_COMPAT_DATA_DIR=$(BACKWARD_COMPAT_DATA_PROJECT)
-TFHE_SPEC:=tfhe
+# sed: -n, do not print input stream, -e means a script/expression
+# 1,/version/ indicates from the first line, to the line matching version at the start of the line
+# p indicates to print, so we keep only the start of the Cargo.toml until we hit the first version
+# entry which should be the version of tfhe
+TFHE_CURRENT_VERSION:=\
+$(shell sed -n -e '1,/^version/p' tfhe/Cargo.toml | \
+grep '^version[[:space:]]*=' | cut -d '=' -f 2 | xargs)
+# Cargo has a hard time distinguishing between our package from the workspace and a package that
+# could be a dependency, so we build an unambiguous spec here
+TFHE_SPEC:=tfhe@$(TFHE_CURRENT_VERSION)
 # This is done to avoid forgetting it, we still precise the RUSTFLAGS in the commands to be able to
 # copy paste the command in the terminal and change them if required without forgetting the flags
 export RUSTFLAGS?=-C target-cpu=native
@@ -111,17 +114,12 @@ install_cargo_nextest: 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 || \
+	cargo $(CARGO_RS_BUILD_TOOLCHAIN) install wasm-pack || \
 	( echo "Unable to install cargo wasm-pack, unknown error." && exit 1 )
 
 .PHONY: install_node # Install last version of NodeJS via nvm
 install_node:
-	curl -o nvm_install.sh https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.3/install.sh
-	@echo "2ed5e94ba12434370f0358800deb69f514e8bce90f13beb0e1b241d42c6abafd nvm_install.sh" > nvm_checksum
-	@sha256sum -c nvm_checksum
-	@rm nvm_checksum
-	$(SHELL) nvm_install.sh
-	@rm nvm_install.sh
+	curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.3/install.sh | $(SHELL)
 	source ~/.bashrc
 	$(SHELL) -i -c 'nvm install $(NODE_VERSION)' || \
 	( echo "Unable to install node, unknown error." && exit 1 )
@@ -141,11 +139,6 @@ 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: check_linelint_installed # Check if linelint newline linter is installed
 check_linelint_installed:
 	@printf "\n" | linelint - > /dev/null 2>&1 || \
@@ -156,57 +149,24 @@ check_actionlint_installed:
 	@actionlint --version > /dev/null 2>&1 || \
 	( echo "Unable to locate actionlint. Try installing it: https://github.com/rhysd/actionlint/releases" && exit 1 )
 
-.PHONY: check_nvm_installed # Check if Node Version Manager is installed
-check_nvm_installed:
-	@source ~/.nvm/nvm.sh && nvm --version > /dev/null 2>&1 || \
-	( echo "Unable to locate Node. Run 'make install_node'" && exit 1 )
-
-.PHONY: install_mlc # Install mlc (Markup Link Checker)
-install_mlc: install_rs_build_toolchain
-	@mlc --version > /dev/null 2>&1 || \
-	cargo $(CARGO_RS_BUILD_TOOLCHAIN) install mlc --locked || \
-	( echo "Unable to install mlc, unknown error." && exit 1 )
-
 .PHONY: fmt # Format rust code
 fmt: install_rs_check_toolchain
 	cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" fmt
 
-.PHONY: fmt_js # Format javascript code
-fmt_js: check_nvm_installed
-	source ~/.nvm/nvm.sh && \
-	nvm install $(NODE_VERSION) && \
-	nvm use $(NODE_VERSION) && \
-	$(MAKE) -C tfhe/web_wasm_parallel_tests fmt
-
 .PHONY: fmt_gpu # Format rust and cuda code
 fmt_gpu: install_rs_check_toolchain
 	cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" fmt
 	cd "$(TFHECUDA_SRC)" && ./format_tfhe_cuda_backend.sh
 
-.PHONY: fmt_c_tests # Format c tests
-fmt_c_tests:
-	find tfhe/c_api_tests/ -regex '.*\.\(cpp\|hpp\|cu\|c\|h\)' -exec clang-format -style=file -i {} \;
-
 .PHONY: check_fmt # Check rust code format
 check_fmt: install_rs_check_toolchain
 	cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" fmt --check
 
-.PHONY: check_fmt_c_tests  # Check C tests format
-check_fmt_c_tests:
-	find tfhe/c_api_tests/ -regex '.*\.\(cpp\|hpp\|cu\|c\|h\)' -exec clang-format --dry-run --Werror -style=file {} \;
-
 .PHONY: check_fmt_gpu # Check rust and cuda code format
 check_fmt_gpu: install_rs_check_toolchain
 	cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" fmt --check
 	cd "$(TFHECUDA_SRC)" && ./format_tfhe_cuda_backend.sh -c
 
-.PHONY: check_fmt_js # Check javascript code format
-check_fmt_js: check_nvm_installed
-	source ~/.nvm/nvm.sh && \
-	nvm install $(NODE_VERSION) && \
-	nvm use $(NODE_VERSION) && \
-	$(MAKE) -C tfhe/web_wasm_parallel_tests check_fmt
-
 .PHONY: clippy_gpu # Run clippy lints on tfhe with "gpu" enabled
 clippy_gpu: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
@@ -265,74 +225,53 @@ clippy: install_rs_check_toolchain
 		--features=$(TARGET_ARCH_FEATURE),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
-clippy_rustdoc: install_rs_check_toolchain
-	if [[ "$(OS)" != "Linux" && "$(OS)" != "Darwin" ]]; then \
-		echo "WARNING: skipped clippy_rustdoc, unsupported OS $(OS)"; \
-		exit 0; \
-	fi && \
-	CLIPPYFLAGS="-D warnings" RUSTDOCFLAGS="--no-run --nocapture --test-builder ./scripts/clippy_driver.sh -Z unstable-options" \
-		cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" test --doc \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,zk-pok,pbs-stats \
-		-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=$(TARGET_ARCH_FEATURE),boolean-c-api,shortint-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
 clippy_js_wasm_api: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
-		--features=boolean-client-js-wasm-api,shortint-client-js-wasm-api,integer-client-js-wasm-api,high-level-client-js-wasm-api \
+		--features=boolean-client-js-wasm-api,shortint-client-js-wasm-api,integer-client-js-wasm-api \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 
 .PHONY: clippy_tasks # Run clippy lints on helper tasks crate.
-clippy_tasks: install_rs_check_toolchain
+clippy_tasks:
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
 		-p tasks -- --no-deps -D warnings
 
 .PHONY: clippy_trivium # Run clippy lints on Trivium app
 clippy_trivium: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy --all-targets \
+	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy \
 		-p tfhe-trivium -- --no-deps -D warnings
 
 .PHONY: clippy_all_targets # Run clippy lints on all targets (benches, examples, etc.)
-clippy_all_targets: install_rs_check_toolchain
+clippy_all_targets:
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy --all-targets \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache,zk-pok \
+		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache \
 		-p $(TFHE_SPEC) -- --no-deps -D warnings
 
 .PHONY: clippy_concrete_csprng # Run clippy lints on concrete-csprng
-clippy_concrete_csprng: install_rs_check_toolchain
+clippy_concrete_csprng:
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy --all-targets \
 		--features=$(TARGET_ARCH_FEATURE) \
 		-p concrete-csprng -- --no-deps -D warnings
 
-.PHONY: clippy_zk_pok # Run clippy lints on tfhe-zk-pok
-clippy_zk_pok: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy --all-targets \
-		-p tfhe-zk-pok -- --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_all: 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_trivium
 
 .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_fast: clippy clippy_all_targets clippy_c_api clippy_js_wasm_api clippy_tasks clippy_core \
+clippy_concrete_csprng
 
 .PHONY: clippy_cuda_backend # Run clippy lints on the tfhe-cuda-backend
 clippy_cuda_backend: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" clippy --all-targets \
 		-p tfhe-cuda-backend -- --no-deps -D warnings
 
-.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 -- -D warnings
-
 .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 $(CARGO_PROFILE) \
@@ -385,21 +324,21 @@ symlink_c_libs_without_fingerprint:
 .PHONY: build_c_api # Build the C API for boolean, shortint and integer
 build_c_api: install_rs_check_toolchain
 	RUSTFLAGS="$(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,$(FORWARD_COMPAT_FEATURE) \
+		--features=$(TARGET_ARCH_FEATURE),boolean-c-api,shortint-c-api,high-level-c-api,$(FORWARD_COMPAT_FEATURE) \
 		-p $(TFHE_SPEC)
 	@"$(MAKE)" symlink_c_libs_without_fingerprint
 
 .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=$(TARGET_ARCH_FEATURE),boolean-c-api,shortint-c-api,high-level-c-api,gpu \
 		-p $(TFHE_SPEC)
 	@"$(MAKE)" symlink_c_libs_without_fingerprint
 
 .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,$(FORWARD_COMPAT_FEATURE) \
+		--features=$(TARGET_ARCH_FEATURE),boolean-c-api,shortint-c-api,high-level-c-api,experimental-force_fft_algo_dif4,$(FORWARD_COMPAT_FEATURE) \
 		-p $(TFHE_SPEC)
 	@"$(MAKE)" symlink_c_libs_without_fingerprint
 
@@ -408,7 +347,7 @@ build_web_js_api: install_rs_build_toolchain install_wasm_pack
 	cd tfhe && \
 	RUSTFLAGS="$(WASM_RUSTFLAGS)" rustup run "$(RS_BUILD_TOOLCHAIN)" \
 		wasm-pack build --release --target=web \
-		-- --features=boolean-client-js-wasm-api,shortint-client-js-wasm-api,integer-client-js-wasm-api,zk-pok
+		-- --features=boolean-client-js-wasm-api,shortint-client-js-wasm-api,integer-client-js-wasm-api
 
 .PHONY: build_web_js_api_parallel # Build the js API targeting the web browser with parallelism support
 build_web_js_api_parallel: install_rs_check_toolchain install_wasm_pack
@@ -416,16 +355,15 @@ build_web_js_api_parallel: install_rs_check_toolchain install_wasm_pack
 	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) \
 		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';|" {} \;
+		-- --features=boolean-client-js-wasm-api,shortint-client-js-wasm-api,integer-client-js-wasm-api,parallel-wasm-api \
+		-Z build-std=panic_abort,std
 
 .PHONY: build_node_js_api # Build the js API targeting nodejs
 build_node_js_api: install_rs_build_toolchain install_wasm_pack
 	cd tfhe && \
 	RUSTFLAGS="$(WASM_RUSTFLAGS)" rustup run "$(RS_BUILD_TOOLCHAIN)" \
 		wasm-pack build --release --target=nodejs \
-		-- --features=boolean-client-js-wasm-api,shortint-client-js-wasm-api,integer-client-js-wasm-api,zk-pok
+		-- --features=boolean-client-js-wasm-api,shortint-client-js-wasm-api,integer-client-js-wasm-api
 
 .PHONY: build_concrete_csprng # Build concrete_csprng
 build_concrete_csprng: install_rs_build_toolchain
@@ -435,10 +373,10 @@ build_concrete_csprng: install_rs_build_toolchain
 .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 $(TFHE_SPEC) -- core_crypto::
+		--features=$(TARGET_ARCH_FEATURE),experimental -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,$(AVX512_FEATURE) -p $(TFHE_SPEC) -- core_crypto::; \
+			--features=$(TARGET_ARCH_FEATURE),experimental,$(AVX512_FEATURE) -p $(TFHE_SPEC) -- core_crypto::; \
 	fi
 
 .PHONY: test_core_crypto_cov # Run the tests of the core_crypto module with code coverage
@@ -461,80 +399,26 @@ test_cuda_backend:
 	mkdir -p "$(TFHECUDA_BUILD)" && \
 		cd "$(TFHECUDA_BUILD)" && \
 		cmake .. -DCMAKE_BUILD_TYPE=Release -DTFHE_CUDA_BACKEND_BUILD_TESTS=ON && \
-		"$(MAKE)" -j "$(CPU_COUNT)" && \
-		"$(MAKE)" test
+		make -j && \
+		make test
 
 .PHONY: test_gpu # Run the tests of the core_crypto module including experimental on the gpu backend
 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
+test_core_crypto_gpu: install_rs_build_toolchain install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
 		--features=$(TARGET_ARCH_FEATURE),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::
 
 .PHONY: test_integer_gpu # Run the tests of the integer module including experimental on the gpu backend
-test_integer_gpu: install_rs_build_toolchain
+test_integer_gpu: install_rs_build_toolchain install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		--features=$(TARGET_ARCH_FEATURE),integer,gpu -p $(TFHE_SPEC) -- integer::gpu::server_key:: --test-threads=6
+		--features=$(TARGET_ARCH_FEATURE),integer,gpu -p $(TFHE_SPEC) -- integer::gpu::server_key::
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --doc --profile $(CARGO_PROFILE) \
 		--features=$(TARGET_ARCH_FEATURE),integer,gpu -p $(TFHE_SPEC) -- integer::gpu::server_key::
 
-.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
-	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
-	FAST_TESTS="$(FAST_TESTS)" \
-	NIGHTLY_TESTS="$(NIGHTLY_TESTS)" \
-		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_CHECK_TOOLCHAIN) \
-		--cargo-profile "$(CARGO_PROFILE)" --backend "gpu" \
-		--tfhe-package "$(TFHE_SPEC)"
-
-.PHONY: test_unsigned_integer_gpu_ci # Run the tests for unsigned integer ci on gpu backend
-test_unsigned_integer_gpu_ci: install_rs_check_toolchain install_cargo_nextest
-	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
-	FAST_TESTS="$(FAST_TESTS)" \
-	NIGHTLY_TESTS="$(NIGHTLY_TESTS)" \
-		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_CHECK_TOOLCHAIN) \
-		--cargo-profile "$(CARGO_PROFILE)" --backend "gpu" \
-		--unsigned-only --tfhe-package "$(TFHE_SPEC)"
-
-.PHONY: test_signed_integer_gpu_ci # Run the tests for signed integer ci on gpu backend
-test_signed_integer_gpu_ci: install_rs_check_toolchain install_cargo_nextest
-	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
-	FAST_TESTS="$(FAST_TESTS)" \
-	NIGHTLY_TESTS="$(NIGHTLY_TESTS)" \
-		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_CHECK_TOOLCHAIN) \
-		--cargo-profile "$(CARGO_PROFILE)" --backend "gpu" \
-		--signed-only --tfhe-package "$(TFHE_SPEC)"
-
-.PHONY: test_integer_multi_bit_gpu_ci # Run the tests for integer ci on gpu backend running only multibit tests
-test_integer_multi_bit_gpu_ci: install_rs_check_toolchain install_cargo_nextest
-	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
-	FAST_TESTS="$(FAST_TESTS)" \
-	NIGHTLY_TESTS="$(NIGHTLY_TESTS)" \
-		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_CHECK_TOOLCHAIN) \
-		--cargo-profile "$(CARGO_PROFILE)" --multi-bit --backend "gpu" \
-		--tfhe-package "$(TFHE_SPEC)"
-
-.PHONY: test_unsigned_integer_multi_bit_gpu_ci # Run the tests for unsigned integer ci on gpu backend running only multibit tests
-test_unsigned_integer_multi_bit_gpu_ci: install_rs_check_toolchain install_cargo_nextest
-	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
-	FAST_TESTS="$(FAST_TESTS)" \
-	NIGHTLY_TESTS="$(NIGHTLY_TESTS)" \
-		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_CHECK_TOOLCHAIN) \
-		--cargo-profile "$(CARGO_PROFILE)" --multi-bit --backend "gpu" \
-		--unsigned-only --tfhe-package "$(TFHE_SPEC)"
-
-.PHONY: test_signed_integer_multi_bit_gpu_ci # Run the tests for signed integer ci on gpu backend running only multibit tests
-test_signed_integer_multi_bit_gpu_ci: install_rs_check_toolchain install_cargo_nextest
-	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
-	FAST_TESTS="$(FAST_TESTS)" \
-	NIGHTLY_TESTS="$(NIGHTLY_TESTS)" \
-		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_CHECK_TOOLCHAIN) \
-		--cargo-profile "$(CARGO_PROFILE)" --multi-bit --backend "gpu" \
-		--signed-only --tfhe-package "$(TFHE_SPEC)"
-
 .PHONY: test_boolean # Run the tests of the boolean module
 test_boolean: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
@@ -557,14 +441,14 @@ test_c_api_rs: install_rs_check_toolchain
 
 .PHONY: test_c_api_c # Run the C tests for the C API
 test_c_api_c: build_c_api
-	./scripts/c_api_tests.sh --cargo-profile "$(CARGO_PROFILE)"
+	./scripts/c_api_tests.sh
 
 .PHONY: test_c_api # Run all the tests for the C API
 test_c_api: test_c_api_rs test_c_api_c
 
 .PHONY: test_c_api_gpu # Run the C tests for the C API
 test_c_api_gpu: build_c_api_gpu
-	./scripts/c_api_tests.sh --gpu --cargo-profile "$(CARGO_PROFILE)"
+	./scripts/c_api_tests.sh --gpu
 
 .PHONY: test_shortint_ci # Run the tests for shortint ci
 test_shortint_ci: install_rs_build_toolchain install_cargo_nextest
@@ -597,7 +481,6 @@ test_shortint_cov: install_rs_check_toolchain install_tarpaulin
 test_integer_ci: install_rs_check_toolchain install_cargo_nextest
 	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
 	FAST_TESTS="$(FAST_TESTS)" \
-	NIGHTLY_TESTS="$(NIGHTLY_TESTS)" \
 		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_CHECK_TOOLCHAIN) \
 		--cargo-profile "$(CARGO_PROFILE)" --avx512-support "$(AVX512_SUPPORT)" \
 		--tfhe-package "$(TFHE_SPEC)"
@@ -606,7 +489,6 @@ test_integer_ci: install_rs_check_toolchain install_cargo_nextest
 test_unsigned_integer_ci: install_rs_check_toolchain install_cargo_nextest
 	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
 	FAST_TESTS="$(FAST_TESTS)" \
-	NIGHTLY_TESTS="$(NIGHTLY_TESTS)" \
 		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_CHECK_TOOLCHAIN) \
 		--cargo-profile "$(CARGO_PROFILE)" --avx512-support "$(AVX512_SUPPORT)" \
 		--unsigned-only --tfhe-package "$(TFHE_SPEC)"
@@ -615,7 +497,6 @@ test_unsigned_integer_ci: install_rs_check_toolchain install_cargo_nextest
 test_signed_integer_ci: install_rs_check_toolchain install_cargo_nextest
 	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
 	FAST_TESTS="$(FAST_TESTS)" \
-	NIGHTLY_TESTS="$(NIGHTLY_TESTS)" \
 		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_CHECK_TOOLCHAIN) \
 		--cargo-profile "$(CARGO_PROFILE)" --avx512-support "$(AVX512_SUPPORT)" \
 		--signed-only --tfhe-package "$(TFHE_SPEC)"
@@ -624,25 +505,22 @@ test_signed_integer_ci: install_rs_check_toolchain install_cargo_nextest
 test_integer_multi_bit_ci: install_rs_check_toolchain install_cargo_nextest
 	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
 	FAST_TESTS="$(FAST_TESTS)" \
-	NIGHTLY_TESTS="$(NIGHTLY_TESTS)" \
 		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_CHECK_TOOLCHAIN) \
 		--cargo-profile "$(CARGO_PROFILE)" --multi-bit --avx512-support "$(AVX512_SUPPORT)" \
 		--tfhe-package "$(TFHE_SPEC)"
 
-.PHONY: test_unsigned_integer_multi_bit_ci # Run the tests for unsigned integer ci running only multibit tests
+.PHONY: test_unsigned_integer_multi_bit_ci # Run the tests for nsigned integer ci running only multibit tests
 test_unsigned_integer_multi_bit_ci: install_rs_check_toolchain install_cargo_nextest
 	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
 	FAST_TESTS="$(FAST_TESTS)" \
-	NIGHTLY_TESTS="$(NIGHTLY_TESTS)" \
 		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_CHECK_TOOLCHAIN) \
 		--cargo-profile "$(CARGO_PROFILE)" --multi-bit --avx512-support "$(AVX512_SUPPORT)" \
 		--unsigned-only --tfhe-package "$(TFHE_SPEC)"
 
-.PHONY: test_signed_integer_multi_bit_ci # Run the tests for signed integer ci running only multibit tests
+.PHONY: test_signed_integer_multi_bit_ci # Run the tests for nsigned integer ci running only multibit tests
 test_signed_integer_multi_bit_ci: install_rs_check_toolchain install_cargo_nextest
 	BIG_TESTS_INSTANCE="$(BIG_TESTS_INSTANCE)" \
 	FAST_TESTS="$(FAST_TESTS)" \
-	NIGHTLY_TESTS="$(NIGHTLY_TESTS)" \
 		./scripts/integer-tests.sh --rust-toolchain $(CARGO_RS_CHECK_TOOLCHAIN) \
 		--cargo-profile "$(CARGO_PROFILE)" --multi-bit --avx512-support "$(AVX512_SUPPORT)" \
 		--signed-only --tfhe-package "$(TFHE_SPEC)"
@@ -669,7 +547,7 @@ test_integer_cov: install_rs_check_toolchain install_tarpaulin
 .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 $(TFHE_SPEC) \
+		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache -p $(TFHE_SPEC) \
 		-- high_level_api::
 
 test_high_level_api_gpu: install_rs_build_toolchain install_cargo_nextest
@@ -680,14 +558,13 @@ test_high_level_api_gpu: install_rs_build_toolchain install_cargo_nextest
 .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 \
-		-p $(TFHE_SPEC) \
+		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache -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 $(TFHE_SPEC) \
+		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,internal-keycache,gpu -p $(TFHE_SPEC) \
 		-- test_user_docs::
 
 .PHONY: test_fhe_strings # Run tests for fhe_strings example
@@ -722,87 +599,46 @@ test_kreyvium: install_rs_build_toolchain
 		-p tfhe-trivium -- --test-threads=1 kreyvium::
 
 .PHONY: test_concrete_csprng # Run concrete-csprng tests
-test_concrete_csprng: install_rs_build_toolchain
+test_concrete_csprng:
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
 		--features=$(TARGET_ARCH_FEATURE) -p concrete-csprng
 
-.PHONY: test_zk_pok # Run tfhe-zk-pok tests
-test_zk_pok: install_rs_build_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		-p tfhe-zk-pok
-
-.PHONY: test_versionable # Run tests for tfhe-versionable subcrate
-test_versionable: install_rs_build_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --profile $(CARGO_PROFILE) \
-		-p tfhe-versionable
-
-# The backward compat data repo holds historical binary data but also rust code to generate and load them.
-# Here we use the "patch" functionality of Cargo to make sure the repo used for the data is the same as the one used for the code.
-.PHONY: test_backward_compatibility_ci
-test_backward_compatibility_ci: install_rs_build_toolchain
-	TFHE_BACKWARD_COMPAT_DATA_DIR="$(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 -p $(TFHE_SPEC) test_backward_compatibility -- --nocapture
-
-.PHONY: test_backward_compatibility # Same as test_backward_compatibility_ci but tries to clone the data repo first if needed
-test_backward_compatibility: tfhe/$(BACKWARD_COMPAT_DATA_DIR) test_backward_compatibility_ci
-
-.PHONY: backward_compat_branch # Prints the required backward compatibility branch
-backward_compat_branch:
-	@echo "$(BACKWARD_COMPAT_DATA_BRANCH)"
-
 .PHONY: doc # Build rust doc
 doc: install_rs_check_toolchain
-	@# Even though we are not in docs.rs, this allows to "just" build the doc
-	DOCS_RS=1 \
 	RUSTDOCFLAGS="--html-in-header katex-header.html" \
 	cargo "$(CARGO_RS_CHECK_TOOLCHAIN)" doc \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,gpu,internal-keycache,experimental,zk-pok --no-deps -p $(TFHE_SPEC)
+		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer --no-deps -p $(TFHE_SPEC)
 
 .PHONY: docs # Build rust doc alias for doc
 docs: doc
 
 .PHONY: lint_doc # Build rust doc with linting enabled
 lint_doc: 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 "$(CARGO_RS_CHECK_TOOLCHAIN)" doc \
-		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer,gpu,internal-keycache,experimental,zk-pok -p $(TFHE_SPEC) --no-deps
+		--features=$(TARGET_ARCH_FEATURE),boolean,shortint,integer -p $(TFHE_SPEC) --no-deps
 
 .PHONY: lint_docs # Build rust doc with linting enabled alias for lint_doc
 lint_docs: lint_doc
 
 .PHONY: format_doc_latex # Format the documentation latex equations to avoid broken rendering.
 format_doc_latex:
-	RUSTFLAGS="" cargo xtask format_latex_doc
+	cargo xtask format_latex_doc
 	@"$(MAKE)" --no-print-directory fmt
 	@printf "\n===============================\n\n"
 	@printf "Please manually inspect changes made by format_latex_doc, rustfmt can break equations \
 	if the line length is exceeded\n"
 	@printf "\n===============================\n"
 
-.PHONY: check_md_docs_are_tested # Checks that the rust codeblocks in our .md files are tested
-check_md_docs_are_tested:
-	RUSTFLAGS="" cargo xtask check_tfhe_docs_are_tested
-
-.PHONY: check_intra_md_links # Checks broken internal links in Markdown docs
-check_intra_md_links: install_mlc
-	mlc --offline --match-file-extension tfhe/docs
-
-.PHONY: check_md_links # Checks all broken links in Markdown docs
-check_md_links: install_mlc
-	mlc --match-file-extension tfhe/docs
-
 .PHONY: check_compile_tests # Build tests in debug without running them
-check_compile_tests: install_rs_build_toolchain
+check_compile_tests:
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_BUILD_TOOLCHAIN) test --no-run \
 		--features=$(TARGET_ARCH_FEATURE),experimental,boolean,shortint,integer,internal-keycache \
 		-p $(TFHE_SPEC)
 
 	@if [[ "$(OS)" == "Linux" || "$(OS)" == "Darwin" ]]; then \
 		"$(MAKE)" build_c_api && \
-		./scripts/c_api_tests.sh --build-only --cargo-profile "$(CARGO_PROFILE)"; \
+		./scripts/c_api_tests.sh --build-only; \
 	fi
 
 .PHONY: check_compile_tests_benches_gpu # Build tests in debug without running them
@@ -813,7 +649,7 @@ check_compile_tests_benches_gpu: install_rs_build_toolchain
 	mkdir -p "$(TFHECUDA_BUILD)" && \
 		cd "$(TFHECUDA_BUILD)" && \
 		cmake .. -DCMAKE_BUILD_TYPE=Debug -DTFHE_CUDA_BACKEND_BUILD_TESTS=ON -DTFHE_CUDA_BACKEND_BUILD_BENCHMARKS=ON && \
-		"$(MAKE)" -j "$(CPU_COUNT)"
+		make -j
 
 .PHONY: build_nodejs_test_docker # Build a docker image with tools to run nodejs tests for wasm API
 build_nodejs_test_docker:
@@ -833,14 +669,14 @@ test_nodejs_wasm_api_in_docker: build_nodejs_test_docker
 
 .PHONY: test_nodejs_wasm_api # Run tests for the nodejs on wasm API
 test_nodejs_wasm_api: build_node_js_api
-	cd tfhe/js_on_wasm_tests && npm run test
+	cd tfhe && node --test js_on_wasm_tests
 
 .PHONY: test_web_js_api_parallel # Run tests for the web wasm api
 test_web_js_api_parallel: build_web_js_api_parallel
 	$(MAKE) -C tfhe/web_wasm_parallel_tests test
 
-.PHONY: test_web_js_api_parallel_ci # Run tests for the web wasm api
-test_web_js_api_parallel_ci: build_web_js_api_parallel
+.PHONY: ci_test_web_js_api_parallel # Run tests for the web wasm api
+ci_test_web_js_api_parallel: build_web_js_api_parallel
 	source ~/.nvm/nvm.sh && \
 	nvm install $(NODE_VERSION) && \
 	nvm use $(NODE_VERSION) && \
@@ -907,22 +743,6 @@ bench_integer_multi_bit_gpu: install_rs_check_toolchain
 	--bench integer-bench \
 	--features=$(TARGET_ARCH_FEATURE),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) \
-	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
-	--bench integer-bench \
-	--features=$(TARGET_ARCH_FEATURE),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)" \
-	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
-	--bench zk-pke-bench \
-	--features=$(TARGET_ARCH_FEATURE),integer,internal-keycache,zk-pok,nightly-avx512 \
-	-p $(TFHE_SPEC) --
-
 .PHONY: bench_shortint # Run benchmarks for shortint
 bench_shortint: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_BENCH_OP_FLAVOR=$(BENCH_OP_FLAVOR) \
@@ -930,12 +750,18 @@ bench_shortint: install_rs_check_toolchain
 	--bench shortint-bench \
 	--features=$(TARGET_ARCH_FEATURE),shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
 
-.PHONY: bench_shortint_oprf # Run benchmarks for shortint
-bench_shortint_oprf: install_rs_check_toolchain
+.PHONY: bench_oprf # Run benchmarks for shortint
+bench_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 $(TFHE_SPEC)
+	RUSTFLAGS="$(RUSTFLAGS)" \
+	cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
+	--bench oprf-integer-bench \
+	--features=$(TARGET_ARCH_FEATURE),integer,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
@@ -945,6 +771,7 @@ bench_shortint_multi_bit: install_rs_check_toolchain
 	--bench shortint-bench \
 	--features=$(TARGET_ARCH_FEATURE),shortint,internal-keycache,nightly-avx512 -p $(TFHE_SPEC) --
 
+
 .PHONY: bench_boolean # Run benchmarks for boolean
 bench_boolean: install_rs_check_toolchain
 	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
@@ -957,15 +784,9 @@ bench_pbs: install_rs_check_toolchain
 	--bench pbs-bench \
 	--features=$(TARGET_ARCH_FEATURE),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 $(TFHE_SPEC)
-
 .PHONY: bench_pbs_gpu # Run benchmarks for PBS on GPU backend
 bench_pbs_gpu: install_rs_check_toolchain
-	RUSTFLAGS="$(RUSTFLAGS)" __TFHE_RS_FAST_BENCH=$(FAST_BENCH) cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
+	RUSTFLAGS="$(RUSTFLAGS)" cargo $(CARGO_RS_CHECK_TOOLCHAIN) bench \
 	--bench pbs-bench \
 	--features=$(TARGET_ARCH_FEATURE),boolean,shortint,gpu,internal-keycache,nightly-avx512 -p $(TFHE_SPEC)
 
@@ -985,16 +806,15 @@ bench_ks_gpu: install_rs_check_toolchain
 bench_web_js_api_parallel: build_web_js_api_parallel
 	$(MAKE) -C tfhe/web_wasm_parallel_tests bench
 
-.PHONY: bench_web_js_api_parallel_ci # Run benchmarks for the web wasm api
-bench_web_js_api_parallel_ci: build_web_js_api_parallel
+.PHONY: ci_bench_web_js_api_parallel # Run benchmarks for the web wasm api
+ci_bench_web_js_api_parallel: build_web_js_api_parallel
 	source ~/.nvm/nvm.sh && \
-	nvm use $(NODE_VERSION) && \
+	nvm use node && \
 	$(MAKE) -C tfhe/web_wasm_parallel_tests bench-ci
 
 #
 # Utility tools
 #
-
 .PHONY: gen_key_cache # Run the script to generate keys and cache them for shortint tests
 gen_key_cache: install_rs_build_toolchain
 	RUSTFLAGS="$(RUSTFLAGS) --cfg tarpaulin" cargo $(CARGO_RS_BUILD_TOOLCHAIN) run --profile $(CARGO_PROFILE) \
@@ -1045,12 +865,6 @@ write_params_to_file: install_rs_check_toolchain
 	--example write_params_to_file \
 	--features=$(TARGET_ARCH_FEATURE),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)
-
-tfhe/$(BACKWARD_COMPAT_DATA_DIR): clone_backward_compat_data
-
 #
 # Real use case examples
 #
@@ -1076,15 +890,13 @@ sha256_bool: install_rs_check_toolchain
 	--features=$(TARGET_ARCH_FEATURE),boolean
 
 .PHONY: pcc # pcc stands for pre commit checks (except GPU)
-pcc: no_tfhe_typo no_dbg_log check_fmt lint_doc check_md_docs_are_tested check_intra_md_links \
-clippy_all tfhe_lints check_compile_tests
+pcc: no_tfhe_typo no_dbg_log check_fmt lint_doc clippy_all check_compile_tests
 
 .PHONY: pcc_gpu # pcc stands for pre commit checks for GPU compilation
 pcc_gpu: clippy_gpu clippy_cuda_backend check_compile_tests_benches_gpu
 
 .PHONY: fpcc # pcc stands for pre commit checks, the f stands for fast
-fpcc: no_tfhe_typo no_dbg_log check_fmt lint_doc check_md_docs_are_tested clippy_fast \
-check_compile_tests
+fpcc: no_tfhe_typo no_dbg_log check_fmt lint_doc clippy_fast check_compile_tests
 
 .PHONY: conformance # Automatically fix problems that can be fixed
 conformance: fix_newline fmt

README.md

@@ -1,10 +1,6 @@
 <p align="center">
 <!-- product name logo -->
-<picture>
-  <source media="(prefers-color-scheme: dark)" srcset="https://github.com/zama-ai/tfhe-rs/assets/157474013/5283e0ba-da1e-43af-9f2a-c5221367a12b">
-  <source media="(prefers-color-scheme: light)" srcset="https://github.com/zama-ai/tfhe-rs/assets/157474013/b94a8c96-7595-400b-9311-70765c706955">
-  <img width=600 alt="Zama TFHE-rs">
-</picture>
+  <img width=600 src="https://user-images.githubusercontent.com/5758427/231206749-8f146b97-3c5a-4201-8388-3ffa88580415.png">
 </p>
 
 <hr/>
@@ -50,7 +46,7 @@ production-ready library for all the advanced features of TFHE.
 <br></br>
 
 ## Table of Contents
-- **[Getting started](#getting-started)**
+- **[Getting Started](#getting-started)**
    - [Cargo.toml configuration](#cargotoml-configuration)
    - [A simple example](#a-simple-example)
 - **[Resources](#resources)**
@@ -65,7 +61,7 @@ production-ready library for all the advanced features of TFHE.
 - **[Support](#support)**
 <br></br>
 
-## Getting started
+## Getting Started
 
 ### 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`:
@@ -131,13 +127,13 @@ fn main() -> Result<(), Box<dyn std::error::Error>> {
     // Clear equivalent computations: 1344 * 5 = 6720
     let encrypted_res_mul = &encrypted_a * &encrypted_b;
 
-    // Clear equivalent computations: 6720 >> 5 = 210
+    // Clear equivalent computations: 1344 >> 5 = 42
     encrypted_a = &encrypted_res_mul >> &encrypted_b;
 
     // Clear equivalent computations: let casted_a = a as u8;
     let casted_a: FheUint8 = encrypted_a.cast_into();
 
-    // Clear equivalent computations: min(210, 7) = 7
+    // Clear equivalent computations: min(42, 7) = 7
     let encrypted_res_min = &casted_a.min(&encrypted_c);
 
     // Operation between clear and encrypted data:
@@ -177,12 +173,12 @@ to run in release mode with cargo's `--release` flag to have the best performanc
 <br></br>
 
 ### Tutorials
-- [[Video tutorial] Implement signed integers using TFHE-rs ](https://www.zama.ai/post/video-tutorial-implement-signed-integers-ssing-tfhe-rs)
-- [Homomorphic parity bit](https://docs.zama.ai/tfhe-rs/tutorials/parity_bit)
-- [Homomorphic case changing on Ascii string](https://docs.zama.ai/tfhe-rs/tutorials/ascii_fhe_string)
+- [Homomorphic Parity Bit](https://docs.zama.ai/tfhe-rs/tutorials/parity_bit)
+- [Homomorphic Case Changing on Ascii String](https://docs.zama.ai/tfhe-rs/tutorials/ascii_fhe_string)
 - [Boolean SHA256 with TFHE-rs](https://www.zama.ai/post/boolean-sha256-tfhe-rs)
-- [Dark market with TFHE-rs](https://www.zama.ai/post/dark-market-tfhe-rs)
-- [Regular expression engine with TFHE-rs](https://www.zama.ai/post/regex-engine-tfhe-rs)
+- [Dark Market with TFHE-rs](https://www.zama.ai/post/dark-market-tfhe-rs)
+- [Regular Expression Engine with TFHE-rs](https://www.zama.ai/post/regex-engine-tfhe-rs)
+
 
 *Explore more useful resources in [TFHE-rs tutorials](https://docs.zama.ai/tfhe-rs/tutorials) and [Awesome Zama repo](https://github.com/zama-ai/awesome-zama)*
 <br></br>
@@ -198,7 +194,7 @@ Full, comprehensive documentation is available here: [https://docs.zama.ai/tfhe-
 
 ### Disclaimers
 
-#### Security estimation
+#### Security Estimation
 
 Security estimations are done using the
 [Lattice Estimator](https://github.com/malb/lattice-estimator)
@@ -206,13 +202,7 @@ with `red_cost_model = reduction.RC.BDGL16`.
 
 When a new update is published in the Lattice Estimator, we update parameters accordingly.
 
-### Security model
-
-The default parameters for the TFHE-rs library are chosen considering the IND-CPA security model, and are selected with a bootstrapping failure probability fixed at p_error = $2^{-64}$. In particular, it is assumed that the results of decrypted computations are not shared by the secret key owner with any third parties, as such an action can lead to leakage of the secret encryption key. If you are designing an application where decryptions must be shared, you will need to craft custom encryption parameters which are chosen in consideration of the IND-CPA^D security model [1]. 
-
-[1] Li, Baiyu, et al. "Securing approximate homomorphic encryption using differential privacy." Annual International Cryptology Conference. Cham: Springer Nature Switzerland, 2022. https://eprint.iacr.org/2022/816.pdf
-
-#### Side-channel attacks
+#### Side-Channel Attacks
 
 Mitigation for side-channel attacks has not yet been implemented in TFHE-rs,
 and will be released in upcoming versions.
@@ -241,23 +231,7 @@ Becoming an approved contributor involves signing our Contributor License Agreem
 <br></br>
 
 ### License
-This software is distributed under the **BSD-3-Clause-Clear** license. Read [this](LICENSE) for more details.
-
-#### FAQ
-**Is Zama’s technology free to use?**
->Zama’s libraries are free to use under the BSD 3-Clause Clear license only for development, research, prototyping, and experimentation purposes. However, for any commercial use of Zama's open source code, companies must purchase Zama’s commercial patent license.
->
->Everything we do is open source and we are very transparent on what it means for our users, you can read more about how we monetize our open source products at Zama in [this blogpost](https://www.zama.ai/post/open-source).
-
-**What do I need to do if I want to use Zama’s technology for commercial purposes?**
->To commercially use Zama’s technology you need to be granted Zama’s patent license. Please contact us hello@zama.ai for more information.
-
-**Do you file IP on your technology?**
->Yes, all Zama’s technologies are patented.
-
-**Can you customize a solution for my specific use case?**
->We are open to collaborating and advancing the FHE space with our partners. If you have specific needs, please email us at hello@zama.ai.
-
+This software is distributed under the **BSD-3-Clause-Clear** license. If you have any questions, please contact us at hello@zama.ai.
 <p align="right">
   <a href="#about" > ↑ Back to top </a> 
 </p>
@@ -266,11 +240,7 @@ This software is distributed under the **BSD-3-Clause-Clear** license. Read [thi
 ## Support
 
 <a target="_blank" href="https://community.zama.ai">
-<picture>
-  <source media="(prefers-color-scheme: dark)" srcset="https://github.com/zama-ai/tfhe-rs/assets/157474013/08656d0a-3f44-4126-b8b6-8c601dff5380">
-  <source media="(prefers-color-scheme: light)" srcset="https://github.com/zama-ai/tfhe-rs/assets/157474013/1c9c9308-50ac-4aab-a4b9-469bb8c536a4">
-  <img alt="Support">
-</picture>
+  <img src="https://github.com/zama-ai/tfhe-rs/assets/157474013/8da6cf5b-51a0-4c86-9e75-fd0e4a4c64a4">
 </a>
 
 🌟 If you find this project helpful or interesting, please consider giving it a star on GitHub! Your support helps to grow the community and motivates further development. 

apps/trivium/README.md

@@ -15,6 +15,7 @@ Example of a Rust main below:
 ```rust
 use tfhe::{ConfigBuilder, generate_keys, FheBool};
 use tfhe::prelude::*;
+
 use tfhe_trivium::TriviumStream;
 
 fn get_hexadecimal_string_from_lsb_first_stream(a: Vec<bool>) -> String {
@@ -71,7 +72,7 @@ fn get_hexadecimal_string_from_lsb_first_stream(a: Vec<bool>) -> String {
 }
 
 fn main() {
-	let config = ConfigBuilder::default().build();
+	let config = ConfigBuilder::all_disabled().enable_default_bool().build();
 	let (client_key, server_key) = generate_keys(config);
 
 	let key_string = "0053A6F94C9FF24598EB".to_string();
@@ -138,12 +139,14 @@ Example code:
 ```rust
 use tfhe::shortint::prelude::*;
 use tfhe::shortint::CastingKey;
+
 use tfhe::{ConfigBuilder, generate_keys, FheUint64};
 use tfhe::prelude::*;
+
 use tfhe_trivium::TriviumStreamShortint;
 
 fn test_shortint() {
-	let config = ConfigBuilder::default().build();
+	let config = ConfigBuilder::all_disabled().enable_default_integers().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));

apps/trivium/benches/kreyvium_bool.rs

@@ -1,8 +1,10 @@
-use criterion::Criterion;
 use tfhe::prelude::*;
 use tfhe::{generate_keys, ConfigBuilder, FheBool};
+
 use tfhe_trivium::KreyviumStream;
 
+use criterion::Criterion;
+
 pub fn kreyvium_bool_gen(c: &mut Criterion) {
     let config = ConfigBuilder::default().build();
     let (client_key, server_key) = generate_keys(config);

apps/trivium/benches/kreyvium_byte.rs

@@ -1,11 +1,14 @@
-use criterion::Criterion;
 use tfhe::prelude::*;
 use tfhe::{generate_keys, ConfigBuilder, FheUint64, FheUint8};
+
 use tfhe_trivium::{KreyviumStreamByte, TransCiphering};
 
-pub fn kreyvium_byte_gen(c: &mut Criterion) {
-    let config = ConfigBuilder::default().build();
+use criterion::Criterion;
 
+pub fn kreyvium_byte_gen(c: &mut Criterion) {
+    let config = ConfigBuilder::default()
+        .enable_function_evaluation()
+        .build();
     let (client_key, server_key) = generate_keys(config);
 
     let key_string = "0053A6F94C9FF24598EB000000000000".to_string();
@@ -32,8 +35,9 @@ pub fn kreyvium_byte_gen(c: &mut Criterion) {
 }
 
 pub fn kreyvium_byte_trans(c: &mut Criterion) {
-    let config = ConfigBuilder::default().build();
-
+    let config = ConfigBuilder::default()
+        .enable_function_evaluation()
+        .build();
     let (client_key, server_key) = generate_keys(config);
 
     let key_string = "0053A6F94C9FF24598EB000000000000".to_string();
@@ -61,8 +65,9 @@ pub fn kreyvium_byte_trans(c: &mut Criterion) {
 }
 
 pub fn kreyvium_byte_warmup(c: &mut Criterion) {
-    let config = ConfigBuilder::default().build();
-
+    let config = ConfigBuilder::default()
+        .enable_function_evaluation()
+        .build();
     let (client_key, server_key) = generate_keys(config);
 
     let key_string = "0053A6F94C9FF24598EB000000000000".to_string();

apps/trivium/benches/kreyvium_shortint.rs

@@ -1,9 +1,12 @@
-use criterion::Criterion;
 use tfhe::prelude::*;
 use tfhe::shortint::prelude::*;
+use tfhe::shortint::KeySwitchingKey;
 use tfhe::{generate_keys, ConfigBuilder, FheUint64};
+
 use tfhe_trivium::{KreyviumStreamShortint, TransCiphering};
 
+use criterion::Criterion;
+
 pub fn kreyvium_shortint_warmup(c: &mut Criterion) {
     let config = ConfigBuilder::default().build();
     let (hl_client_key, hl_server_key) = generate_keys(config);
@@ -13,7 +16,7 @@ pub fn kreyvium_shortint_warmup(c: &mut Criterion) {
     let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
-        (&client_key, Some(&server_key)),
+        (&client_key, &server_key),
         (&underlying_ck, &underlying_sk),
         PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
@@ -63,7 +66,7 @@ pub fn kreyvium_shortint_gen(c: &mut Criterion) {
     let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
-        (&client_key, Some(&server_key)),
+        (&client_key, &server_key),
         (&underlying_ck, &underlying_sk),
         PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
@@ -108,7 +111,7 @@ pub fn kreyvium_shortint_trans(c: &mut Criterion) {
     let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
-        (&client_key, Some(&server_key)),
+        (&client_key, &server_key),
         (&underlying_ck, &underlying_sk),
         PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );

apps/trivium/benches/trivium_bool.rs

@@ -1,8 +1,10 @@
-use criterion::Criterion;
 use tfhe::prelude::*;
 use tfhe::{generate_keys, ConfigBuilder, FheBool};
+
 use tfhe_trivium::TriviumStream;
 
+use criterion::Criterion;
+
 pub fn trivium_bool_gen(c: &mut Criterion) {
     let config = ConfigBuilder::default().build();
     let (client_key, server_key) = generate_keys(config);

apps/trivium/benches/trivium_byte.rs

@@ -1,8 +1,10 @@
-use criterion::Criterion;
 use tfhe::prelude::*;
 use tfhe::{generate_keys, ConfigBuilder, FheUint64, FheUint8};
+
 use tfhe_trivium::{TransCiphering, TriviumStreamByte};
 
+use criterion::Criterion;
+
 pub fn trivium_byte_gen(c: &mut Criterion) {
     let config = ConfigBuilder::default().build();
     let (client_key, server_key) = generate_keys(config);

apps/trivium/benches/trivium_shortint.rs

@@ -1,9 +1,12 @@
-use criterion::Criterion;
 use tfhe::prelude::*;
 use tfhe::shortint::prelude::*;
+use tfhe::shortint::KeySwitchingKey;
 use tfhe::{generate_keys, ConfigBuilder, FheUint64};
+
 use tfhe_trivium::{TransCiphering, TriviumStreamShortint};
 
+use criterion::Criterion;
+
 pub fn trivium_shortint_warmup(c: &mut Criterion) {
     let config = ConfigBuilder::default().build();
     let (hl_client_key, hl_server_key) = generate_keys(config);
@@ -13,7 +16,7 @@ pub fn trivium_shortint_warmup(c: &mut Criterion) {
     let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
-        (&client_key, Some(&server_key)),
+        (&client_key, &server_key),
         (&underlying_ck, &underlying_sk),
         PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
@@ -63,7 +66,7 @@ pub fn trivium_shortint_gen(c: &mut Criterion) {
     let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
-        (&client_key, Some(&server_key)),
+        (&client_key, &server_key),
         (&underlying_ck, &underlying_sk),
         PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
@@ -108,7 +111,7 @@ pub fn trivium_shortint_trans(c: &mut Criterion) {
     let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
-        (&client_key, Some(&server_key)),
+        (&client_key, &server_key),
         (&underlying_ck, &underlying_sk),
         PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );

apps/trivium/src/kreyvium/kreyvium.rs

@@ -2,10 +2,12 @@
 //! for the representation of the inner bits.
 
 use crate::static_deque::StaticDeque;
-use rayon::prelude::*;
+
 use tfhe::prelude::*;
 use tfhe::{set_server_key, unset_server_key, FheBool, ServerKey};
 
+use rayon::prelude::*;
+
 /// Internal trait specifying which operations are necessary for KreyviumStream generic type
 pub trait KreyviumBoolInput<OpOutput>:
     Sized

apps/trivium/src/kreyvium/kreyvium_byte.rs

@@ -2,10 +2,12 @@
 //! for the representation of the inner bits.
 
 use crate::static_deque::{StaticByteDeque, StaticByteDequeInput};
-use rayon::prelude::*;
+
 use tfhe::prelude::*;
 use tfhe::{set_server_key, unset_server_key, FheUint8, ServerKey};
 
+use rayon::prelude::*;
+
 /// Internal trait specifying which operations are necessary for KreyviumStreamByte generic type
 pub trait KreyviumByteInput<OpOutput>:
     Sized
@@ -119,7 +121,7 @@ impl KreyviumStreamByte<FheUint8> {
         }
 
         // Key and iv are stored in reverse in their shift registers
-        let mut key = key_bytes.map(|b| b.reverse_bits());
+        let mut key = key_bytes.map(|b| b.map(|x| (x as u8).reverse_bits() as u64));
         let mut iv = iv_bytes.map(|x| FheUint8::encrypt_trivial(x.reverse_bits()));
         key.reverse();
         iv.reverse();

apps/trivium/src/kreyvium/kreyvium_shortint.rs

@@ -1,7 +1,9 @@
 use crate::static_deque::StaticDeque;
-use rayon::prelude::*;
+
 use tfhe::shortint::prelude::*;
 
+use rayon::prelude::*;
+
 /// KreyviumStreamShortint: a struct implementing the Kreyvium stream cipher, using a generic
 /// Ciphertext for the internal representation of bits (intended to represent a single bit). To be
 /// able to compute FHE operations, it also owns a ServerKey.
@@ -34,7 +36,7 @@ impl KreyviumStreamShortint {
         let mut c_register: [Ciphertext; 111] = [0; 111].map(|x| sk.create_trivial(x));
 
         for i in 0..93 {
-            a_register[i].clone_from(&key[128 - 93 + i]);
+            a_register[i] = key[128 - 93 + i].clone();
         }
         for i in 0..84 {
             b_register[i] = sk.create_trivial(iv[128 - 84 + i]);

apps/trivium/src/kreyvium/test.rs

@@ -1,7 +1,8 @@
-use crate::{KreyviumStream, KreyviumStreamByte, KreyviumStreamShortint, TransCiphering};
 use tfhe::prelude::*;
 use tfhe::{generate_keys, ConfigBuilder, FheBool, FheUint64, FheUint8};
 
+use crate::{KreyviumStream, KreyviumStreamByte, KreyviumStreamShortint, TransCiphering};
+
 // Values for these tests come from the github repo renaud1239/Kreyvium,
 // commit fd6828f68711276c25f55e605935028f5e843f43
 
@@ -224,7 +225,7 @@ fn kreyvium_test_shortint_long() {
     let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
-        (&client_key, Some(&server_key)),
+        (&client_key, &server_key),
         (&underlying_ck, &underlying_sk),
         PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );
@@ -299,8 +300,9 @@ fn kreyvium_test_clear_byte() {
 
 #[test]
 fn kreyvium_test_byte_long() {
-    let config = ConfigBuilder::default().build();
-
+    let config = ConfigBuilder::default()
+        .enable_function_evaluation()
+        .build();
     let (client_key, server_key) = generate_keys(config);
 
     let key_string = "0053A6F94C9FF24598EB000000000000".to_string();
@@ -337,8 +339,9 @@ fn kreyvium_test_byte_long() {
 
 #[test]
 fn kreyvium_test_fhe_byte_transciphering_long() {
-    let config = ConfigBuilder::default().build();
-
+    let config = ConfigBuilder::default()
+        .enable_function_evaluation()
+        .build();
     let (client_key, server_key) = generate_keys(config);
 
     let key_string = "0053A6F94C9FF24598EB000000000000".to_string();

apps/trivium/src/static_deque/mod.rs

@@ -1,6 +1,5 @@
 #[allow(clippy::module_inception)]
 mod static_deque;
 pub use static_deque::StaticDeque;
-
 mod static_byte_deque;
 pub use static_byte_deque::{StaticByteDeque, StaticByteDequeInput};

apps/trivium/src/static_deque/static_byte_deque.rs

@@ -4,6 +4,7 @@
 //! This is pretending to store bits, and allows accessing bits in chunks of 8 consecutive.
 
 use crate::static_deque::StaticDeque;
+
 use tfhe::FheUint8;
 
 /// Internal trait specifying which operations are needed by StaticByteDeque

apps/trivium/src/trans_ciphering/mod.rs

@@ -2,11 +2,13 @@
 //! when trans ciphering is available to them.
 
 use crate::{KreyviumStreamByte, KreyviumStreamShortint, TriviumStreamByte, TriviumStreamShortint};
-use rayon::prelude::*;
-use tfhe::prelude::*;
 use tfhe::shortint::Ciphertext;
+
+use tfhe::prelude::*;
 use tfhe::{set_server_key, unset_server_key, FheUint64, FheUint8, ServerKey};
 
+use rayon::prelude::*;
+
 /// Triat specifying the interface for trans ciphering a FheUint64 object. Since it is meant
 /// to be used with stream ciphers, encryption and decryption are by default the same.
 pub trait TransCiphering {

apps/trivium/src/trivium/test.rs

@@ -1,7 +1,8 @@
-use crate::{TransCiphering, TriviumStream, TriviumStreamByte, TriviumStreamShortint};
 use tfhe::prelude::*;
 use tfhe::{generate_keys, ConfigBuilder, FheBool, FheUint64, FheUint8};
 
+use crate::{TransCiphering, TriviumStream, TriviumStreamByte, TriviumStreamShortint};
+
 // Values for these tests come from the github repo cantora/avr-crypto-lib, commit 2a5b018,
 // file testvectors/trivium-80.80.test-vectors
 
@@ -360,7 +361,7 @@ fn trivium_test_shortint_long() {
     let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
 
     let ksk = KeySwitchingKey::new(
-        (&client_key, Some(&server_key)),
+        (&client_key, &server_key),
         (&underlying_ck, &underlying_sk),
         PARAM_KEYSWITCH_1_1_KS_PBS_TO_2_2_KS_PBS,
     );

apps/trivium/src/trivium/trivium_bool.rs

@@ -2,10 +2,12 @@
 //! for the representation of the inner bits.
 
 use crate::static_deque::StaticDeque;
-use rayon::prelude::*;
+
 use tfhe::prelude::*;
 use tfhe::{set_server_key, unset_server_key, FheBool, ServerKey};
 
+use rayon::prelude::*;
+
 /// Internal trait specifying which operations are necessary for TriviumStream generic type
 pub trait TriviumBoolInput<OpOutput>:
     Sized

apps/trivium/src/trivium/trivium_byte.rs

@@ -2,10 +2,12 @@
 //! for the representation of the inner bits.
 
 use crate::static_deque::{StaticByteDeque, StaticByteDequeInput};
-use rayon::prelude::*;
+
 use tfhe::prelude::*;
 use tfhe::{set_server_key, unset_server_key, FheUint8, ServerKey};
 
+use rayon::prelude::*;
+
 /// Internal trait specifying which operations are necessary for TriviumStreamByte generic type
 pub trait TriviumByteInput<OpOutput>:
     Sized

apps/trivium/src/trivium/trivium_shortint.rs

@@ -1,7 +1,9 @@
 use crate::static_deque::StaticDeque;
-use rayon::prelude::*;
+
 use tfhe::shortint::prelude::*;
 
+use rayon::prelude::*;
+
 /// TriviumStreamShortint: a struct implementing the Trivium stream cipher, using a generic
 /// Ciphertext for the internal representation of bits (intended to represent a single bit). To be
 /// able to compute FHE operations, it also owns a ServerKey.
@@ -32,7 +34,7 @@ impl TriviumStreamShortint {
         let mut c_register: [Ciphertext; 111] = [0; 111].map(|x| sk.create_trivial(x));
 
         for i in 0..80 {
-            a_register[93 - 80 + i].clone_from(&key[i]);
+            a_register[93 - 80 + i] = key[i].clone();
             b_register[84 - 80 + i] = sk.create_trivial(iv[i]);
         }
 

backends/tfhe-cuda-backend/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "tfhe-cuda-backend"
-version = "0.4.0-alpha.0"
+version = "0.2.0"
 edition = "2021"
 authors = ["Zama team"]
 license = "BSD-3-Clause-Clear"
@@ -13,4 +13,6 @@ keywords = ["fully", "homomorphic", "encryption", "fhe", "cryptography"]
 
 [build-dependencies]
 cmake = { version = "0.1" }
-pkg-config = { version = "0.3" }
+
+[dependencies]
+thiserror = "1.0"

backends/tfhe-cuda-backend/build.rs

@@ -2,30 +2,7 @@ use std::env;
 use std::process::Command;
 
 fn main() {
-    if let Ok(val) = env::var("DOCS_RS") {
-        if val.parse::<u32>() == Ok(1) {
-            return;
-        }
-    }
-
-    // This is a workaround to the current nightly toolchain (2024-06-27 which started with
-    // toolchain 2024-05-05) build issue
-    // Essentially if cbindgen is running, a wrong argument ends up forwarded to the cuda backend
-    // "make" command during macro expansions for TFHE-rs C API, crashing make for make < 4.4 and
-    // thus crashing the build
-    // On the other hand, this speeds up C API build greatly given we don't have macro expansions
-    // in the CUDA backend so this skips the second compilation of TFHE-rs for macro inspection by
-    // cbindgen
-    if std::env::var("_CBINDGEN_IS_RUNNING").is_ok() {
-        return;
-    }
-
     println!("Build tfhe-cuda-backend");
-    println!("cargo::rerun-if-changed=cuda/include");
-    println!("cargo::rerun-if-changed=cuda/src");
-    println!("cargo::rerun-if-changed=cuda/tests_and_benchmarks");
-    println!("cargo::rerun-if-changed=cuda/CMakeLists.txt");
-    println!("cargo::rerun-if-changed=src");
     if env::consts::OS == "linux" {
         let output = Command::new("./get_os_name.sh").output().unwrap();
         let distribution = String::from_utf8(output.stdout).unwrap();
@@ -38,15 +15,7 @@ fn main() {
         let dest = cmake::build("cuda");
         println!("cargo:rustc-link-search=native={}", dest.display());
         println!("cargo:rustc-link-lib=static=tfhe_cuda_backend");
-
-        // Try to find the cuda libs with pkg-config, default to the path used by the nvidia runfile
-        if pkg_config::Config::new()
-            .atleast_version("10")
-            .probe("cuda")
-            .is_err()
-        {
-            println!("cargo:rustc-link-search=native=/usr/local/cuda/lib64");
-        }
+        println!("cargo:rustc-link-search=native=/usr/local/cuda/lib64");
         println!("cargo:rustc-link-lib=gomp");
         println!("cargo:rustc-link-lib=cudart");
         println!("cargo:rustc-link-search=native=/usr/lib/x86_64-linux-gnu/");

backends/tfhe-cuda-backend/cuda/.gitignore

@@ -1,2 +1 @@
 /build/
-include/cuda_config.h

backends/tfhe-cuda-backend/cuda/CMakeLists.txt

@@ -58,15 +58,10 @@ set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${OpenMP_CXX_FLAGS}")
 set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Xcompiler ${OpenMP_CXX_FLAGS}")
 if(${CUDA_SUCCESS})
   set(CMAKE_CUDA_ARCHITECTURES native)
-  string(REPLACE "-arch=sm_" "" CUDA_ARCH "${ARCH}")
-  set(CUDA_ARCH "${CUDA_ARCH}0")
 else()
   set(CMAKE_CUDA_ARCHITECTURES 70)
-  set(CUDA_ARCH "700")
 endif()
 
-add_compile_definitions(CUDA_ARCH=${CUDA_ARCH})
-
 # in production, should use -arch=sm_70 --ptxas-options=-v to see register spills -lineinfo for better debugging
 set(CMAKE_CUDA_FLAGS
     "${CMAKE_CUDA_FLAGS} -ccbin ${CMAKE_CXX_COMPILER} -O3 \

backends/tfhe-cuda-backend/cuda/format_tfhe_cuda_backend.sh

@@ -1,4 +1,4 @@
-#!/usr/bin/env bash
+#!/bin/bash
 
 set -e
 
@@ -6,14 +6,14 @@ while getopts ":c" option; do
   case $option in
     c)
       # code to execute when flag1 is provided
-      find ./{include,src,tests_and_benchmarks/include,tests_and_benchmarks/tests,tests_and_benchmarks/benchmarks} -iregex '^.*\.\(cpp\|cu\|h\|cuh\)$' -print | xargs clang-format-15 -i -style='file' --dry-run --Werror
+      find ./{include,src,tests_and_benchmarks/tests,tests_and_benchmarks/benchmarks} -iregex '^.*\.\(cpp\|cu\|h\|cuh\)$' -print | xargs clang-format-15 -i -style='file' --dry-run --Werror
       cmake-format -i CMakeLists.txt -c .cmake-format-config.py
-      find ./{include,src,tests_and_benchmarks/include,tests_and_benchmarks/tests,tests_and_benchmarks/benchmarks} -type f -name "CMakeLists.txt" | xargs -I % sh -c 'cmake-format -i % -c .cmake-format-config.py'
+      find ./{include,src,tests_and_benchmarks/tests,tests_and_benchmarks/benchmarks} -type f -name "CMakeLists.txt" | xargs -I % sh -c 'cmake-format -i % -c .cmake-format-config.py'
       git diff --exit-code
       exit
       ;;
   esac
 done
-find ./{include,src,tests_and_benchmarks/include,tests_and_benchmarks/tests,tests_and_benchmarks/benchmarks} -iregex '^.*\.\(cpp\|cu\|h\|cuh\)$' -print | xargs clang-format-15 -i -style='file'
+find ./{include,src,tests_and_benchmarks/tests,tests_and_benchmarks/benchmarks} -iregex '^.*\.\(cpp\|cu\|h\|cuh\)$' -print | xargs clang-format-15 -i -style='file'
 cmake-format -i CMakeLists.txt -c .cmake-format-config.py
-find ./{include,src,tests_and_benchmarks/include,tests_and_benchmarks/tests,tests_and_benchmarks/benchmarks} -type f -name "CMakeLists.txt" | xargs -I % sh -c 'cmake-format -i % -c .cmake-format-config.py'
+find ./{include,src,tests_and_benchmarks/tests,tests_and_benchmarks/benchmarks} -type f -name "CMakeLists.txt" | xargs -I % sh -c 'cmake-format -i % -c .cmake-format-config.py'

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

@@ -0,0 +1,328 @@
+#ifndef CUDA_BOOTSTRAP_H
+#define CUDA_BOOTSTRAP_H
+
+#include "device.h"
+#include <cstdint>
+
+enum PBS_TYPE { MULTI_BIT = 0, LOW_LAT = 1, AMORTIZED = 2 };
+enum PBS_VARIANT { DEFAULT = 0, FAST = 1 };
+
+extern "C" {
+void cuda_fourier_polynomial_mul(void *input1, void *input2, void *output,
+                                 cuda_stream_t *stream,
+                                 uint32_t polynomial_size,
+                                 uint32_t total_polynomials);
+
+void cuda_convert_lwe_bootstrap_key_32(void *dest, void *src,
+                                       cuda_stream_t *stream,
+                                       uint32_t input_lwe_dim,
+                                       uint32_t glwe_dim, uint32_t level_count,
+                                       uint32_t polynomial_size);
+
+void cuda_convert_lwe_bootstrap_key_64(void *dest, void *src,
+                                       cuda_stream_t *stream,
+                                       uint32_t input_lwe_dim,
+                                       uint32_t glwe_dim, uint32_t level_count,
+                                       uint32_t polynomial_size);
+
+void scratch_cuda_bootstrap_amortized_32(
+    cuda_stream_t *stream, int8_t **pbs_buffer, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t input_lwe_ciphertext_count,
+    uint32_t max_shared_memory, bool allocate_gpu_memory);
+
+void scratch_cuda_bootstrap_amortized_64(
+    cuda_stream_t *stream, int8_t **pbs_buffer, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t input_lwe_ciphertext_count,
+    uint32_t max_shared_memory, bool allocate_gpu_memory);
+
+void cuda_bootstrap_amortized_lwe_ciphertext_vector_32(
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_output_indexes,
+    void *lut_vector, void *lut_vector_indexes, void *lwe_array_in,
+    void *lwe_input_indexes, void *bootstrapping_key, int8_t *pbs_buffer,
+    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t base_log, uint32_t level_count, uint32_t num_samples,
+    uint32_t num_luts, uint32_t lwe_idx, uint32_t max_shared_memory);
+
+void cuda_bootstrap_amortized_lwe_ciphertext_vector_64(
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_output_indexes,
+    void *lut_vector, void *lut_vector_indexes, void *lwe_array_in,
+    void *lwe_input_indexes, void *bootstrapping_key, int8_t *pbs_buffer,
+    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t base_log, uint32_t level_count, uint32_t num_samples,
+    uint32_t num_luts, uint32_t lwe_idx, uint32_t max_shared_memory);
+
+void cleanup_cuda_bootstrap_amortized(cuda_stream_t *stream,
+                                      int8_t **pbs_buffer);
+
+void scratch_cuda_bootstrap_low_latency_32(
+    cuda_stream_t *stream, int8_t **buffer, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t level_count,
+    uint32_t input_lwe_ciphertext_count, uint32_t max_shared_memory,
+    bool allocate_gpu_memory);
+
+void scratch_cuda_bootstrap_low_latency_64(
+    cuda_stream_t *stream, int8_t **buffer, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t level_count,
+    uint32_t input_lwe_ciphertext_count, uint32_t max_shared_memory,
+    bool allocate_gpu_memory);
+
+void cuda_bootstrap_low_latency_lwe_ciphertext_vector_32(
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_output_indexes,
+    void *lut_vector, void *lut_vector_indexes, void *lwe_array_in,
+    void *lwe_input_indexes, void *bootstrapping_key, int8_t *buffer,
+    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t base_log, uint32_t level_count, uint32_t num_samples,
+    uint32_t num_luts, uint32_t lwe_idx, uint32_t max_shared_memory);
+
+void cuda_bootstrap_low_latency_lwe_ciphertext_vector_64(
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_output_indexes,
+    void *lut_vector, void *lut_vector_indexes, void *lwe_array_in,
+    void *lwe_input_indexes, void *bootstrapping_key, int8_t *buffer,
+    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t base_log, uint32_t level_count, uint32_t num_samples,
+    uint32_t num_luts, uint32_t lwe_idx, uint32_t max_shared_memory);
+
+void cleanup_cuda_bootstrap_low_latency_32(cuda_stream_t *stream,
+                                           int8_t **pbs_buffer);
+
+void cleanup_cuda_bootstrap_low_latency_64(cuda_stream_t *stream,
+                                           int8_t **pbs_buffer);
+
+uint64_t get_buffer_size_bootstrap_amortized_64(
+    uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t input_lwe_ciphertext_count, uint32_t max_shared_memory);
+
+uint64_t get_buffer_size_bootstrap_low_latency_64(
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
+    uint32_t input_lwe_ciphertext_count, uint32_t max_shared_memory);
+}
+
+template <typename Torus>
+__host__ __device__ uint64_t
+get_buffer_size_full_sm_bootstrap_low_latency_step_one(
+    uint32_t polynomial_size) {
+  return sizeof(Torus) * polynomial_size +      // accumulator_rotated
+         sizeof(double2) * polynomial_size / 2; // accumulator fft
+}
+template <typename Torus>
+__host__ __device__ uint64_t
+get_buffer_size_full_sm_bootstrap_low_latency_step_two(
+    uint32_t polynomial_size) {
+  return sizeof(Torus) * polynomial_size +      // accumulator
+         sizeof(double2) * polynomial_size / 2; // accumulator fft
+}
+
+template <typename Torus>
+__host__ __device__ uint64_t
+get_buffer_size_partial_sm_bootstrap_low_latency(uint32_t polynomial_size) {
+  return sizeof(double2) * polynomial_size / 2; // accumulator fft
+}
+
+template <typename Torus>
+__host__ __device__ uint64_t
+get_buffer_size_full_sm_bootstrap_fast_low_latency(uint32_t polynomial_size) {
+  return sizeof(Torus) * polynomial_size +      // accumulator_rotated
+         sizeof(Torus) * polynomial_size +      // accumulator
+         sizeof(double2) * polynomial_size / 2; // accumulator fft
+}
+
+template <typename Torus>
+__host__ __device__ uint64_t
+get_buffer_size_partial_sm_bootstrap_fast_low_latency(
+    uint32_t polynomial_size) {
+  return sizeof(double2) * polynomial_size / 2; // accumulator fft mask & body
+}
+
+template <typename Torus, PBS_TYPE pbs_type> struct pbs_buffer;
+
+template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::LOW_LAT> {
+  int8_t *d_mem;
+
+  Torus *global_accumulator;
+  double2 *global_accumulator_fft;
+
+  PBS_VARIANT pbs_variant;
+
+  pbs_buffer(cuda_stream_t *stream, uint32_t glwe_dimension,
+             uint32_t polynomial_size, uint32_t level_count,
+             uint32_t input_lwe_ciphertext_count, PBS_VARIANT pbs_variant,
+             bool allocate_gpu_memory) {
+    this->pbs_variant = pbs_variant;
+
+    auto max_shared_memory = cuda_get_max_shared_memory(stream->gpu_index);
+
+    if (allocate_gpu_memory) {
+      switch (pbs_variant) {
+      case PBS_VARIANT::DEFAULT: {
+        uint64_t full_sm_step_one =
+            get_buffer_size_full_sm_bootstrap_low_latency_step_one<Torus>(
+                polynomial_size);
+        uint64_t full_sm_step_two =
+            get_buffer_size_full_sm_bootstrap_low_latency_step_two<Torus>(
+                polynomial_size);
+        uint64_t partial_sm =
+            get_buffer_size_partial_sm_bootstrap_low_latency<Torus>(
+                polynomial_size);
+
+        uint64_t partial_dm_step_one = full_sm_step_one - partial_sm;
+        uint64_t partial_dm_step_two = full_sm_step_two - partial_sm;
+        uint64_t full_dm = full_sm_step_one;
+
+        uint64_t device_mem = 0;
+        if (max_shared_memory < partial_sm) {
+          device_mem = full_dm * input_lwe_ciphertext_count * level_count *
+                       (glwe_dimension + 1);
+        } else if (max_shared_memory < full_sm_step_two) {
+          device_mem =
+              (partial_dm_step_two + partial_dm_step_one * level_count) *
+              input_lwe_ciphertext_count * (glwe_dimension + 1);
+        } else if (max_shared_memory < full_sm_step_one) {
+          device_mem = partial_dm_step_one * input_lwe_ciphertext_count *
+                       level_count * (glwe_dimension + 1);
+        }
+        // Otherwise, both kernels run all in shared memory
+        d_mem = (int8_t *)cuda_malloc_async(device_mem, stream);
+
+        global_accumulator_fft = (double2 *)cuda_malloc_async(
+            (glwe_dimension + 1) * level_count * input_lwe_ciphertext_count *
+                (polynomial_size / 2) * sizeof(double2),
+            stream);
+
+        global_accumulator = (Torus *)cuda_malloc_async(
+            (glwe_dimension + 1) * input_lwe_ciphertext_count *
+                polynomial_size * sizeof(Torus),
+            stream);
+      } break;
+      case PBS_VARIANT::FAST: {
+        uint64_t full_sm =
+            get_buffer_size_full_sm_bootstrap_fast_low_latency<Torus>(
+                polynomial_size);
+        uint64_t partial_sm =
+            get_buffer_size_partial_sm_bootstrap_fast_low_latency<Torus>(
+                polynomial_size);
+
+        uint64_t partial_dm = full_sm - partial_sm;
+        uint64_t full_dm = full_sm;
+        uint64_t device_mem = 0;
+
+        if (max_shared_memory < partial_sm) {
+          device_mem = full_dm * input_lwe_ciphertext_count * level_count *
+                       (glwe_dimension + 1);
+        } else if (max_shared_memory < full_sm) {
+          device_mem = partial_dm * input_lwe_ciphertext_count * level_count *
+                       (glwe_dimension + 1);
+        }
+
+        // Otherwise, both kernels run all in shared memory
+        d_mem = (int8_t *)cuda_malloc_async(device_mem, stream);
+
+        global_accumulator_fft = (double2 *)cuda_malloc_async(
+            (glwe_dimension + 1) * level_count * input_lwe_ciphertext_count *
+                polynomial_size / 2 * sizeof(double2),
+            stream);
+      } break;
+      default:
+        PANIC("Cuda error (PBS): unsupported implementation variant.")
+      }
+    }
+  }
+
+  void release(cuda_stream_t *stream) {
+    cuda_drop_async(d_mem, stream);
+    cuda_drop_async(global_accumulator_fft, stream);
+
+    if (pbs_variant == DEFAULT)
+      cuda_drop_async(global_accumulator, stream);
+  }
+};
+
+template <typename Torus>
+__host__ __device__ uint64_t get_buffer_size_bootstrap_fast_low_latency(
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
+    uint32_t input_lwe_ciphertext_count, uint32_t max_shared_memory) {
+
+  uint64_t full_sm = get_buffer_size_full_sm_bootstrap_fast_low_latency<Torus>(
+      polynomial_size);
+  uint64_t partial_sm =
+      get_buffer_size_partial_sm_bootstrap_fast_low_latency<Torus>(
+          polynomial_size);
+  uint64_t partial_dm = full_sm - partial_sm;
+  uint64_t full_dm = full_sm;
+  uint64_t device_mem = 0;
+  if (max_shared_memory < partial_sm) {
+    device_mem = full_dm * input_lwe_ciphertext_count * level_count *
+                 (glwe_dimension + 1);
+  } else if (max_shared_memory < full_sm) {
+    device_mem = partial_dm * input_lwe_ciphertext_count * level_count *
+                 (glwe_dimension + 1);
+  }
+  uint64_t buffer_size = device_mem + (glwe_dimension + 1) * level_count *
+                                          input_lwe_ciphertext_count *
+                                          polynomial_size / 2 * sizeof(double2);
+  return buffer_size + buffer_size % sizeof(double2);
+}
+
+template <typename Torus>
+bool has_support_to_cuda_bootstrap_fast_low_latency(uint32_t glwe_dimension,
+                                                    uint32_t polynomial_size,
+                                                    uint32_t level_count,
+                                                    uint32_t num_samples,
+                                                    uint32_t max_shared_memory);
+
+template <typename Torus>
+void cuda_bootstrap_fast_low_latency_lwe_ciphertext_vector(
+    cuda_stream_t *stream, Torus *lwe_array_out, Torus *lwe_output_indexes,
+    Torus *lut_vector, Torus *lut_vector_indexes, Torus *lwe_array_in,
+    Torus *lwe_input_indexes, double2 *bootstrapping_key,
+    pbs_buffer<Torus, LOW_LAT> *buffer, uint32_t lwe_dimension,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t base_log,
+    uint32_t level_count, uint32_t num_samples, uint32_t num_luts,
+    uint32_t lwe_idx, uint32_t max_shared_memory);
+
+template <typename Torus>
+void cuda_bootstrap_low_latency_lwe_ciphertext_vector(
+    cuda_stream_t *stream, Torus *lwe_array_out, Torus *lwe_output_indexes,
+    Torus *lut_vector, Torus *lut_vector_indexes, Torus *lwe_array_in,
+    Torus *lwe_input_indexes, double2 *bootstrapping_key,
+    pbs_buffer<Torus, LOW_LAT> *buffer, uint32_t lwe_dimension,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t base_log,
+    uint32_t level_count, uint32_t num_samples, uint32_t num_luts,
+    uint32_t lwe_idx, uint32_t max_shared_memory);
+
+template <typename Torus, typename STorus>
+void scratch_cuda_fast_bootstrap_low_latency(
+    cuda_stream_t *stream, pbs_buffer<Torus, LOW_LAT> **pbs_buffer,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
+    uint32_t input_lwe_ciphertext_count, uint32_t max_shared_memory,
+    bool allocate_gpu_memory);
+
+template <typename Torus, typename STorus>
+void scratch_cuda_bootstrap_low_latency(
+    cuda_stream_t *stream, pbs_buffer<Torus, LOW_LAT> **buffer,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
+    uint32_t input_lwe_ciphertext_count, uint32_t max_shared_memory,
+    bool allocate_gpu_memory);
+
+#ifdef __CUDACC__
+__device__ inline int get_start_ith_ggsw(int i, uint32_t polynomial_size,
+                                         int glwe_dimension,
+                                         uint32_t level_count);
+
+template <typename T>
+__device__ T *get_ith_mask_kth_block(T *ptr, int i, int k, int level,
+                                     uint32_t polynomial_size,
+                                     int glwe_dimension, uint32_t level_count);
+
+template <typename T>
+__device__ T *get_ith_body_kth_block(T *ptr, int i, int k, int level,
+                                     uint32_t polynomial_size,
+                                     int glwe_dimension, uint32_t level_count);
+
+template <typename T>
+__device__ T *get_multi_bit_ith_lwe_gth_group_kth_block(
+    T *ptr, int g, int i, int k, int level, uint32_t grouping_factor,
+    uint32_t polynomial_size, uint32_t glwe_dimension, uint32_t level_count);
+
+#endif
+
+#endif // CUDA_BOOTSTRAP_H

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

@@ -0,0 +1,155 @@
+#ifndef CUDA_MULTI_BIT_H
+#define CUDA_MULTI_BIT_H
+
+#include "bootstrap.h"
+#include <cstdint>
+
+extern "C" {
+
+bool has_support_to_cuda_bootstrap_fast_multi_bit(uint32_t glwe_dimension,
+                                                  uint32_t polynomial_size,
+                                                  uint32_t level_count,
+                                                  uint32_t num_samples,
+                                                  uint32_t max_shared_memory);
+
+void cuda_convert_lwe_multi_bit_bootstrap_key_64(
+    void *dest, void *src, cuda_stream_t *stream, uint32_t input_lwe_dim,
+    uint32_t glwe_dim, uint32_t level_count, uint32_t polynomial_size,
+    uint32_t grouping_factor);
+
+void scratch_cuda_multi_bit_pbs_64(
+    cuda_stream_t *stream, int8_t **pbs_buffer, uint32_t lwe_dimension,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
+    uint32_t grouping_factor, uint32_t input_lwe_ciphertext_count,
+    uint32_t max_shared_memory, bool allocate_gpu_memory,
+    uint32_t chunk_size = 0);
+
+void cuda_multi_bit_pbs_lwe_ciphertext_vector_64(
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_output_indexes,
+    void *lut_vector, void *lut_vector_indexes, void *lwe_array_in,
+    void *lwe_input_indexes, void *bootstrapping_key, int8_t *buffer,
+    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t grouping_factor, uint32_t base_log, uint32_t level_count,
+    uint32_t num_samples, uint32_t num_luts, uint32_t lwe_idx,
+    uint32_t max_shared_memory, uint32_t lwe_chunk_size = 0);
+
+void scratch_cuda_generic_multi_bit_pbs_64(
+    cuda_stream_t *stream, int8_t **pbs_buffer, uint32_t lwe_dimension,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
+    uint32_t grouping_factor, uint32_t input_lwe_ciphertext_count,
+    uint32_t max_shared_memory, bool allocate_gpu_memory,
+    uint32_t lwe_chunk_size = 0);
+
+void cuda_generic_multi_bit_pbs_lwe_ciphertext_vector_64(
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_output_indexes,
+    void *lut_vector, void *lut_vector_indexes, void *lwe_array_in,
+    void *lwe_input_indexes, void *bootstrapping_key, int8_t *pbs_buffer,
+    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t grouping_factor, uint32_t base_log, uint32_t level_count,
+    uint32_t num_samples, uint32_t num_luts, uint32_t lwe_idx,
+    uint32_t max_shared_memory, uint32_t lwe_chunk_size = 0);
+
+void cleanup_cuda_multi_bit_pbs_32(cuda_stream_t *stream, int8_t **pbs_buffer);
+void cleanup_cuda_multi_bit_pbs_64(cuda_stream_t *stream, int8_t **pbs_buffer);
+}
+
+template <typename Torus, typename STorus>
+void scratch_cuda_fast_multi_bit_pbs(
+    cuda_stream_t *stream, pbs_buffer<Torus, MULTI_BIT> **pbs_buffer,
+    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t level_count, uint32_t grouping_factor,
+    uint32_t input_lwe_ciphertext_count, uint32_t max_shared_memory,
+    bool allocate_gpu_memory, uint32_t lwe_chunk_size = 0);
+
+template <typename Torus>
+void cuda_fast_multi_bit_pbs_lwe_ciphertext_vector(
+    cuda_stream_t *stream, Torus *lwe_array_out, Torus *lwe_output_indexes,
+    Torus *lut_vector, Torus *lut_vector_indexes, Torus *lwe_array_in,
+    Torus *lwe_input_indexes, Torus *bootstrapping_key,
+    pbs_buffer<Torus, MULTI_BIT> *pbs_buffer, uint32_t lwe_dimension,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t grouping_factor,
+    uint32_t base_log, uint32_t level_count, uint32_t num_samples,
+    uint32_t num_luts, uint32_t lwe_idx, uint32_t max_shared_memory,
+    uint32_t lwe_chunk_size = 0);
+
+template <typename Torus, typename STorus>
+void scratch_cuda_multi_bit_pbs(
+    cuda_stream_t *stream, pbs_buffer<Torus, MULTI_BIT> **pbs_buffer,
+    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t level_count, uint32_t grouping_factor,
+    uint32_t input_lwe_ciphertext_count, uint32_t max_shared_memory,
+    bool allocate_gpu_memory, uint32_t lwe_chunk_size = 0);
+
+template <typename Torus>
+void cuda_multi_bit_pbs_lwe_ciphertext_vector(
+    cuda_stream_t *stream, Torus *lwe_array_out, Torus *lwe_output_indexes,
+    Torus *lut_vector, Torus *lut_vector_indexes, Torus *lwe_array_in,
+    Torus *lwe_input_indexes, Torus *bootstrapping_key,
+    pbs_buffer<Torus, MULTI_BIT> *pbs_buffer, uint32_t lwe_dimension,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t grouping_factor,
+    uint32_t base_log, uint32_t level_count, uint32_t num_samples,
+    uint32_t num_luts, uint32_t lwe_idx, uint32_t max_shared_memory,
+    uint32_t lwe_chunk_size = 0);
+
+template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::MULTI_BIT> {
+  double2 *keybundle_fft;
+  Torus *global_accumulator;
+  double2 *global_accumulator_fft;
+
+  PBS_VARIANT pbs_variant;
+
+  pbs_buffer(cuda_stream_t *stream, uint32_t glwe_dimension,
+             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) {
+    this->pbs_variant = pbs_variant;
+    auto max_shared_memory = cuda_get_max_shared_memory(stream->gpu_index);
+
+    if (allocate_gpu_memory) {
+      switch (pbs_variant) {
+      case DEFAULT:
+      case FAST:
+        keybundle_fft = (double2 *)cuda_malloc_async(
+            input_lwe_ciphertext_count * lwe_chunk_size * level_count *
+                (glwe_dimension + 1) * (glwe_dimension + 1) *
+                (polynomial_size / 2) * sizeof(double2),
+            stream);
+        global_accumulator = (Torus *)cuda_malloc_async(
+            input_lwe_ciphertext_count * (glwe_dimension + 1) *
+                polynomial_size * sizeof(Torus),
+            stream);
+        global_accumulator_fft = (double2 *)cuda_malloc_async(
+            input_lwe_ciphertext_count * (glwe_dimension + 1) * level_count *
+                (polynomial_size / 2) * sizeof(double2),
+            stream);
+        break;
+      default:
+        PANIC("Cuda error (PBS): unsupported implementation variant.")
+      }
+    }
+  }
+
+  void release(cuda_stream_t *stream) {
+    cuda_drop_async(keybundle_fft, stream);
+    cuda_drop_async(global_accumulator, stream);
+    cuda_drop_async(global_accumulator_fft, stream);
+  }
+};
+
+#ifdef __CUDACC__
+__host__ uint32_t get_lwe_chunk_size(uint32_t lwe_dimension,
+                                     uint32_t level_count,
+                                     uint32_t glwe_dimension,
+                                     uint32_t num_samples);
+
+__host__ uint32_t get_average_lwe_chunk_size(uint32_t lwe_dimension,
+                                             uint32_t level_count,
+                                             uint32_t glwe_dimension,
+                                             uint32_t ct_count);
+
+__host__ uint64_t get_max_buffer_size_multibit_bootstrap(
+    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t level_count, uint32_t max_input_lwe_ciphertext_count);
+#endif
+
+#endif // CUDA_MULTI_BIT_H

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

@@ -1,25 +1,18 @@
 #ifndef CUDA_CIPHERTEXT_H
 #define CUDA_CIPHERTEXT_H
 
-#include "device.h"
 #include <cstdint>
 
 extern "C" {
-void cuda_convert_lwe_ciphertext_vector_to_gpu_64(void *stream,
+void cuda_convert_lwe_ciphertext_vector_to_gpu_64(void *dest, void *src,
+                                                  void *v_stream,
                                                   uint32_t gpu_index,
-                                                  void *dest, void *src,
                                                   uint32_t number_of_cts,
                                                   uint32_t lwe_dimension);
-void cuda_convert_lwe_ciphertext_vector_to_cpu_64(void *stream,
+void cuda_convert_lwe_ciphertext_vector_to_cpu_64(void *dest, void *src,
+                                                  void *v_stream,
                                                   uint32_t gpu_index,
-                                                  void *dest, void *src,
                                                   uint32_t number_of_cts,
                                                   uint32_t lwe_dimension);
-
-void cuda_glwe_sample_extract_64(void *stream, uint32_t gpu_index,
-                                 void *lwe_array_out, void *glwe_array_in,
-                                 uint32_t *nth_array, uint32_t num_glwes,
-                                 uint32_t glwe_dimension,
-                                 uint32_t polynomial_size);
 };
 #endif

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

@@ -6,9 +6,9 @@
 #include <cstdlib>
 #include <cstring>
 #include <cuda_runtime.h>
-#include <vector>
 
 #define synchronize_threads_in_block() __syncthreads()
+
 extern "C" {
 
 #define check_cuda_error(ans)                                                  \
@@ -27,33 +27,47 @@ inline void cuda_error(cudaError_t code, const char *file, int line) {
     std::abort();                                                              \
   }
 
-cudaStream_t cuda_create_stream(uint32_t gpu_index);
+struct cuda_stream_t {
+  cudaStream_t stream;
+  uint32_t gpu_index;
 
-void cuda_destroy_stream(cudaStream_t stream, uint32_t gpu_index);
+  cuda_stream_t(uint32_t gpu_index) {
+    this->gpu_index = gpu_index;
 
-void cuda_synchronize_stream(cudaStream_t stream, uint32_t gpu_index);
+    check_cuda_error(cudaStreamCreate(&stream));
+  }
+
+  void release() {
+    check_cuda_error(cudaSetDevice(gpu_index));
+    check_cuda_error(cudaStreamDestroy(stream));
+  }
+
+  void synchronize() { check_cuda_error(cudaStreamSynchronize(stream)); }
+};
+
+cuda_stream_t *cuda_create_stream(uint32_t gpu_index);
+
+void cuda_destroy_stream(cuda_stream_t *stream);
 
 void *cuda_malloc(uint64_t size, uint32_t gpu_index);
 
-void *cuda_malloc_async(uint64_t size, cudaStream_t stream, uint32_t gpu_index);
+void *cuda_malloc_async(uint64_t size, cuda_stream_t *stream);
 
 void cuda_check_valid_malloc(uint64_t size, uint32_t gpu_index);
 
 bool cuda_check_support_cooperative_groups();
 
-bool cuda_check_support_thread_block_clusters();
-
 void cuda_memcpy_async_to_gpu(void *dest, void *src, uint64_t size,
-                              cudaStream_t stream, uint32_t gpu_index);
+                              cuda_stream_t *stream);
 
 void cuda_memcpy_async_gpu_to_gpu(void *dest, void *src, uint64_t size,
-                                  cudaStream_t stream, uint32_t gpu_index);
+                                  cuda_stream_t *stream);
 
 void cuda_memcpy_async_to_cpu(void *dest, const void *src, uint64_t size,
-                              cudaStream_t stream, uint32_t gpu_index);
+                              cuda_stream_t *stream);
 
 void cuda_memset_async(void *dest, uint64_t val, uint64_t size,
-                       cudaStream_t stream, uint32_t gpu_index);
+                       cuda_stream_t *stream);
 
 int cuda_get_number_of_gpus();
 
@@ -61,12 +75,20 @@ void cuda_synchronize_device(uint32_t gpu_index);
 
 void cuda_drop(void *ptr, uint32_t gpu_index);
 
-void cuda_drop_async(void *ptr, cudaStream_t stream, uint32_t gpu_index);
+void cuda_drop_async(void *ptr, cuda_stream_t *stream);
 
 int cuda_get_max_shared_memory(uint32_t gpu_index);
+
+void cuda_synchronize_stream(cuda_stream_t *stream);
+
+void cuda_stream_add_callback(cuda_stream_t *stream,
+                              cudaStreamCallback_t callback, void *user_data);
+
+void host_free_on_stream_callback(cudaStream_t stream, cudaError_t status,
+                                  void *host_pointer);
 }
 
 template <typename Torus>
-void cuda_set_value_async(cudaStream_t stream, uint32_t gpu_index,
-                          Torus *d_array, Torus value, Torus n);
+void cuda_set_value_async(cudaStream_t *stream, Torus *d_array, Torus value,
+                          Torus n);
 #endif

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

@@ -1,34 +0,0 @@
-#ifndef HELPER_MULTI_GPU_H
-#define HELPER_MULTI_GPU_H
-#include <mutex>
-#include <variant>
-#include <vector>
-
-extern std::mutex m;
-extern bool p2p_enabled;
-
-extern "C" {
-int cuda_setup_multi_gpu();
-}
-
-// Define a variant type that can be either a vector or a single pointer
-template <typename Torus>
-using LweArrayVariant = std::variant<std::vector<Torus *>, Torus *>;
-
-// Macro to define the visitor logic using std::holds_alternative for vectors
-#define GET_VARIANT_ELEMENT(variant, index)                                    \
-  [&] {                                                                        \
-    if (std::holds_alternative<std::vector<Torus *>>(variant)) {               \
-      return std::get<std::vector<Torus *>>(variant)[index];                   \
-    } else {                                                                   \
-      return std::get<Torus *>(variant);                                       \
-    }                                                                          \
-  }()
-
-int get_active_gpu_count(int num_inputs, int gpu_count);
-
-int get_num_inputs_on_gpu(int total_num_inputs, int gpu_index, int gpu_count);
-
-int get_gpu_offset(int total_num_inputs, int gpu_index, int gpu_count);
-
-#endif

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

@@ -6,16 +6,16 @@
 extern "C" {
 
 void cuda_keyswitch_lwe_ciphertext_vector_32(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void *lwe_output_indexes, void *lwe_array_in, void *lwe_input_indexes,
-    void *ksk, uint32_t lwe_dimension_in, uint32_t lwe_dimension_out,
-    uint32_t base_log, uint32_t level_count, uint32_t num_samples);
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_output_indexes,
+    void *lwe_array_in, void *lwe_input_indexes, void *ksk,
+    uint32_t lwe_dimension_in, uint32_t lwe_dimension_out, uint32_t base_log,
+    uint32_t level_count, uint32_t num_samples);
 
 void cuda_keyswitch_lwe_ciphertext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void *lwe_output_indexes, void *lwe_array_in, void *lwe_input_indexes,
-    void *ksk, uint32_t lwe_dimension_in, uint32_t lwe_dimension_out,
-    uint32_t base_log, uint32_t level_count, uint32_t num_samples);
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_output_indexes,
+    void *lwe_array_in, void *lwe_input_indexes, void *ksk,
+    uint32_t lwe_dimension_in, uint32_t lwe_dimension_out, uint32_t base_log,
+    uint32_t level_count, uint32_t num_samples);
 }
 
 #endif // CNCRT_KS_H_

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

@@ -1,48 +1,48 @@
 #ifndef CUDA_LINALG_H_
 #define CUDA_LINALG_H_
 
-#include "programmable_bootstrap.h"
+#include "bootstrap.h"
 #include <cstdint>
 #include <device.h>
 
 extern "C" {
 
-void cuda_negate_lwe_ciphertext_vector_32(void *stream, uint32_t gpu_index,
+void cuda_negate_lwe_ciphertext_vector_32(cuda_stream_t *stream,
                                           void *lwe_array_out,
                                           void *lwe_array_in,
                                           uint32_t input_lwe_dimension,
                                           uint32_t input_lwe_ciphertext_count);
-void cuda_negate_lwe_ciphertext_vector_64(void *stream, uint32_t gpu_index,
+void cuda_negate_lwe_ciphertext_vector_64(cuda_stream_t *stream,
                                           void *lwe_array_out,
                                           void *lwe_array_in,
                                           uint32_t input_lwe_dimension,
                                           uint32_t input_lwe_ciphertext_count);
-void cuda_add_lwe_ciphertext_vector_32(void *stream, uint32_t gpu_index,
+void cuda_add_lwe_ciphertext_vector_32(cuda_stream_t *stream,
                                        void *lwe_array_out,
                                        void *lwe_array_in_1,
                                        void *lwe_array_in_2,
                                        uint32_t input_lwe_dimension,
                                        uint32_t input_lwe_ciphertext_count);
-void cuda_add_lwe_ciphertext_vector_64(void *stream, uint32_t gpu_index,
+void cuda_add_lwe_ciphertext_vector_64(cuda_stream_t *stream,
                                        void *lwe_array_out,
                                        void *lwe_array_in_1,
                                        void *lwe_array_in_2,
                                        uint32_t input_lwe_dimension,
                                        uint32_t input_lwe_ciphertext_count);
 void cuda_add_lwe_ciphertext_vector_plaintext_vector_32(
-    void *stream, uint32_t gpu_index, void *lwe_array_out, void *lwe_array_in,
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_array_in,
     void *plaintext_array_in, uint32_t input_lwe_dimension,
     uint32_t input_lwe_ciphertext_count);
 void cuda_add_lwe_ciphertext_vector_plaintext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out, void *lwe_array_in,
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_array_in,
     void *plaintext_array_in, uint32_t input_lwe_dimension,
     uint32_t input_lwe_ciphertext_count);
 void cuda_mult_lwe_ciphertext_vector_cleartext_vector_32(
-    void *stream, uint32_t gpu_index, void *lwe_array_out, void *lwe_array_in,
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_array_in,
     void *cleartext_array_in, uint32_t input_lwe_dimension,
     uint32_t input_lwe_ciphertext_count);
 void cuda_mult_lwe_ciphertext_vector_cleartext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out, void *lwe_array_in,
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_array_in,
     void *cleartext_array_in, uint32_t input_lwe_dimension,
     uint32_t input_lwe_ciphertext_count);
 }

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

@@ -1,416 +0,0 @@
-#ifndef CUDA_BOOTSTRAP_H
-#define CUDA_BOOTSTRAP_H
-
-#include "device.h"
-#include <cstdint>
-
-enum PBS_TYPE { MULTI_BIT = 0, CLASSICAL = 1 };
-enum PBS_VARIANT { DEFAULT = 0, CG = 1, TBC = 2 };
-
-extern "C" {
-void cuda_fourier_polynomial_mul(cudaStream_t stream, uint32_t gpu_index,
-                                 void *input1, void *input2, void *output,
-                                 uint32_t polynomial_size,
-                                 uint32_t total_polynomials);
-
-void cuda_convert_lwe_programmable_bootstrap_key_32(
-    void *stream, uint32_t gpu_index, void *dest, void *src,
-    uint32_t input_lwe_dim, uint32_t glwe_dim, uint32_t level_count,
-    uint32_t polynomial_size);
-
-void cuda_convert_lwe_programmable_bootstrap_key_64(
-    void *stream, uint32_t gpu_index, void *dest, void *src,
-    uint32_t input_lwe_dim, uint32_t glwe_dim, uint32_t level_count,
-    uint32_t polynomial_size);
-
-void scratch_cuda_programmable_bootstrap_amortized_32(
-    void *stream, uint32_t gpu_index, int8_t **pbs_buffer,
-    uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-void scratch_cuda_programmable_bootstrap_amortized_64(
-    void *stream, uint32_t gpu_index, int8_t **pbs_buffer,
-    uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-void cuda_programmable_bootstrap_amortized_lwe_ciphertext_vector_32(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void *lwe_output_indexes, void *lut_vector, void *lut_vector_indexes,
-    void *lwe_array_in, void *lwe_input_indexes, void *bootstrapping_key,
-    int8_t *pbs_buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples);
-
-void cuda_programmable_bootstrap_amortized_lwe_ciphertext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void *lwe_output_indexes, void *lut_vector, void *lut_vector_indexes,
-    void *lwe_array_in, void *lwe_input_indexes, void *bootstrapping_key,
-    int8_t *pbs_buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples);
-
-void cleanup_cuda_programmable_bootstrap_amortized(void *stream,
-                                                   uint32_t gpu_index,
-                                                   int8_t **pbs_buffer);
-
-void scratch_cuda_programmable_bootstrap_32(
-    void *stream, uint32_t gpu_index, int8_t **buffer, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t level_count,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-void scratch_cuda_programmable_bootstrap_64(
-    void *stream, uint32_t gpu_index, int8_t **buffer, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t level_count,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-void cuda_programmable_bootstrap_lwe_ciphertext_vector_32(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void *lwe_output_indexes, void *lut_vector, void *lut_vector_indexes,
-    void *lwe_array_in, void *lwe_input_indexes, void *bootstrapping_key,
-    int8_t *buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples);
-
-void cuda_programmable_bootstrap_lwe_ciphertext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void *lwe_output_indexes, void *lut_vector, void *lut_vector_indexes,
-    void *lwe_array_in, void *lwe_input_indexes, void *bootstrapping_key,
-    int8_t *buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples);
-
-void cleanup_cuda_programmable_bootstrap(void *stream, uint32_t gpu_index,
-                                         int8_t **pbs_buffer);
-
-uint64_t get_buffer_size_programmable_bootstrap_amortized_64(
-    uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t input_lwe_ciphertext_count);
-
-uint64_t get_buffer_size_programmable_bootstrap_64(
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
-    uint32_t input_lwe_ciphertext_count);
-}
-
-template <typename Torus>
-uint64_t get_buffer_size_full_sm_programmable_bootstrap_step_one(
-    uint32_t polynomial_size) {
-  return sizeof(Torus) * polynomial_size +      // accumulator_rotated
-         sizeof(double2) * polynomial_size / 2; // accumulator fft
-}
-template <typename Torus>
-uint64_t get_buffer_size_full_sm_programmable_bootstrap_step_two(
-    uint32_t polynomial_size) {
-  return sizeof(Torus) * polynomial_size +      // accumulator
-         sizeof(double2) * polynomial_size / 2; // accumulator fft
-}
-
-template <typename Torus>
-uint64_t
-get_buffer_size_partial_sm_programmable_bootstrap(uint32_t polynomial_size) {
-  return sizeof(double2) * polynomial_size / 2; // accumulator fft
-}
-
-template <typename Torus>
-uint64_t
-get_buffer_size_full_sm_programmable_bootstrap_tbc(uint32_t polynomial_size) {
-  return sizeof(Torus) * polynomial_size +      // accumulator_rotated
-         sizeof(Torus) * polynomial_size +      // accumulator
-         sizeof(double2) * polynomial_size / 2; // accumulator fft
-}
-
-template <typename Torus>
-uint64_t get_buffer_size_partial_sm_programmable_bootstrap_tbc(
-    uint32_t polynomial_size) {
-  return sizeof(double2) * polynomial_size / 2; // accumulator fft mask & body
-}
-
-template <typename Torus>
-uint64_t get_buffer_size_sm_dsm_plus_tbc_classic_programmable_bootstrap(
-    uint32_t polynomial_size) {
-  return sizeof(double2) * polynomial_size / 2; // tbc
-}
-
-template <typename Torus>
-uint64_t
-get_buffer_size_full_sm_programmable_bootstrap_cg(uint32_t polynomial_size) {
-  return sizeof(Torus) * polynomial_size +      // accumulator_rotated
-         sizeof(Torus) * polynomial_size +      // accumulator
-         sizeof(double2) * polynomial_size / 2; // accumulator fft
-}
-
-template <typename Torus>
-uint64_t
-get_buffer_size_partial_sm_programmable_bootstrap_cg(uint32_t polynomial_size) {
-  return sizeof(double2) * polynomial_size / 2; // accumulator fft mask & body
-}
-
-template <typename Torus>
-bool supports_distributed_shared_memory_on_classic_programmable_bootstrap(
-    uint32_t polynomial_size);
-
-template <typename Torus, PBS_TYPE pbs_type> struct pbs_buffer;
-
-template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
-  int8_t *d_mem;
-
-  Torus *global_accumulator;
-  double2 *global_accumulator_fft;
-
-  PBS_VARIANT pbs_variant;
-
-  pbs_buffer(cudaStream_t stream, uint32_t gpu_index, uint32_t glwe_dimension,
-             uint32_t polynomial_size, uint32_t level_count,
-             uint32_t input_lwe_ciphertext_count, PBS_VARIANT pbs_variant,
-             bool allocate_gpu_memory) {
-
-    this->pbs_variant = pbs_variant;
-
-    auto max_shared_memory = cuda_get_max_shared_memory(0);
-
-    if (allocate_gpu_memory) {
-      switch (pbs_variant) {
-      case PBS_VARIANT::DEFAULT: {
-        uint64_t full_sm_step_one =
-            get_buffer_size_full_sm_programmable_bootstrap_step_one<Torus>(
-                polynomial_size);
-        uint64_t full_sm_step_two =
-            get_buffer_size_full_sm_programmable_bootstrap_step_two<Torus>(
-                polynomial_size);
-        uint64_t partial_sm =
-            get_buffer_size_partial_sm_programmable_bootstrap<Torus>(
-                polynomial_size);
-
-        uint64_t partial_dm_step_one = full_sm_step_one - partial_sm;
-        uint64_t partial_dm_step_two = full_sm_step_two - partial_sm;
-        uint64_t full_dm = full_sm_step_one;
-
-        uint64_t device_mem = 0;
-        if (max_shared_memory < partial_sm) {
-          device_mem = full_dm * input_lwe_ciphertext_count * level_count *
-                       (glwe_dimension + 1);
-        } else if (max_shared_memory < full_sm_step_two) {
-          device_mem =
-              (partial_dm_step_two + partial_dm_step_one * level_count) *
-              input_lwe_ciphertext_count * (glwe_dimension + 1);
-        } else if (max_shared_memory < full_sm_step_one) {
-          device_mem = partial_dm_step_one * input_lwe_ciphertext_count *
-                       level_count * (glwe_dimension + 1);
-        }
-        // Otherwise, both kernels run all in shared memory
-        d_mem = (int8_t *)cuda_malloc_async(device_mem, stream, gpu_index);
-
-        global_accumulator_fft = (double2 *)cuda_malloc_async(
-            (glwe_dimension + 1) * level_count * input_lwe_ciphertext_count *
-                (polynomial_size / 2) * sizeof(double2),
-            stream, gpu_index);
-
-        global_accumulator = (Torus *)cuda_malloc_async(
-            (glwe_dimension + 1) * input_lwe_ciphertext_count *
-                polynomial_size * sizeof(Torus),
-            stream, gpu_index);
-      } break;
-      case PBS_VARIANT::CG: {
-        uint64_t full_sm =
-            get_buffer_size_full_sm_programmable_bootstrap_cg<Torus>(
-                polynomial_size);
-        uint64_t partial_sm =
-            get_buffer_size_partial_sm_programmable_bootstrap_cg<Torus>(
-                polynomial_size);
-
-        uint64_t partial_dm = full_sm - partial_sm;
-        uint64_t full_dm = full_sm;
-        uint64_t device_mem = 0;
-
-        if (max_shared_memory < partial_sm) {
-          device_mem = full_dm * input_lwe_ciphertext_count * level_count *
-                       (glwe_dimension + 1);
-        } else if (max_shared_memory < full_sm) {
-          device_mem = partial_dm * input_lwe_ciphertext_count * level_count *
-                       (glwe_dimension + 1);
-        }
-
-        // Otherwise, both kernels run all in shared memory
-        d_mem = (int8_t *)cuda_malloc_async(device_mem, stream, gpu_index);
-
-        global_accumulator_fft = (double2 *)cuda_malloc_async(
-            (glwe_dimension + 1) * level_count * input_lwe_ciphertext_count *
-                polynomial_size / 2 * sizeof(double2),
-            stream, gpu_index);
-      } break;
-#if CUDA_ARCH >= 900
-      case PBS_VARIANT::TBC: {
-
-        bool supports_dsm =
-            supports_distributed_shared_memory_on_classic_programmable_bootstrap<
-                Torus>(polynomial_size);
-
-        uint64_t full_sm =
-            get_buffer_size_full_sm_programmable_bootstrap_tbc<Torus>(
-                polynomial_size);
-        uint64_t partial_sm =
-            get_buffer_size_partial_sm_programmable_bootstrap_tbc<Torus>(
-                polynomial_size);
-        uint64_t minimum_sm_tbc = 0;
-        if (supports_dsm)
-          minimum_sm_tbc =
-              get_buffer_size_sm_dsm_plus_tbc_classic_programmable_bootstrap<
-                  Torus>(polynomial_size);
-
-        uint64_t partial_dm = full_sm - partial_sm;
-        uint64_t full_dm = full_sm;
-        uint64_t device_mem = 0;
-
-        // There is a minimum amount of memory we need to run the TBC PBS, which
-        // is minimum_sm_tbc. We know that minimum_sm_tbc bytes are available
-        // because otherwise the previous check would have redirected
-        // computation to some other variant. If over that we don't have more
-        // partial_sm bytes, TBC PBS will run on NOSM. If we have partial_sm but
-        // not full_sm bytes, it will run on PARTIALSM. Otherwise, FULLSM.
-        //
-        // NOSM mode actually requires minimum_sm_tbc shared memory bytes.
-        if (max_shared_memory < partial_sm + minimum_sm_tbc) {
-          device_mem = full_dm * input_lwe_ciphertext_count * level_count *
-                       (glwe_dimension + 1);
-        } else if (max_shared_memory < full_sm + minimum_sm_tbc) {
-          device_mem = partial_dm * input_lwe_ciphertext_count * level_count *
-                       (glwe_dimension + 1);
-        }
-
-        // Otherwise, both kernels run all in shared memory
-        d_mem = (int8_t *)cuda_malloc_async(device_mem, stream, gpu_index);
-
-        global_accumulator_fft = (double2 *)cuda_malloc_async(
-            (glwe_dimension + 1) * level_count * input_lwe_ciphertext_count *
-                polynomial_size / 2 * sizeof(double2),
-            stream, gpu_index);
-      } break;
-#endif
-      default:
-        PANIC("Cuda error (PBS): unsupported implementation variant.")
-      }
-    }
-  }
-
-  void release(cudaStream_t stream, uint32_t gpu_index) {
-    cuda_drop_async(d_mem, stream, gpu_index);
-    cuda_drop_async(global_accumulator_fft, stream, gpu_index);
-
-    if (pbs_variant == DEFAULT)
-      cuda_drop_async(global_accumulator, stream, gpu_index);
-  }
-};
-
-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);
-  uint64_t full_sm =
-      get_buffer_size_full_sm_programmable_bootstrap_cg<Torus>(polynomial_size);
-  uint64_t partial_sm =
-      get_buffer_size_partial_sm_programmable_bootstrap_cg<Torus>(
-          polynomial_size);
-  uint64_t partial_dm = full_sm - partial_sm;
-  uint64_t full_dm = full_sm;
-  uint64_t device_mem = 0;
-  if (max_shared_memory < partial_sm) {
-    device_mem = full_dm * input_lwe_ciphertext_count * level_count *
-                 (glwe_dimension + 1);
-  } else if (max_shared_memory < full_sm) {
-    device_mem = partial_dm * input_lwe_ciphertext_count * level_count *
-                 (glwe_dimension + 1);
-  }
-  uint64_t buffer_size = device_mem + (glwe_dimension + 1) * level_count *
-                                          input_lwe_ciphertext_count *
-                                          polynomial_size / 2 * sizeof(double2);
-  return buffer_size + buffer_size % sizeof(double2);
-}
-
-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);
-
-template <typename Torus>
-void cuda_programmable_bootstrap_cg_lwe_ciphertext_vector(
-    void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
-    Torus *lwe_array_in, Torus *lwe_input_indexes, double2 *bootstrapping_key,
-    pbs_buffer<Torus, CLASSICAL> *buffer, uint32_t lwe_dimension,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t base_log,
-    uint32_t level_count, uint32_t num_samples);
-
-template <typename Torus>
-void cuda_programmable_bootstrap_lwe_ciphertext_vector(
-    void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
-    Torus *lwe_array_in, Torus *lwe_input_indexes, double2 *bootstrapping_key,
-    pbs_buffer<Torus, CLASSICAL> *buffer, uint32_t lwe_dimension,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t base_log,
-    uint32_t level_count, uint32_t num_samples);
-
-#if (CUDA_ARCH >= 900)
-template <typename Torus>
-void cuda_programmable_bootstrap_tbc_lwe_ciphertext_vector(
-    void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
-    Torus *lwe_array_in, Torus *lwe_input_indexes, double2 *bootstrapping_key,
-    pbs_buffer<Torus, CLASSICAL> *buffer, uint32_t lwe_dimension,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t base_log,
-    uint32_t level_count, uint32_t num_samples);
-
-template <typename Torus>
-void scratch_cuda_programmable_bootstrap_tbc(
-    void *stream, uint32_t gpu_index, pbs_buffer<Torus, CLASSICAL> **pbs_buffer,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-#endif
-
-template <typename Torus>
-void scratch_cuda_programmable_bootstrap_cg(
-    void *stream, uint32_t gpu_index, pbs_buffer<Torus, CLASSICAL> **pbs_buffer,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-template <typename Torus>
-void scratch_cuda_programmable_bootstrap(
-    void *stream, uint32_t gpu_index, pbs_buffer<Torus, CLASSICAL> **buffer,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-template <typename Torus>
-bool has_support_to_cuda_programmable_bootstrap_tbc(uint32_t num_samples,
-                                                    uint32_t glwe_dimension,
-                                                    uint32_t polynomial_size,
-                                                    uint32_t level_count);
-
-#ifdef __CUDACC__
-__device__ inline int get_start_ith_ggsw(int i, uint32_t polynomial_size,
-                                         int glwe_dimension,
-                                         uint32_t level_count);
-
-template <typename T>
-__device__ const T *get_ith_mask_kth_block(const T *ptr, int i, int k,
-                                           int level, uint32_t polynomial_size,
-                                           int glwe_dimension,
-                                           uint32_t level_count);
-
-template <typename T>
-__device__ T *get_ith_mask_kth_block(T *ptr, int i, int k, int level,
-                                     uint32_t polynomial_size,
-                                     int glwe_dimension, uint32_t level_count);
-
-template <typename T>
-__device__ T *get_ith_body_kth_block(T *ptr, int i, int k, int level,
-                                     uint32_t polynomial_size,
-                                     int glwe_dimension, uint32_t level_count);
-
-template <typename T>
-__device__ const T *get_multi_bit_ith_lwe_gth_group_kth_block(
-    const T *ptr, int g, int i, int k, int level, uint32_t grouping_factor,
-    uint32_t polynomial_size, uint32_t glwe_dimension, uint32_t level_count);
-
-#endif
-
-#endif // CUDA_BOOTSTRAP_H

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

@@ -1,293 +0,0 @@
-#ifndef CUDA_MULTI_BIT_H
-#define CUDA_MULTI_BIT_H
-
-#include "programmable_bootstrap.h"
-#include <cstdint>
-
-extern "C" {
-
-bool has_support_to_cuda_programmable_bootstrap_cg_multi_bit(
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
-    uint32_t num_samples);
-
-void cuda_convert_lwe_multi_bit_programmable_bootstrap_key_64(
-    void *stream, uint32_t gpu_index, void *dest, void *src,
-    uint32_t input_lwe_dim, uint32_t glwe_dim, uint32_t level_count,
-    uint32_t polynomial_size, uint32_t grouping_factor);
-
-void scratch_cuda_multi_bit_programmable_bootstrap_64(
-    void *stream, uint32_t gpu_index, int8_t **pbs_buffer,
-    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t level_count, uint32_t grouping_factor,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-void cuda_multi_bit_programmable_bootstrap_lwe_ciphertext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void *lwe_output_indexes, void *lut_vector, void *lut_vector_indexes,
-    void *lwe_array_in, void *lwe_input_indexes, void *bootstrapping_key,
-    int8_t *buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t grouping_factor, uint32_t base_log,
-    uint32_t level_count, uint32_t num_samples);
-
-void cleanup_cuda_multi_bit_programmable_bootstrap(void *stream,
-                                                   uint32_t gpu_index,
-                                                   int8_t **pbs_buffer);
-}
-
-template <typename Torus>
-bool supports_distributed_shared_memory_on_multibit_programmable_bootstrap(
-    uint32_t polynomial_size);
-
-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);
-
-#if CUDA_ARCH >= 900
-template <typename Torus>
-void scratch_cuda_tbc_multi_bit_programmable_bootstrap(
-    void *stream, uint32_t gpu_index, pbs_buffer<Torus, MULTI_BIT> **buffer,
-    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t level_count, uint32_t grouping_factor,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-template <typename Torus>
-void cuda_tbc_multi_bit_programmable_bootstrap_lwe_ciphertext_vector(
-    void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
-    Torus *lwe_array_in, Torus *lwe_input_indexes, Torus *bootstrapping_key,
-    pbs_buffer<Torus, MULTI_BIT> *pbs_buffer, uint32_t lwe_dimension,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t grouping_factor,
-    uint32_t base_log, uint32_t level_count, uint32_t num_samples);
-#endif
-
-template <typename Torus>
-void scratch_cuda_cg_multi_bit_programmable_bootstrap(
-    void *stream, uint32_t gpu_index, pbs_buffer<Torus, MULTI_BIT> **pbs_buffer,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-template <typename Torus>
-void cuda_cg_multi_bit_programmable_bootstrap_lwe_ciphertext_vector(
-    void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
-    Torus *lwe_array_in, Torus *lwe_input_indexes, Torus *bootstrapping_key,
-    pbs_buffer<Torus, MULTI_BIT> *pbs_buffer, uint32_t lwe_dimension,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t grouping_factor,
-    uint32_t base_log, uint32_t level_count, uint32_t num_samples);
-
-template <typename Torus>
-void scratch_cuda_multi_bit_programmable_bootstrap(
-    void *stream, uint32_t gpu_index, pbs_buffer<Torus, MULTI_BIT> **pbs_buffer,
-    uint32_t lwe_dimension, uint32_t glwe_dimension, uint32_t polynomial_size,
-    uint32_t level_count, uint32_t grouping_factor,
-    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
-
-template <typename Torus>
-void cuda_multi_bit_programmable_bootstrap_lwe_ciphertext_vector(
-    void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
-    Torus *lwe_array_in, Torus *lwe_input_indexes, Torus *bootstrapping_key,
-    pbs_buffer<Torus, MULTI_BIT> *pbs_buffer, uint32_t lwe_dimension,
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t grouping_factor,
-    uint32_t base_log, uint32_t level_count, uint32_t num_samples);
-
-template <typename Torus>
-uint64_t get_buffer_size_full_sm_multibit_programmable_bootstrap_keybundle(
-    uint32_t polynomial_size);
-template <typename Torus>
-uint64_t get_buffer_size_full_sm_multibit_programmable_bootstrap_step_one(
-    uint32_t polynomial_size);
-template <typename Torus>
-uint64_t get_buffer_size_full_sm_multibit_programmable_bootstrap_step_two(
-    uint32_t polynomial_size);
-template <typename Torus>
-uint64_t get_buffer_size_partial_sm_multibit_programmable_bootstrap_step_one(
-    uint32_t polynomial_size);
-template <typename Torus>
-uint64_t get_buffer_size_full_sm_cg_multibit_programmable_bootstrap(
-    uint32_t polynomial_size);
-template <typename Torus>
-uint64_t get_buffer_size_partial_sm_cg_multibit_programmable_bootstrap(
-    uint32_t polynomial_size);
-template <typename Torus>
-uint64_t get_buffer_size_sm_dsm_plus_tbc_multibit_programmable_bootstrap(
-    uint32_t polynomial_size);
-template <typename Torus>
-uint64_t get_buffer_size_partial_sm_tbc_multibit_programmable_bootstrap(
-    uint32_t polynomial_size);
-template <typename Torus>
-uint64_t get_buffer_size_full_sm_tbc_multibit_programmable_bootstrap(
-    uint32_t polynomial_size);
-
-template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::MULTI_BIT> {
-  int8_t *d_mem_keybundle = NULL;
-  int8_t *d_mem_acc_step_one = NULL;
-  int8_t *d_mem_acc_step_two = NULL;
-  int8_t *d_mem_acc_cg = NULL;
-  int8_t *d_mem_acc_tbc = NULL;
-
-  double2 *keybundle_fft;
-  Torus *global_accumulator;
-  double2 *global_accumulator_fft;
-
-  PBS_VARIANT pbs_variant;
-
-  pbs_buffer(cudaStream_t stream, uint32_t gpu_index, uint32_t glwe_dimension,
-             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) {
-    this->pbs_variant = pbs_variant;
-    auto max_shared_memory = cuda_get_max_shared_memory(gpu_index);
-
-    // default
-    uint64_t full_sm_keybundle =
-        get_buffer_size_full_sm_multibit_programmable_bootstrap_keybundle<
-            Torus>(polynomial_size);
-    uint64_t full_sm_accumulate_step_one =
-        get_buffer_size_full_sm_multibit_programmable_bootstrap_step_one<Torus>(
-            polynomial_size);
-    uint64_t full_sm_accumulate_step_two =
-        get_buffer_size_full_sm_multibit_programmable_bootstrap_step_two<Torus>(
-            polynomial_size);
-    uint64_t partial_sm_accumulate_step_one =
-        get_buffer_size_partial_sm_multibit_programmable_bootstrap_step_one<
-            Torus>(polynomial_size);
-    // cg
-    uint64_t full_sm_cg_accumulate =
-        get_buffer_size_full_sm_cg_multibit_programmable_bootstrap<Torus>(
-            polynomial_size);
-    uint64_t partial_sm_cg_accumulate =
-        get_buffer_size_partial_sm_cg_multibit_programmable_bootstrap<Torus>(
-            polynomial_size);
-
-    auto num_blocks_keybundle = input_lwe_ciphertext_count * lwe_chunk_size *
-                                (glwe_dimension + 1) * (glwe_dimension + 1) *
-                                level_count;
-    auto num_blocks_acc_step_one =
-        level_count * (glwe_dimension + 1) * input_lwe_ciphertext_count;
-    auto num_blocks_acc_step_two =
-        input_lwe_ciphertext_count * (glwe_dimension + 1);
-    auto num_blocks_acc_cg =
-        level_count * (glwe_dimension + 1) * input_lwe_ciphertext_count;
-
-#if CUDA_ARCH >= 900
-    uint64_t full_sm_tbc_accumulate =
-        get_buffer_size_full_sm_tbc_multibit_programmable_bootstrap<Torus>(
-            polynomial_size);
-    uint64_t partial_sm_tbc_accumulate =
-        get_buffer_size_partial_sm_tbc_multibit_programmable_bootstrap<Torus>(
-            polynomial_size);
-    uint64_t minimum_sm_tbc =
-        get_buffer_size_sm_dsm_plus_tbc_multibit_programmable_bootstrap<Torus>(
-            polynomial_size);
-    auto num_blocks_acc_tbc = num_blocks_acc_cg;
-#endif
-
-    if (allocate_gpu_memory) {
-      // Keybundle
-      if (max_shared_memory < full_sm_keybundle)
-        d_mem_keybundle = (int8_t *)cuda_malloc_async(
-            num_blocks_keybundle * full_sm_keybundle, stream, gpu_index);
-
-      switch (pbs_variant) {
-      case PBS_VARIANT::CG:
-        // Accumulator CG
-        if (max_shared_memory < partial_sm_cg_accumulate)
-          d_mem_acc_cg = (int8_t *)cuda_malloc_async(
-              num_blocks_acc_cg * full_sm_cg_accumulate, stream, gpu_index);
-        else if (max_shared_memory < full_sm_cg_accumulate)
-          d_mem_acc_cg = (int8_t *)cuda_malloc_async(
-              num_blocks_acc_cg * partial_sm_cg_accumulate, stream, gpu_index);
-        break;
-      case PBS_VARIANT::DEFAULT:
-        // Accumulator step one
-        if (max_shared_memory < partial_sm_accumulate_step_one)
-          d_mem_acc_step_one = (int8_t *)cuda_malloc_async(
-              num_blocks_acc_step_one * full_sm_accumulate_step_one, stream,
-              gpu_index);
-        else if (max_shared_memory < full_sm_accumulate_step_one)
-          d_mem_acc_step_one = (int8_t *)cuda_malloc_async(
-              num_blocks_acc_step_one * partial_sm_accumulate_step_one, stream,
-              gpu_index);
-
-        // Accumulator step two
-        if (max_shared_memory < full_sm_accumulate_step_two)
-          d_mem_acc_step_two = (int8_t *)cuda_malloc_async(
-              num_blocks_acc_step_two * full_sm_accumulate_step_two, stream,
-              gpu_index);
-        break;
-#if CUDA_ARCH >= 900
-      case TBC:
-        // There is a minimum amount of memory we need to run the TBC PBS, which
-        // is minimum_sm_tbc. We know that minimum_sm_tbc bytes are available
-        // because otherwise the previous check would have redirected
-        // computation to some other variant. If over that we don't have more
-        // partial_sm_tbc_accumulate bytes, TBC PBS will run on NOSM. If we have
-        // partial_sm_tbc_accumulate but not full_sm_tbc_accumulate bytes, it
-        // will run on PARTIALSM. Otherwise, FULLSM.
-        //
-        // NOSM mode actually requires minimum_sm_tbc shared memory bytes.
-
-        // Accumulator TBC
-        if (max_shared_memory < partial_sm_tbc_accumulate + minimum_sm_tbc)
-          d_mem_acc_tbc = (int8_t *)cuda_malloc_async(
-              num_blocks_acc_tbc * full_sm_tbc_accumulate, stream, gpu_index);
-        else if (max_shared_memory < full_sm_tbc_accumulate + minimum_sm_tbc)
-          d_mem_acc_tbc = (int8_t *)cuda_malloc_async(
-              num_blocks_acc_tbc * partial_sm_tbc_accumulate, stream,
-              gpu_index);
-        break;
-#endif
-      default:
-        PANIC("Cuda error (PBS): unsupported implementation variant.")
-      }
-
-      keybundle_fft = (double2 *)cuda_malloc_async(
-          num_blocks_keybundle * (polynomial_size / 2) * sizeof(double2),
-          stream, gpu_index);
-      global_accumulator = (Torus *)cuda_malloc_async(
-          num_blocks_acc_step_one * polynomial_size * sizeof(Torus), stream,
-          gpu_index);
-      global_accumulator_fft = (double2 *)cuda_malloc_async(
-          num_blocks_acc_step_one * (polynomial_size / 2) * sizeof(double2),
-          stream, gpu_index);
-    }
-  }
-
-  void release(cudaStream_t stream, uint32_t gpu_index) {
-
-    if (d_mem_keybundle)
-      cuda_drop_async(d_mem_keybundle, stream, gpu_index);
-    switch (pbs_variant) {
-    case DEFAULT:
-      if (d_mem_acc_step_one)
-        cuda_drop_async(d_mem_acc_step_one, stream, gpu_index);
-      if (d_mem_acc_step_two)
-        cuda_drop_async(d_mem_acc_step_two, stream, gpu_index);
-      break;
-    case CG:
-      if (d_mem_acc_cg)
-        cuda_drop_async(d_mem_acc_cg, stream, gpu_index);
-      break;
-#if CUDA_ARCH >= 900
-    case TBC:
-      if (d_mem_acc_tbc)
-        cuda_drop_async(d_mem_acc_tbc, stream, gpu_index);
-      break;
-#endif
-    default:
-      PANIC("Cuda error (PBS): unsupported implementation variant.")
-    }
-
-    cuda_drop_async(keybundle_fft, stream, gpu_index);
-    cuda_drop_async(global_accumulator, stream, gpu_index);
-    cuda_drop_async(global_accumulator_fft, stream, gpu_index);
-  }
-};
-
-template <typename Torus, class params>
-uint32_t get_lwe_chunk_size(uint32_t gpu_index, uint32_t max_num_pbs,
-                            uint32_t polynomial_size);
-
-#endif // CUDA_MULTI_BIT_H

backends/tfhe-cuda-backend/cuda/src/CMakeLists.txt

@@ -10,8 +10,7 @@ set(SOURCES
     ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/keyswitch.h
     ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/linear_algebra.h
     ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/shifts.h
-    ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/vertical_packing.h
-    ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/helper_multi_gpu.h)
+    ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/vertical_packing.h)
 file(GLOB_RECURSE SOURCES "*.cu")
 add_library(tfhe_cuda_backend STATIC ${SOURCES})
 set_target_properties(tfhe_cuda_backend PROPERTIES CUDA_SEPARABLE_COMPILATION ON CUDA_RESOLVE_DEVICE_SYMBOLS ON)

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

@@ -1,77 +1 @@
 #include "ciphertext.cuh"
-#include "polynomial/parameters.cuh"
-
-void cuda_convert_lwe_ciphertext_vector_to_gpu_64(void *stream,
-                                                  uint32_t gpu_index,
-                                                  void *dest, void *src,
-                                                  uint32_t number_of_cts,
-                                                  uint32_t lwe_dimension) {
-  cuda_convert_lwe_ciphertext_vector_to_gpu<uint64_t>(
-      static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)dest,
-      (uint64_t *)src, number_of_cts, lwe_dimension);
-}
-
-void cuda_convert_lwe_ciphertext_vector_to_cpu_64(void *stream,
-                                                  uint32_t gpu_index,
-                                                  void *dest, void *src,
-                                                  uint32_t number_of_cts,
-                                                  uint32_t lwe_dimension) {
-  cuda_convert_lwe_ciphertext_vector_to_cpu<uint64_t>(
-      static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)dest,
-      (uint64_t *)src, number_of_cts, lwe_dimension);
-}
-
-void cuda_glwe_sample_extract_64(void *stream, uint32_t gpu_index,
-                                 void *lwe_array_out, void *glwe_array_in,
-                                 uint32_t *nth_array, uint32_t num_glwes,
-                                 uint32_t glwe_dimension,
-                                 uint32_t polynomial_size) {
-
-  switch (polynomial_size) {
-  case 256:
-    host_sample_extract<uint64_t, AmortizedDegree<256>>(
-        static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_glwes,
-        glwe_dimension);
-    break;
-  case 512:
-    host_sample_extract<uint64_t, AmortizedDegree<512>>(
-        static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_glwes,
-        glwe_dimension);
-    break;
-  case 1024:
-    host_sample_extract<uint64_t, AmortizedDegree<1024>>(
-        static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_glwes,
-        glwe_dimension);
-    break;
-  case 2048:
-    host_sample_extract<uint64_t, AmortizedDegree<2048>>(
-        static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_glwes,
-        glwe_dimension);
-    break;
-  case 4096:
-    host_sample_extract<uint64_t, AmortizedDegree<4096>>(
-        static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_glwes,
-        glwe_dimension);
-    break;
-  case 8192:
-    host_sample_extract<uint64_t, AmortizedDegree<8192>>(
-        static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_glwes,
-        glwe_dimension);
-    break;
-  case 16384:
-    host_sample_extract<uint64_t, AmortizedDegree<16384>>(
-        static_cast<cudaStream_t>(stream), gpu_index, (uint64_t *)lwe_array_out,
-        (uint64_t *)glwe_array_in, (uint32_t *)nth_array, num_glwes,
-        glwe_dimension);
-    break;
-  default:
-    PANIC("Cuda error: unsupported polynomial size. Supported "
-          "N's are powers of two in the interval [256..16384].")
-  }
-}

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

@@ -3,62 +3,42 @@
 
 #include "ciphertext.h"
 #include "device.h"
-#include "polynomial/functions.cuh"
 #include <cstdint>
 
 template <typename T>
-void cuda_convert_lwe_ciphertext_vector_to_gpu(cudaStream_t stream,
-                                               uint32_t gpu_index, T *dest,
-                                               T *src, uint32_t number_of_cts,
+void cuda_convert_lwe_ciphertext_vector_to_gpu(T *dest, T *src,
+                                               cuda_stream_t *stream,
+                                               uint32_t number_of_cts,
                                                uint32_t lwe_dimension) {
-  cudaSetDevice(gpu_index);
+  cudaSetDevice(stream->gpu_index);
   uint64_t size = number_of_cts * (lwe_dimension + 1) * sizeof(T);
-  cuda_memcpy_async_to_gpu(dest, src, size, stream, gpu_index);
+  cuda_memcpy_async_to_gpu(dest, src, size, stream);
+}
+
+void cuda_convert_lwe_ciphertext_vector_to_gpu_64(void *dest, void *src,
+                                                  cuda_stream_t *stream,
+                                                  uint32_t number_of_cts,
+                                                  uint32_t lwe_dimension) {
+  cuda_convert_lwe_ciphertext_vector_to_gpu<uint64_t>(
+      (uint64_t *)dest, (uint64_t *)src, stream, number_of_cts, lwe_dimension);
 }
 
 template <typename T>
-void cuda_convert_lwe_ciphertext_vector_to_cpu(cudaStream_t stream,
-                                               uint32_t gpu_index, T *dest,
-                                               T *src, uint32_t number_of_cts,
+void cuda_convert_lwe_ciphertext_vector_to_cpu(T *dest, T *src,
+                                               cuda_stream_t *stream,
+                                               uint32_t number_of_cts,
                                                uint32_t lwe_dimension) {
-  cudaSetDevice(gpu_index);
+  cudaSetDevice(stream->gpu_index);
   uint64_t size = number_of_cts * (lwe_dimension + 1) * sizeof(T);
-  cuda_memcpy_async_to_cpu(dest, src, size, stream, gpu_index);
+  cuda_memcpy_async_to_cpu(dest, src, size, stream);
 }
 
-template <typename Torus, class params>
-__global__ void sample_extract(Torus *lwe_array_out, Torus *glwe_array_in,
-                               uint32_t *nth_array, uint32_t glwe_dimension) {
-
-  const int input_id = blockIdx.x;
-
-  const int glwe_input_size = (glwe_dimension + 1) * params::degree;
-  const int lwe_output_size = glwe_dimension * params::degree + 1;
-
-  auto lwe_out = lwe_array_out + input_id * lwe_output_size;
-
-  // We assume each GLWE will store the first polynomial_size inputs
-  uint32_t nth_per_glwe = params::degree;
-  auto glwe_in = glwe_array_in + (input_id / nth_per_glwe) * glwe_input_size;
-
-  auto nth = nth_array[input_id];
-
-  sample_extract_mask<Torus, params>(lwe_out, glwe_in, glwe_dimension, nth);
-  sample_extract_body<Torus, params>(lwe_out, glwe_in, glwe_dimension, nth);
-}
-
-template <typename Torus, class params>
-__host__ void host_sample_extract(cudaStream_t stream, uint32_t gpu_index,
-                                  Torus *lwe_array_out, Torus *glwe_array_in,
-                                  uint32_t *nth_array, uint32_t num_glwes,
-                                  uint32_t glwe_dimension) {
-  cudaSetDevice(gpu_index);
-
-  dim3 grid(num_glwes);
-  dim3 thds(params::degree / params::opt);
-  sample_extract<Torus, params><<<grid, thds, 0, stream>>>(
-      lwe_array_out, glwe_array_in, nth_array, glwe_dimension);
-  check_cuda_error(cudaGetLastError());
+void cuda_convert_lwe_ciphertext_vector_to_cpu_64(void *dest, void *src,
+                                                  cuda_stream_t *stream,
+                                                  uint32_t number_of_cts,
+                                                  uint32_t lwe_dimension) {
+  cuda_convert_lwe_ciphertext_vector_to_cpu<uint64_t>(
+      (uint64_t *)dest, (uint64_t *)src, stream, number_of_cts, lwe_dimension);
 }
 
 #endif

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

@@ -20,7 +20,9 @@ private:
   uint32_t level_count;
   uint32_t base_log;
   uint32_t mask;
+  uint32_t halfbg;
   uint32_t num_poly;
+  T offset;
   int current_level;
   T mask_mod_b;
   T *state;
@@ -80,12 +82,72 @@ public:
     synchronize_threads_in_block();
   }
 
+  // Decomposes a single polynomial
+  __device__ void
+  decompose_and_compress_next_polynomial_elements(double2 *result, int j) {
+    if (j == 0)
+      current_level -= 1;
+
+    int tid = threadIdx.x;
+    auto state_slice = state + j * params::degree;
+    for (int i = 0; i < params::opt / 2; i++) {
+      T res_re = state_slice[tid] & mask_mod_b;
+      T res_im = state_slice[tid + params::degree / 2] & mask_mod_b;
+      state_slice[tid] >>= base_log;
+      state_slice[tid + params::degree / 2] >>= base_log;
+      T carry_re = ((res_re - 1ll) | state_slice[tid]) & res_re;
+      T carry_im =
+          ((res_im - 1ll) | state_slice[tid + params::degree / 2]) & res_im;
+      carry_re >>= (base_log - 1);
+      carry_im >>= (base_log - 1);
+      state_slice[tid] += carry_re;
+      state_slice[tid + params::degree / 2] += carry_im;
+      res_re -= carry_re << base_log;
+      res_im -= carry_im << base_log;
+
+      result[i].x = (int32_t)res_re;
+      result[i].y = (int32_t)res_im;
+
+      tid += params::degree / params::opt;
+    }
+    synchronize_threads_in_block();
+  }
+
   __device__ void decompose_and_compress_level(double2 *result, int level) {
     for (int i = 0; i < level_count - level; i++)
       decompose_and_compress_next(result);
   }
 };
 
+template <typename T> class GadgetMatrixSingle {
+private:
+  uint32_t level_count;
+  uint32_t base_log;
+  uint32_t mask;
+  uint32_t halfbg;
+  T offset;
+
+public:
+  __device__ GadgetMatrixSingle(uint32_t base_log, uint32_t level_count)
+      : base_log(base_log), level_count(level_count) {
+    uint32_t bg = 1 << base_log;
+    this->halfbg = bg / 2;
+    this->mask = bg - 1;
+    T temp = 0;
+    for (int i = 0; i < this->level_count; i++) {
+      temp += 1ULL << (sizeof(T) * 8 - (i + 1) * this->base_log);
+    }
+    this->offset = temp * this->halfbg;
+  }
+
+  __device__ T decompose_one_level_single(T element, uint32_t level) {
+    T s = element + this->offset;
+    uint32_t decal = (sizeof(T) * 8 - (level + 1) * this->base_log);
+    T temp1 = (s >> decal) & this->mask;
+    return (T)(temp1 - this->halfbg);
+  }
+};
+
 template <typename Torus>
 __device__ Torus decompose_one(Torus &state, Torus mask_mod_b, int base_log) {
   Torus res = state & mask_mod_b;

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

@@ -49,15 +49,11 @@ __global__ void device_batch_fft_ggsw_vector(double2 *dest, T *src,
  * global memory
  */
 template <typename T, typename ST, class params>
-void batch_fft_ggsw_vector(cudaStream_t *streams, uint32_t *gpu_indexes,
-                           uint32_t gpu_count, double2 *dest, T *src,
+void batch_fft_ggsw_vector(cuda_stream_t *stream, double2 *dest, T *src,
                            int8_t *d_mem, uint32_t r, uint32_t glwe_dim,
                            uint32_t polynomial_size, uint32_t level_count,
-                           uint32_t max_shared_memory) {
-  if (gpu_count != 1)
-    PANIC("GPU error (batch_fft_ggsw_vector): multi-GPU execution is not "
-          "supported yet.")
-  cudaSetDevice(gpu_indexes[0]);
+                           uint32_t gpu_index, uint32_t max_shared_memory) {
+  cudaSetDevice(stream->gpu_index);
 
   int shared_memory_size = sizeof(double) * polynomial_size;
 
@@ -66,11 +62,11 @@ void batch_fft_ggsw_vector(cudaStream_t *streams, uint32_t *gpu_indexes,
 
   if (max_shared_memory < shared_memory_size) {
     device_batch_fft_ggsw_vector<T, ST, params, NOSM>
-        <<<gridSize, blockSize, 0, streams[0]>>>(dest, src, d_mem);
+        <<<gridSize, blockSize, 0, stream->stream>>>(dest, src, d_mem);
   } else {
     device_batch_fft_ggsw_vector<T, ST, params, FULLSM>
-        <<<gridSize, blockSize, shared_memory_size, streams[0]>>>(dest, src,
-                                                                  d_mem);
+        <<<gridSize, blockSize, shared_memory_size, stream->stream>>>(dest, src,
+                                                                      d_mem);
   }
   check_cuda_error(cudaGetLastError());
 }

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

@@ -6,13 +6,12 @@
  * Head out to the equivalent operation on 64 bits for more details.
  */
 void cuda_keyswitch_lwe_ciphertext_vector_32(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void *lwe_output_indexes, void *lwe_array_in, void *lwe_input_indexes,
-    void *ksk, uint32_t lwe_dimension_in, uint32_t lwe_dimension_out,
-    uint32_t base_log, uint32_t level_count, uint32_t num_samples) {
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_output_indexes,
+    void *lwe_array_in, void *lwe_input_indexes, void *ksk,
+    uint32_t lwe_dimension_in, uint32_t lwe_dimension_out, uint32_t base_log,
+    uint32_t level_count, uint32_t num_samples) {
   cuda_keyswitch_lwe_ciphertext_vector(
-      static_cast<cudaStream_t>(stream), gpu_index,
-      static_cast<uint32_t *>(lwe_array_out),
+      stream, static_cast<uint32_t *>(lwe_array_out),
       static_cast<uint32_t *>(lwe_output_indexes),
       static_cast<uint32_t *>(lwe_array_in),
       static_cast<uint32_t *>(lwe_input_indexes), static_cast<uint32_t *>(ksk),
@@ -36,13 +35,12 @@ void cuda_keyswitch_lwe_ciphertext_vector_32(
  * 	- num_samples blocks of threads are launched
  */
 void cuda_keyswitch_lwe_ciphertext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void *lwe_output_indexes, void *lwe_array_in, void *lwe_input_indexes,
-    void *ksk, uint32_t lwe_dimension_in, uint32_t lwe_dimension_out,
-    uint32_t base_log, uint32_t level_count, uint32_t num_samples) {
+    cuda_stream_t *stream, void *lwe_array_out, void *lwe_output_indexes,
+    void *lwe_array_in, void *lwe_input_indexes, void *ksk,
+    uint32_t lwe_dimension_in, uint32_t lwe_dimension_out, uint32_t base_log,
+    uint32_t level_count, uint32_t num_samples) {
   cuda_keyswitch_lwe_ciphertext_vector(
-      static_cast<cudaStream_t>(stream), gpu_index,
-      static_cast<uint64_t *>(lwe_array_out),
+      stream, static_cast<uint64_t *>(lwe_array_out),
       static_cast<uint64_t *>(lwe_output_indexes),
       static_cast<uint64_t *>(lwe_array_in),
       static_cast<uint64_t *>(lwe_input_indexes), static_cast<uint64_t *>(ksk),

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

@@ -3,11 +3,8 @@
 
 #include "device.h"
 #include "gadget.cuh"
-#include "helper_multi_gpu.h"
-#include "polynomial/functions.cuh"
 #include "polynomial/polynomial_math.cuh"
 #include "torus.cuh"
-#include "utils/kernel_dimensions.cuh"
 #include <thread>
 #include <vector>
 
@@ -34,124 +31,110 @@ __device__ Torus *get_ith_block(Torus *ksk, int i, int level,
  * scaling factor) under key s2 instead of s1, with an increased noise
  *
  */
-// Each thread in x are used to calculate one output.
-// threads in y are used to paralelize the lwe_dimension_in loop.
-// shared memory is used to store intermediate results of the reduction.
 template <typename Torus>
 __global__ void
-keyswitch(Torus *lwe_array_out, const Torus *__restrict__ lwe_output_indexes,
-          const Torus *__restrict__ lwe_array_in,
-          const Torus *__restrict__ lwe_input_indexes,
-          const Torus *__restrict__ ksk, uint32_t lwe_dimension_in,
-          uint32_t lwe_dimension_out, uint32_t base_log, uint32_t level_count) {
-  const int tid = threadIdx.x + blockIdx.x * blockDim.x;
-  const int shmem_index = threadIdx.x + threadIdx.y * blockDim.x;
+keyswitch(Torus *lwe_array_out, Torus *lwe_output_indexes, Torus *lwe_array_in,
+          Torus *lwe_input_indexes, Torus *ksk, uint32_t lwe_dimension_in,
+          uint32_t lwe_dimension_out, uint32_t base_log, uint32_t level_count,
+          int lwe_lower, int lwe_upper, int cutoff) {
+  int tid = threadIdx.x;
 
   extern __shared__ int8_t sharedmem[];
-  Torus *lwe_acc_out = (Torus *)sharedmem;
+
+  Torus *local_lwe_array_out = (Torus *)sharedmem;
+
+  auto block_lwe_array_in = get_chunk(
+      lwe_array_in, lwe_input_indexes[blockIdx.x], lwe_dimension_in + 1);
   auto block_lwe_array_out = get_chunk(
-      lwe_array_out, lwe_output_indexes[blockIdx.y], lwe_dimension_out + 1);
+      lwe_array_out, lwe_output_indexes[blockIdx.x], lwe_dimension_out + 1);
 
-  if (tid <= lwe_dimension_out) {
+  auto gadget = GadgetMatrixSingle<Torus>(base_log, level_count);
 
-    Torus local_lwe_out = 0;
-    auto block_lwe_array_in = get_chunk(
-        lwe_array_in, lwe_input_indexes[blockIdx.y], lwe_dimension_in + 1);
+  int lwe_part_per_thd;
+  if (tid < cutoff) {
+    lwe_part_per_thd = lwe_upper;
+  } else {
+    lwe_part_per_thd = lwe_lower;
+  }
+  __syncthreads();
 
-    if (tid == lwe_dimension_out && threadIdx.y == 0) {
-      local_lwe_out = block_lwe_array_in[lwe_dimension_in];
-    }
-    const Torus mask_mod_b = (1ll << base_log) - 1ll;
+  for (int k = 0; k < lwe_part_per_thd; k++) {
+    int idx = tid + k * blockDim.x;
+    local_lwe_array_out[idx] = 0;
+  }
+  __syncthreads();
 
-    const int pack_size = (lwe_dimension_in + blockDim.y - 1) / blockDim.y;
-    const int start_i = pack_size * threadIdx.y;
-    const int end_i = SEL(lwe_dimension_in, pack_size * (threadIdx.y + 1),
-                          pack_size * (threadIdx.y + 1) <= lwe_dimension_in);
-
-    // This loop distribution seems to benefit the global mem reads
-    for (int i = start_i; i < end_i; i++) {
-      Torus a_i = round_to_closest_multiple(block_lwe_array_in[i], base_log,
-                                            level_count);
-      Torus state = a_i >> (sizeof(Torus) * 8 - base_log * level_count);
-
-      for (int j = 0; j < level_count; j++) {
-        auto ksk_block =
-            get_ith_block(ksk, i, j, lwe_dimension_out, level_count);
-        Torus decomposed = decompose_one<Torus>(state, mask_mod_b, base_log);
-        local_lwe_out -= (Torus)ksk_block[tid] * decomposed;
-      }
-    }
-
-    lwe_acc_out[shmem_index] = local_lwe_out;
+  if (tid == 0) {
+    local_lwe_array_out[lwe_dimension_out] =
+        block_lwe_array_in[lwe_dimension_in];
   }
 
-  if (tid <= lwe_dimension_out) {
-    for (int offset = blockDim.y / 2; offset > 0 && threadIdx.y < offset;
-         offset /= 2) {
-      __syncthreads();
-      lwe_acc_out[shmem_index] +=
-          lwe_acc_out[shmem_index + offset * blockDim.x];
+  for (int i = 0; i < lwe_dimension_in; i++) {
+
+    __syncthreads();
+
+    Torus a_i =
+        round_to_closest_multiple(block_lwe_array_in[i], base_log, level_count);
+
+    Torus state = a_i >> (sizeof(Torus) * 8 - base_log * level_count);
+    Torus mask_mod_b = (1ll << base_log) - 1ll;
+
+    for (int j = 0; j < level_count; j++) {
+      auto ksk_block = get_ith_block(ksk, i, j, lwe_dimension_out, level_count);
+      Torus decomposed = decompose_one<Torus>(state, mask_mod_b, base_log);
+      for (int k = 0; k < lwe_part_per_thd; k++) {
+        int idx = tid + k * blockDim.x;
+        local_lwe_array_out[idx] -= (Torus)ksk_block[idx] * decomposed;
+      }
     }
-    if (threadIdx.y == 0)
-      block_lwe_array_out[tid] = lwe_acc_out[shmem_index];
+  }
+
+  for (int k = 0; k < lwe_part_per_thd; k++) {
+    int idx = tid + k * blockDim.x;
+    block_lwe_array_out[idx] = local_lwe_array_out[idx];
   }
 }
 
+/// assume lwe_array_in in the gpu
 template <typename Torus>
 __host__ void cuda_keyswitch_lwe_ciphertext_vector(
-    cudaStream_t stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus *lwe_output_indexes, Torus *lwe_array_in, Torus *lwe_input_indexes,
-    Torus *ksk, uint32_t lwe_dimension_in, uint32_t lwe_dimension_out,
-    uint32_t base_log, uint32_t level_count, uint32_t num_samples) {
+    cuda_stream_t *stream, Torus *lwe_array_out, Torus *lwe_output_indexes,
+    Torus *lwe_array_in, Torus *lwe_input_indexes, Torus *ksk,
+    uint32_t lwe_dimension_in, uint32_t lwe_dimension_out, uint32_t base_log,
+    uint32_t level_count, uint32_t num_samples) {
 
-  cudaSetDevice(gpu_index);
+  cudaSetDevice(stream->gpu_index);
+  constexpr int ideal_threads = 128;
 
-  constexpr int num_threads_y = 32;
-  int num_blocks, num_threads_x;
+  int lwe_size = lwe_dimension_out + 1;
+  int lwe_lower, lwe_upper, cutoff;
+  if (lwe_size % ideal_threads == 0) {
+    lwe_lower = lwe_size / ideal_threads;
+    lwe_upper = lwe_size / ideal_threads;
+    cutoff = 0;
+  } else {
+    int y = ceil((double)lwe_size / (double)ideal_threads) * ideal_threads -
+            lwe_size;
+    cutoff = ideal_threads - y;
+    lwe_lower = lwe_size / ideal_threads;
+    lwe_upper = (int)ceil((double)lwe_size / (double)ideal_threads);
+  }
 
-  getNumBlocksAndThreads2D(lwe_dimension_out + 1, 512, num_threads_y,
-                           num_blocks, num_threads_x);
+  int lwe_size_after = lwe_size * num_samples;
 
-  int shared_mem = sizeof(Torus) * num_threads_y * num_threads_x;
-  dim3 grid(num_blocks, num_samples, 1);
-  dim3 threads(num_threads_x, num_threads_y, 1);
+  int shared_mem = sizeof(Torus) * lwe_size;
 
-  keyswitch<Torus><<<grid, threads, shared_mem, stream>>>(
+  cuda_memset_async(lwe_array_out, 0, sizeof(Torus) * lwe_size_after, stream);
+  check_cuda_error(cudaGetLastError());
+
+  dim3 grid(num_samples, 1, 1);
+  dim3 threads(ideal_threads, 1, 1);
+
+  keyswitch<Torus><<<grid, threads, shared_mem, stream->stream>>>(
       lwe_array_out, lwe_output_indexes, lwe_array_in, lwe_input_indexes, ksk,
-      lwe_dimension_in, lwe_dimension_out, base_log, level_count);
+      lwe_dimension_in, lwe_dimension_out, base_log, level_count, lwe_lower,
+      lwe_upper, cutoff);
   check_cuda_error(cudaGetLastError());
 }
 
-template <typename Torus>
-void execute_keyswitch_async(cudaStream_t *streams, uint32_t *gpu_indexes,
-                             uint32_t gpu_count,
-                             const LweArrayVariant<Torus> &lwe_array_out,
-                             const LweArrayVariant<Torus> &lwe_output_indexes,
-                             const LweArrayVariant<Torus> &lwe_array_in,
-                             const LweArrayVariant<Torus> &lwe_input_indexes,
-                             Torus **ksks, uint32_t lwe_dimension_in,
-                             uint32_t lwe_dimension_out, uint32_t base_log,
-                             uint32_t level_count, uint32_t num_samples) {
-
-  /// If the number of radix blocks is lower than the number of GPUs, not all
-  /// GPUs will be active and there will be 1 input per GPU
-  for (uint i = 0; i < gpu_count; i++) {
-    int num_samples_on_gpu = get_num_inputs_on_gpu(num_samples, i, gpu_count);
-
-    Torus *current_lwe_array_out = GET_VARIANT_ELEMENT(lwe_array_out, i);
-    Torus *current_lwe_output_indexes =
-        GET_VARIANT_ELEMENT(lwe_output_indexes, i);
-    Torus *current_lwe_array_in = GET_VARIANT_ELEMENT(lwe_array_in, i);
-    Torus *current_lwe_input_indexes =
-        GET_VARIANT_ELEMENT(lwe_input_indexes, i);
-
-    // Compute Keyswitch
-    cuda_keyswitch_lwe_ciphertext_vector<Torus>(
-        streams[i], gpu_indexes[i], current_lwe_array_out,
-        current_lwe_output_indexes, current_lwe_array_in,
-        current_lwe_input_indexes, ksks[i], lwe_dimension_in, lwe_dimension_out,
-        base_log, level_count, num_samples_on_gpu);
-  }
-}
-
 #endif

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

@@ -39,19 +39,36 @@ __device__ inline T round_to_closest_multiple(T x, uint32_t base_log,
 }
 
 template <typename T>
-__device__ __forceinline__ void modulus_switch(T input, T &output,
-                                               uint32_t log_modulus) {
-  constexpr uint32_t BITS = sizeof(T) * 8;
-
-  output = input + (((T)1) << (BITS - log_modulus - 1));
-  output >>= (BITS - log_modulus);
+__device__ __forceinline__ void rescale_torus_element(T element, T &output,
+                                                      uint32_t log_shift) {
+  output =
+      round((double)element / (double(std::numeric_limits<T>::max()) + 1.0) *
+            (double)log_shift);
 }
 
 template <typename T>
-__device__ __forceinline__ T modulus_switch(T input, uint32_t log_modulus) {
-  T output;
-  modulus_switch(input, output, log_modulus);
-  return output;
+__device__ __forceinline__ T rescale_torus_element(T element,
+                                                   uint32_t log_shift) {
+  return round((double)element / (double(std::numeric_limits<T>::max()) + 1.0) *
+               (double)log_shift);
 }
 
+template <>
+__device__ __forceinline__ void
+rescale_torus_element<uint32_t>(uint32_t element, uint32_t &output,
+                                uint32_t log_shift) {
+  output =
+      round(__uint2double_rn(element) /
+            (__uint2double_rn(std::numeric_limits<uint32_t>::max()) + 1.0) *
+            __uint2double_rn(log_shift));
+}
+
+template <>
+__device__ __forceinline__ void
+rescale_torus_element<uint64_t>(uint64_t element, uint64_t &output,
+                                uint32_t log_shift) {
+  output = round(__ull2double_rn(element) /
+                 (__ull2double_rn(std::numeric_limits<uint64_t>::max()) + 1.0) *
+                 __uint2double_rn(log_shift));
+}
 #endif // CNCRT_TORUS_H