chore(zk): add a shortint wrapper for the CompactPkeCrs

This allows to detect unused dispatch enums

.cargo/config.toml

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

.github/workflows/aws_tfhe_backward_compat_tests.yml

@@ -26,7 +26,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -44,16 +44,13 @@ jobs:
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
@@ -75,7 +72,7 @@ jobs:
           echo "branch=${BRANCH}" >> "${GITHUB_OUTPUT}"
 
       - name: Clone test data
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
           repository: zama-ai/tfhe-backward-compat-data
@@ -90,7 +87,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Backward compatibility tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -103,7 +100,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -114,7 +111,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (backward-compat-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_fast_tests.yml

@@ -26,6 +26,7 @@ jobs:
     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 }}
+      versionable_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.versionable_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 }}
@@ -50,24 +51,31 @@ jobs:
       any_file_changed: ${{ env.IS_PULL_REQUEST == 'false' || steps.aggregated-changes.outputs.any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@6b2903bdce6310cfbddd87c418f253cf29b2dec9
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
         with:
           since_last_remote_commit: true
           files_yaml: |
             dependencies:
               - tfhe/Cargo.toml
-              - concrete-csprng/**
+              - tfhe-csprng/**
+              - tfhe-fft/**
               - tfhe-zk-pok/**
+              - utils/tfhe-versionable/**
+              - utils/tfhe-versionable-derive/**
             csprng:
-              - concrete-csprng/**
+              - tfhe-csprng/**
             zk_pok:
               - tfhe-zk-pok/**
+            versionable:
+              - utils/tfhe-versionable/**
+              - utils/tfhe-versionable-derive/**
             core_crypto:
               - tfhe/src/core_crypto/**
             boolean:
@@ -103,6 +111,7 @@ jobs:
         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.versionable_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' ||
@@ -124,7 +133,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -144,29 +153,31 @@ jobs:
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
-      - name: Run concrete-csprng tests
+      - name: Run tfhe-csprng tests
         if: needs.should-run.outputs.csprng_test == 'true'
         run: |
-          make test_concrete_csprng
+          make test_tfhe_csprng
 
       - name: Run tfhe-zk-pok tests
         if: needs.should-run.outputs.zk_pok_test == 'true'
         run: |
           make test_zk_pok
 
+      - name: Run tfhe-versionable tests
+        if: needs.should-run.outputs.versionable_test == 'true'
+        run: |
+          make test_versionable
+
       - name: Run core tests
         if: needs.should-run.outputs.core_crypto_test == 'true'
         run: |
@@ -182,10 +193,37 @@ jobs:
         run: |
           make test_user_doc
 
+      - name: Get Node version
+        run: |
+          echo "NODE_VERSION=$(make node_version)" >> "${GITHUB_ENV}"
+
+      - name: Node cache restoration
+        id: node-cache
+        uses: actions/cache/restore@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
+        with:
+          path: |
+            ~/.nvm
+            ~/.npm
+          key: node-${{ env.NODE_VERSION }}
+
+      - name: Install Node
+        if: steps.node-cache.outputs.cache-hit != 'true'
+        run: |
+          make install_node
+
+      - name: Node cache save
+        uses: actions/cache/save@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
+        if: steps.node-cache.outputs.cache-hit != 'true'
+        with:
+          path: |
+            ~/.nvm
+            ~/.npm
+          key: node-${{ env.NODE_VERSION }}
+
       - name: Run js on wasm API tests
         if: needs.should-run.outputs.wasm_test == 'true'
         run: |
-          make test_nodejs_wasm_api_in_docker
+          make test_nodejs_wasm_api_ci
 
       - name: Gen Keys if required
         if: needs.should-run.outputs.shortint_test == 'true' ||
@@ -208,14 +246,18 @@ jobs:
         run: |
           make test_high_level_api
 
-      - name: Run safe deserialization tests
+      - name: Run safe serialization tests
         run: |
-          make test_safe_deserialization
+          make test_safe_serialization
+
+      - name: Run zk tests
+        run: |
+          make test_zk
 
       - name: Slack Notification
-        if: ${{ always() }}
+        if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Fast AWS tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -228,7 +270,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -239,7 +281,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (fast-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_integer_tests.yml

@@ -19,28 +19,63 @@ on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
   pull_request:
-    types: [ labeled ]
+    types: [labeled]
   push:
     branches:
       - main
-  schedule:
-    # Nightly tests @ 3AM after each work day
-    - cron: "0 3 * * MON-FRI"
 
 jobs:
+  should-run:
+    if:
+      (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs') ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
+      (github.event_name == 'pull_request' && contains(github.event.label.name, 'approved')) ||
+      github.event_name == 'workflow_dispatch'
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: write
+    outputs:
+      integer_test: ${{ github.event_name == 'workflow_dispatch' ||
+        steps.changed-files.outputs.integer_any_changed }}
+    steps:
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: "false"
+
+      - name: Check for file changes
+        id: changed-files
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
+        with:
+          since_last_remote_commit: true
+          files_yaml: |
+            integer:
+              - tfhe/Cargo.toml
+              - tfhe-csprng/**
+              - tfhe-fft/**
+              - tfhe-zk-pok/**
+              - tfhe/src/core_crypto/**
+              - tfhe/src/shortint/**
+              - tfhe/src/integer/**
+              - .github/workflows/aws_tfhe_integer_tests.yml
+
   setup-instance:
     name: Setup instance (unsigned-integer-tests)
-    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'
+    needs: should-run
+    if:
+      (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs' && needs.should-run.outputs.integer_test == 'true') ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
+      (github.event_name == 'pull_request' && contains(github.event.label.name, 'approved')) ||
+      github.event_name == 'workflow_dispatch'
     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
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -58,16 +93,13 @@ jobs:
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
-          persist-credentials: 'false'
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+          persist-credentials: "false"
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
@@ -93,9 +125,9 @@ jobs:
           AVX512_SUPPORT=ON NO_BIG_PARAMS=${{ env.NO_BIG_PARAMS }} BIG_TESTS_INSTANCE=TRUE make test_unsigned_integer_ci
 
       - name: Slack Notification
-        if: ${{ always() }}
+        if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Unsigned Integer tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -103,12 +135,12 @@ jobs:
   teardown-instance:
     name: Teardown instance (unsigned-integer-tests)
     if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, unsigned-integer-tests ]
+    needs: [setup-instance, 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@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -119,7 +151,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (unsigned-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_signed_integer_tests.yml

@@ -19,28 +19,63 @@ on:
   # Allows you to run this workflow manually from the Actions tab as an alternative.
   workflow_dispatch:
   pull_request:
-    types: [ labeled ]
+    types: [labeled]
   push:
     branches:
       - main
-  schedule:
-    # Nightly tests @ 3AM after each work day
-    - cron: "0 3 * * MON-FRI"
 
 jobs:
+  should-run:
+    if:
+      (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs') ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
+      (github.event_name == 'pull_request' && contains(github.event.label.name, 'approved')) ||
+      github.event_name == 'workflow_dispatch'
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: write
+    outputs:
+      integer_test: ${{ github.event_name == 'workflow_dispatch' ||
+        steps.changed-files.outputs.integer_any_changed }}
+    steps:
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+          persist-credentials: "false"
+
+      - name: Check for file changes
+        id: changed-files
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
+        with:
+          since_last_remote_commit: true
+          files_yaml: |
+            integer:
+              - tfhe/Cargo.toml
+              - tfhe-csprng/**
+              - tfhe-fft/**
+              - tfhe-zk-pok/**
+              - tfhe/src/core_crypto/**
+              - tfhe/src/shortint/**
+              - tfhe/src/integer/**
+              - .github/workflows/aws_tfhe_signed_integer_tests.yml
+
   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'
+    name: Setup instance (unsigned-integer-tests)
+    needs: should-run
+    if:
+      (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs' && needs.should-run.outputs.integer_test == 'true') ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
+      (github.event_name == 'pull_request' && contains(github.event.label.name, 'approved')) ||
+      github.event_name == 'workflow_dispatch'
     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
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -58,16 +93,13 @@ jobs:
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
-          persist-credentials: 'false'
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+          persist-credentials: "false"
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
@@ -97,9 +129,9 @@ jobs:
           AVX512_SUPPORT=ON NO_BIG_PARAMS=${{ env.NO_BIG_PARAMS }} BIG_TESTS_INSTANCE=TRUE make test_signed_integer_ci
 
       - name: Slack Notification
-        if: ${{ always() }}
+        if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Signed Integer tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -107,12 +139,12 @@ jobs:
   teardown-instance:
     name: Teardown instance (signed-integer-tests)
     if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, signed-integer-tests ]
+    needs: [setup-instance, 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@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -123,7 +155,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (signed-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_tests.yml

@@ -40,6 +40,9 @@ jobs:
       shortint_test: ${{ env.IS_PULL_REQUEST == 'false' ||
         steps.changed-files.outputs.shortint_any_changed ||
         steps.changed-files.outputs.dependencies_any_changed }}
+      strings_test: ${{ env.IS_PULL_REQUEST == 'false' ||
+        steps.changed-files.outputs.strings_any_changed ||
+        steps.changed-files.outputs.dependencies_any_changed }}
       high_level_api_test: ${{ env.IS_PULL_REQUEST == 'false' ||
         steps.changed-files.outputs.high_level_api_any_changed ||
         steps.changed-files.outputs.dependencies_any_changed }}
@@ -57,22 +60,24 @@ jobs:
       any_file_changed: ${{ env.IS_PULL_REQUEST == 'false' || steps.aggregated-changes.outputs.any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@6b2903bdce6310cfbddd87c418f253cf29b2dec9
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
         with:
           since_last_remote_commit: true
           files_yaml: |
             dependencies:
               - tfhe/Cargo.toml
-              - concrete-csprng/**
+              - tfhe-csprng/**
+              - tfhe-fft/**
               - tfhe-zk-pok/**
             csprng:
-              - concrete-csprng/**
+              - tfhe-csprng/**
             zk_pok:
               - tfhe-zk-pok/**
             core_crypto:
@@ -83,6 +88,11 @@ jobs:
             shortint:
               - tfhe/src/core_crypto/**
               - tfhe/src/shortint/**
+            strings:
+              - tfhe/src/core_crypto/**
+              - tfhe/src/shortint/**
+              - tfhe/src/integer/**
+              - tfhe/src/strings/**
             high_level_api:
               - tfhe/src/**
               - '!tfhe/src/c_api/**'
@@ -112,6 +122,7 @@ jobs:
           steps.changed-files.outputs.core_crypto_any_changed == 'true' ||
           steps.changed-files.outputs.boolean_any_changed == 'true' ||
           steps.changed-files.outputs.shortint_any_changed == 'true' ||
+          steps.changed-files.outputs.strings_any_changed == 'true' ||
           steps.changed-files.outputs.high_level_api_any_changed == 'true' ||
           steps.changed-files.outputs.c_api_any_changed == 'true' ||
           steps.changed-files.outputs.examples_any_changed == 'true' ||
@@ -131,7 +142,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -151,23 +162,20 @@ jobs:
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
-      - name: Run concrete-csprng tests
+      - name: Run tfhe-csprng tests
         if: needs.should-run.outputs.csprng_test == 'true'
         run: |
-          make test_concrete_csprng
+          make test_tfhe_csprng
 
       - name: Run tfhe-zk-pok tests
         if: needs.should-run.outputs.zk_pok_test == 'true'
@@ -204,6 +212,11 @@ jobs:
         run: |
           BIG_TESTS_INSTANCE=TRUE make test_shortint_ci
 
+      - name: Run strings tests
+        if: needs.should-run.outputs.strings_test == 'true'
+        run: |
+          BIG_TESTS_INSTANCE=TRUE make test_strings
+
       - name: Run high-level API tests
         if: needs.should-run.outputs.high_level_api_test == 'true'
         run: |
@@ -222,9 +235,9 @@ jobs:
           make test_kreyvium
 
       - name: Slack Notification
-        if: ${{ always() }}
+        if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "CPU tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -237,7 +250,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -248,7 +261,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cpu-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/aws_tfhe_wasm_tests.yml

@@ -27,7 +27,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -45,39 +45,68 @@ jobs:
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
+      - name: Get Node version
+        run: |
+          echo "NODE_VERSION=$(make node_version)" >> "${GITHUB_ENV}"
+
+      - name: Node cache restoration
+        id: node-cache
+        uses: actions/cache/restore@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
+        with:
+          path: |
+            ~/.nvm
+            ~/.npm
+          key: node-${{ env.NODE_VERSION }}
+
       - name: Install Node
+        if: steps.node-cache.outputs.cache-hit != 'true'
         run: |
           make install_node
 
+      - name: Node cache save
+        uses: actions/cache/save@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
+        if: steps.node-cache.outputs.cache-hit != 'true'
+        with:
+          path: |
+            ~/.nvm
+            ~/.npm
+          key: node-${{ env.NODE_VERSION }}
+
+      - name: Install web resources
+        run: |
+          make install_chrome_browser
+          make install_chrome_web_driver
+
       - name: Run fmt checks
         run: |
           make check_fmt_js
 
       - name: Run js on wasm API tests
         run: |
-          make test_nodejs_wasm_api_in_docker
+          make test_nodejs_wasm_api_ci
 
       - name: Run parallel wasm tests
         run: |
-          make test_web_js_api_parallel_ci
+          make test_web_js_api_parallel_chrome_ci
+
+      - name: Run x86_64/wasm zk compatibility tests
+        run: |
+          make test_zk_wasm_x86_compat_ci
 
       - name: Slack Notification
-        if: ${{ always() }}
+        if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "WASM tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -90,7 +119,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -101,7 +130,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (wasm-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_boolean.yml

@@ -29,7 +29,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -48,9 +48,10 @@ jobs:
     continue-on-error: true
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -60,13 +61,8 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
       - name: Install rust
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: nightly
 
@@ -84,8 +80,7 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
-          --name-suffix avx512 \
-          --throughput
+          --name-suffix avx512
 
       - name: Measure key sizes
         run: |
@@ -94,17 +89,17 @@ jobs:
       - name: Parse key sizes results
         run: |
           python3 ./ci/benchmark_parser.py tfhe/boolean_key_sizes.csv ${{ env.RESULTS_FILENAME }} \
-          --key-sizes \
+          --object-sizes \
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
         with:
           name: ${{ github.sha }}_boolean
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: zama-ai/slab
           path: slab
@@ -113,21 +108,13 @@ jobs:
       - 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 }}
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Boolean benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -140,7 +127,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -151,7 +138,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (boolean-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_core_crypto.yml

@@ -26,7 +26,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -44,9 +44,10 @@ jobs:
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -56,13 +57,8 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
       - name: Install rust
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: nightly
 
@@ -82,17 +78,16 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --name-suffix avx512 \
-          --walk-subdirs \
-          --throughput
+          --walk-subdirs
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
         with:
           name: ${{ github.sha }}_core_crypto
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: zama-ai/slab
           path: slab
@@ -101,21 +96,13 @@ jobs:
       - name: Send data to Slab
         shell: bash
         run: |
-          echo "Computing HMac on downloaded artifact"
-          SIGNATURE="$(slab/scripts/hmac_calculator.sh ${{ env.RESULTS_FILENAME }} '${{ secrets.JOB_SECRET }}')"
-          echo "Sending results to Slab..."
-          curl -v -k \
-          -H "Content-Type: application/json" \
-          -H "X-Slab-Repository: ${{ github.repository }}" \
-          -H "X-Slab-Command: store_data_v2" \
-          -H "X-Hub-Signature-256: sha256=${SIGNATURE}" \
-          -d @${{ env.RESULTS_FILENAME }} \
-          ${{ secrets.SLAB_URL }}
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "PBS benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -128,7 +115,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -139,7 +126,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (core-crypto-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_erc20.yml

@@ -0,0 +1,141 @@
+# Run all ERC20 benchmarks on an AWS instance and return parsed results to Slab CI bot.
+name: ERC20 benchmarks
+
+on:
+  workflow_dispatch:
+  schedule:
+    # Weekly benchmarks will be triggered each Saturday at 5a.m.
+    - cron: '0 5 * * 6'
+
+env:
+  CARGO_TERM_COLOR: always
+  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
+jobs:
+  setup-instance:
+    name: Setup instance (erc20-benchmarks)
+    runs-on: ubuntu-latest
+    if: github.event_name == 'workflow_dispatch' ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: aws
+          profile: bench
+
+  erc20-benchmarks:
+    name: Execute ERC20 benchmarks
+    needs: setup-instance
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    concurrency:
+      group: ${{ github.workflow }}_${{ github.ref }}
+      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
+    continue-on-error: true
+    timeout-minutes: 720  # 12 hours
+    steps:
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(date --iso-8601=seconds)";
+            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
+            echo "COMMIT_HASH=$(git describe --tags --dirty)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
+        with:
+          toolchain: nightly
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Run benchmarks
+        run: |
+          make bench_hlapi_erc20
+
+      - name: Parse results
+        run: |
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware "hpc7a.96xlarge" \
+          --project-version "${{ env.COMMIT_HASH }}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${{ env.COMMIT_DATE }}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512
+
+      - name: Parse PBS counts
+        run: |
+          python3 ./ci/benchmark_parser.py tfhe/erc20_pbs_count.csv ${{ env.RESULTS_FILENAME }} \
+          --object-sizes \
+          --append-results
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        with:
+          name: ${{ github.sha }}_erc20
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "ERC20 benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (erc20-benchmarks)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, erc20-benchmarks ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (erc20-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_4090.yml

@@ -39,9 +39,10 @@ jobs:
 
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -53,12 +54,12 @@ jobs:
           echo "FAST_BENCH=TRUE" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: zama-ai/slab
           path: slab
@@ -78,11 +79,10 @@ jobs:
           --branch ${{ github.ref_name }} \
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
-          --walk-subdirs \
-          --throughput
+          --walk-subdirs
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
         with:
           name: ${{ github.sha }}_integer_multi_bit_gpu_default
           path: ${{ env.RESULTS_FILENAME }}
@@ -90,21 +90,13 @@ jobs:
       - name: Send data to Slab
         shell: bash
         run: |
-          echo "Computing HMac on results file"
-          SIGNATURE="$(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 }}
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
 
       - name: Slack Notification
-        if: ${{ always() }}
+        if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Integer RTX 4090 full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -121,9 +113,10 @@ jobs:
 
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -134,12 +127,12 @@ jobs:
           } >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: zama-ai/slab
           path: slab
@@ -161,10 +154,10 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
-          --throughput
+      
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
         with:
           name: ${{ github.sha }}_core_crypto
           path: ${{ env.RESULTS_FILENAME }}
@@ -184,9 +177,9 @@ jobs:
           ${{ secrets.SLAB_URL }}
 
       - name: Slack Notification
-        if: ${{ !success() && !cancelled() }}
+        if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Core crypto RTX 4090 full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_core_crypto.yml

@@ -27,7 +27,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -56,7 +56,7 @@ jobs:
       - name: Install dependencies
         run: |
           sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev libclang-dev
           wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
           tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
           cd cmake-${{ env.CMAKE_VERSION }}
@@ -65,9 +65,10 @@ jobs:
           sudo make install
 
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -83,7 +84,7 @@ jobs:
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: nightly
 
@@ -124,17 +125,16 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --name-suffix avx512 \
-          --walk-subdirs \
-          --throughput
+          --walk-subdirs
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
         with:
           name: ${{ github.sha }}_core_crypto
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: zama-ai/slab
           path: slab
@@ -143,16 +143,8 @@ jobs:
       - name: Send data to Slab
         shell: bash
         run: |
-          echo "Computing HMac on downloaded artifact"
-          SIGNATURE="$(slab/scripts/hmac_calculator.sh ${{ env.RESULTS_FILENAME }} '${{ secrets.JOB_SECRET }}')"
-          echo "Sending results to Slab..."
-          curl -v -k \
-          -H "Content-Type: application/json" \
-          -H "X-Slab-Repository: ${{ github.repository }}" \
-          -H "X-Slab-Command: store_data_v2" \
-          -H "X-Hub-Signature-256: sha256=${SIGNATURE}" \
-          -d @${{ env.RESULTS_FILENAME }} \
-          ${{ secrets.SLAB_URL }}
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
 
   slack-notify:
     name: Slack Notification
@@ -162,7 +154,7 @@ jobs:
     continue-on-error: true
     steps:
       - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         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 }})"
@@ -175,7 +167,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -186,7 +178,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-core-crypto-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_erc20.yml

@@ -0,0 +1,41 @@
+# Run CUDA ERC20 benchmarks on a Hyperstack VM and return parsed results to Slab CI bot.
+name: Cuda ERC20 benchmarks
+
+on:
+  workflow_dispatch:
+    inputs:
+      profile:
+        description: "Instance type"
+        required: true
+        type: choice
+        options:
+          - "l40 (n3-L40x1)"
+          - "single-h100 (n3-H100x1)"
+          - "2-h100 (n3-H100x2)"
+          - "multi-h100 (n3-H100x8)"
+
+jobs:
+  parse-inputs:
+    runs-on: ubuntu-latest
+    outputs:
+      profile: ${{ steps.parse_profile.outputs.profile }}
+      hardware_name: ${{ steps.parse_hardware_name.outputs.name }}
+    steps:
+      - name: Parse profile
+        id: parse_profile
+        run: |
+          echo "profile=$(echo '${{ inputs.profile }}' | sed 's|\(.*\)[[:space:]](.*)|\1|')" >> "${GITHUB_OUTPUT}"
+
+      - name: Parse hardware name
+        id: parse_hardware_name
+        run: |
+          echo "name=$(echo '${{ inputs.profile }}' | sed 's|.*[[:space:]](\(.*\))|\1|')" >> "${GITHUB_OUTPUT}"
+
+  run-benchmarks:
+    name: Run benchmarks
+    needs: parse-inputs
+    uses: ./.github/workflows/benchmark_gpu_erc20_common.yml
+    with:
+      profile: ${{ needs.parse-inputs.outputs.profile }}
+      hardware_name: ${{ needs.parse-inputs.outputs.hardware_name }}
+    secrets: inherit

.github/workflows/benchmark_gpu_erc20_common.yml

@@ -1,11 +1,35 @@
-# Run integer benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
-name: Integer GPU benchmarks
+# Run ERC20 benchmarks on an instance with CUDA and return parsed results to Slab CI bot.
+name: Cuda ERC20 benchmarks - common
 
 on:
-  workflow_dispatch:
-  push:
-    branches:
-      - main
+  workflow_call:
+    inputs:
+      backend:
+        type: string
+        default: hyperstack
+      profile:
+        type: string
+        required: true
+      hardware_name:
+        type: string
+        required: true
+    secrets:
+      FHE_ACTIONS_TOKEN:
+        required: true
+      SLAB_ACTION_TOKEN:
+        required: true
+      SLAB_BASE_URL:
+        required: true
+      SLAB_URL:
+        required: true
+      JOB_SECRET:
+        required: true
+      SLACK_CHANNEL:
+        required: true
+      BOT_USERNAME:
+        required: true
+      SLACK_WEBHOOK:
+        required: true
 
 env:
   CARGO_TERM_COLOR: always
@@ -21,26 +45,26 @@ env:
 
 jobs:
   setup-instance:
-    name: Setup instance (cuda-integer-benchmarks)
+    name: Setup instance (cuda-erc20-benchmarks)
     runs-on: ubuntu-latest
     if:  github.event_name == 'workflow_dispatch' ||
-      (github.event_name == 'push' && github.repository == 'zama-ai/tfhe-rs')
+      (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
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         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
+          backend: ${{ inputs.backend }}
+          profile: ${{ inputs.profile }}
 
-  cuda-integer-benchmarks:
-    name: Execute GPU integer benchmarks
+  cuda-erc20-benchmarks:
+    name: Cuda ERC20 benchmarks (${{ inputs.profile }})
     needs: setup-instance
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     strategy:
@@ -68,9 +92,10 @@ jobs:
           sudo make install
 
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -86,7 +111,7 @@ jobs:
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install rust
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: nightly
 
@@ -114,43 +139,31 @@ jobs:
         if: ${{ !cancelled() }}
         run: nvidia-smi
 
-      - name: Run benchmarks with AVX512
+      - name: Run benchmarks
         run: |
-          make FAST_BENCH=TRUE BENCH_OP_FLAVOR=default bench_integer_gpu
-
-      - name: Parse benchmarks to csv
-        run: |
-          make PARSE_INTEGER_BENCH_CSV_FILE=${{ env.PARSE_INTEGER_BENCH_CSV_FILE }} \
-            parse_integer_benches
-
-      - name: Upload csv results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
-        with:
-          name: ${{ github.sha }}_csv_integer
-          path: ${{ env.PARSE_INTEGER_BENCH_CSV_FILE }}
+          make bench_hlapi_erc20_gpu
 
       - name: Parse results
         run: |
           python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
           --database tfhe_rs \
-          --hardware "n3-H100x1" \
+          --hardware "${{ inputs.hardware_name }}" \
           --backend gpu \
           --project-version "${{ env.COMMIT_HASH }}" \
           --branch ${{ github.ref_name }} \
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
-          --name-suffix avx512 \
-          --throughput
+          --name-suffix avx512
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
         with:
-          name: ${{ github.sha }}_integer
+          name: ${{ github.sha }}_erc20
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: zama-ai/slab
           path: slab
@@ -159,39 +172,31 @@ jobs:
       - 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 }}
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
 
   slack-notify:
     name: Slack Notification
-    needs: [ setup-instance, cuda-integer-benchmarks ]
+    needs: [ setup-instance, cuda-erc20-benchmarks ]
     runs-on: ubuntu-latest
-    if: ${{ always() && needs.cuda-integer-benchmarks.result != 'skipped' && failure() }}
+    if: ${{ always() && needs.cuda-erc20-benchmarks.result != 'skipped' && failure() }}
     continue-on-error: true
     steps:
       - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
-          SLACK_COLOR: ${{ needs.cuda-integer-benchmarks.result }}
-          SLACK_MESSAGE: "Integer GPU benchmarks finished with status: ${{ needs.cuda-integer-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_COLOR: ${{ needs.cuda-erc20-benchmarks.result }}
+          SLACK_MESSAGE: "Cuda ERC20 benchmarks (${{ inputs.profile }}) finished with status: ${{ needs.cuda-erc20-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
-    name: Teardown instance (cuda-integer-benchmarks)
+    name: Teardown instance (cuda-erc20-${{ inputs.profile }}-benchmarks)
     if: ${{ always() && needs.setup-instance.result != 'skipped' }}
-    needs: [ setup-instance, cuda-integer-benchmarks, slack-notify ]
+    needs: [ setup-instance, cuda-erc20-benchmarks, slack-notify ]
     runs-on: ubuntu-latest
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -202,7 +207,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-integer-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "Instance teardown (cuda-erc20-${{ inputs.profile }}-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_erc20_weekly.yml

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

.github/workflows/benchmark_gpu_integer.yml

@@ -0,0 +1,78 @@
+# Run CUDA benchmarks on a Hyperstack VM and return parsed results to Slab CI bot.
+name: Cuda benchmarks
+
+on:
+  workflow_dispatch:
+    inputs:
+      profile:
+        description: "Instance type"
+        required: true
+        type: choice
+        options:
+          - "l40 (n3-L40x1)"
+          - "single-h100 (n3-H100x1)"
+          - "2-h100 (n3-H100x2)"
+          - "4-h100 (n3-H100x4)"
+          - "multi-h100 (n3-H100x8)"
+          - "multi-h100-nvlink (n3-H100x8-NVLink)"
+          - "multi-a100-nvlink (n3-A100x8-NVLink)"
+      command:
+        description: "Benchmark command to run"
+        type: choice
+        default: integer_multi_bit
+        options:
+          - integer
+          - integer_multi_bit
+          - integer_compression
+          - pbs
+          - ks
+      op_flavor:
+        description: "Operations set to run"
+        type: choice
+        default: default
+        options:
+          - default
+          - fast_default
+          - unchecked
+      all_precisions:
+        description: "Run all precisions"
+        type: boolean
+        default: false
+      bench_type:
+        description: "Benchmarks type"
+        type: choice
+        default: latency
+        options:
+          - latency
+          - throughput
+          - both
+
+jobs:
+  parse-inputs:
+    runs-on: ubuntu-latest
+    outputs:
+      profile: ${{ steps.parse_profile.outputs.profile }}
+      hardware_name: ${{ steps.parse_hardware_name.outputs.name }}
+    steps:
+      - name: Parse profile
+        id: parse_profile
+        run: |
+          echo "profile=$(echo '${{ inputs.profile }}' | sed 's|\(.*\)[[:space:]](.*)|\1|')" >> "${GITHUB_OUTPUT}"
+
+      - name: Parse hardware name
+        id: parse_hardware_name
+        run: |
+          echo "name=$(echo '${{ inputs.profile }}' | sed 's|.*[[:space:]](\(.*\))|\1|')" >> "${GITHUB_OUTPUT}"
+
+  run-benchmarks:
+    name: Run benchmarks
+    needs: parse-inputs
+    uses: ./.github/workflows/benchmark_gpu_integer_common.yml
+    with:
+      profile: ${{ needs.parse-inputs.outputs.profile }}
+      hardware_name: ${{ needs.parse-inputs.outputs.hardware_name }}
+      command: ${{ inputs.command }}
+      op_flavor: ${{ inputs.op_flavor }}
+      bench_type: ${{ inputs.bench_type }}
+      all_precisions: ${{ inputs.all_precisions }}
+    secrets: inherit

.github/workflows/benchmark_gpu_integer_common.yml

@@ -0,0 +1,289 @@
+# Run integer benchmarks on CUDA instance and return parsed results to Slab CI bot.
+name: Cuda benchmarks - common
+
+on:
+  workflow_call:
+    inputs:
+      backend:
+        type: string
+        default: hyperstack
+      profile:
+        type: string
+        required: true
+      hardware_name:
+        type: string
+        required: true
+      command: # Use a comma separated values to generate an array
+        type: string
+        required: true
+      op_flavor: # Use a comma separated values to generate an array
+        type: string
+        required: true
+      bench_type:
+        type: string
+        default: latency
+      all_precisions:
+        type: boolean
+        default: false
+    secrets:
+      FHE_ACTIONS_TOKEN:
+        required: true
+      SLAB_ACTION_TOKEN:
+        required: true
+      SLAB_BASE_URL:
+        required: true
+      SLAB_URL:
+        required: true
+      JOB_SECRET:
+        required: true
+      SLACK_CHANNEL:
+        required: true
+      BOT_USERNAME:
+        required: true
+      SLACK_WEBHOOK:
+        required: true
+
+env:
+  CARGO_TERM_COLOR: always
+  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  FAST_BENCH: TRUE
+
+jobs:
+  prepare-matrix:
+    name: Prepare operations matrix
+    runs-on: ubuntu-latest
+    outputs:
+      command: ${{ steps.set_command.outputs.command }}
+      op_flavor: ${{ steps.set_op_flavor.outputs.op_flavor }}
+      bench_type: ${{ steps.set_bench_type.outputs.bench_type }}
+    steps:
+      - name: Set single command
+        if: ${{ !contains(inputs.command, ',')}}
+        run: |
+          echo "COMMAND=[\"${{ inputs.command }}\"]" >> "${GITHUB_ENV}"
+
+      - name: Set multiple commands
+        if: ${{ contains(inputs.command, ',')}}
+        run: |
+          PARSED_COMMAND=$(echo "${{ inputs.command }}" | sed 's/[[:space:]]*,[[:space:]]*/\\", \\"/g')
+          echo "COMMAND=[\"${PARSED_COMMAND}\"]" >> "${GITHUB_ENV}"
+
+      - name: Set single operations flavor
+        if: ${{ !contains(inputs.op_flavor, ',')}}
+        run: |
+          echo "OP_FLAVOR=[\"${{ inputs.op_flavor }}\"]" >> "${GITHUB_ENV}"
+
+      - name: Set multiple operations flavors
+        if: ${{ contains(inputs.op_flavor, ',')}}
+        run: |
+          PARSED_OP_FLAVOR=$(echo "${{ inputs.op_flavor }}" | sed 's/[[:space:]]*,[[:space:]]*/", "/g')
+          echo "OP_FLAVOR=[\"${PARSED_OP_FLAVOR}\"]" >> "${GITHUB_ENV}"
+
+      - name: Set benchmark types
+        run: |
+          if [[ "${{ inputs.bench_type }}" == "both" ]]; then
+            echo "BENCH_TYPE=[\"latency\", \"throughput\"]" >> "${GITHUB_ENV}"
+          else
+            echo "BENCH_TYPE=[\"${{ inputs.bench_type }}\"]" >> "${GITHUB_ENV}"
+          fi
+
+      - name: Set command output
+        id: set_command
+        run: |
+          echo "command=${{ toJSON(env.COMMAND) }}" >> "${GITHUB_OUTPUT}"
+
+      - name: Set operation flavor output
+        id: set_op_flavor
+        run: |
+          echo "op_flavor=${{ toJSON(env.OP_FLAVOR) }}" >> "${GITHUB_OUTPUT}"
+
+      - name: Set benchmark types output
+        id: set_bench_type
+        run: |
+          echo "bench_type=${{ toJSON(env.BENCH_TYPE) }}" >> "${GITHUB_OUTPUT}"
+
+  setup-instance:
+    name: Setup instance (cuda-${{ inputs.profile }}-benchmarks)
+    needs: prepare-matrix
+    runs-on: ubuntu-latest
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: ${{ inputs.backend }}
+          profile: ${{ inputs.profile }}
+
+  cuda-benchmarks:
+    name: Cuda benchmarks (${{ inputs.profile }})
+    needs: [ prepare-matrix, setup-instance ]
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    timeout-minutes: 1440 # 24 hours
+    continue-on-error: true
+    strategy:
+      fail-fast: false
+      max-parallel: 1
+      matrix:
+        command: ${{ fromJSON(needs.prepare-matrix.outputs.command) }}
+        op_flavor: ${{ fromJSON(needs.prepare-matrix.outputs.op_flavor) }}
+        bench_type: ${{ fromJSON(needs.prepare-matrix.outputs.bench_type) }}
+        # explicit include-based build matrix, of known valid options
+        include:
+          - os: ubuntu-22.04
+            cuda: "12.2"
+            gcc: 11
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
+    steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev libclang-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(date --iso-8601=seconds)";
+            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
+            echo "COMMIT_HASH=$(git describe --tags --dirty)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Set up home
+        # "Install rust" step require root user to have a HOME directory which is not set.
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
+        with:
+          toolchain: nightly
+
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CUDA_PATH=$CUDA_PATH";
+            echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH";
+            echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc";
+          } >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+          } >> "${GITHUB_ENV}"
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
+      - name: Should run benchmarks with all precisions
+        if: inputs.all_precisions
+        run: |
+          echo "FAST_BENCH=FALSE" >> "${GITHUB_ENV}"
+
+      - name: Run benchmarks
+        run: |
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} BENCH_TYPE=${{ matrix.bench_type }} bench_${{ matrix.command }}_gpu
+
+      - name: Parse results
+        run: |
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --hardware "${{ inputs.hardware_name }}" \
+          --backend gpu \
+          --project-version "${{ env.COMMIT_HASH }}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${{ env.COMMIT_DATE }}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512 \
+          --bench-type ${{ matrix.bench_type }}
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        with:
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
+
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-benchmarks ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-benchmarks.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ needs.cuda-benchmarks.result }}
+          SLACK_MESSAGE: "Cuda benchmarks (${{ inputs.profile }}) finished with status: ${{ needs.cuda-benchmarks.result }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (cuda-${{ inputs.profile }}-benchmarks)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, cuda-benchmarks, slack-notify ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cuda-${{ inputs.profile }}-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_gpu_integer_weekly.yml

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

.github/workflows/benchmark_integer.yml

@@ -8,6 +8,14 @@ on:
         description: "Run all precisions"
         type: boolean
         default: false
+      bench_type:
+        description: "Benchmarks type"
+        type: choice
+        default: latency
+        options:
+          - latency
+          - throughput
+          - both
 
   schedule:
     # Weekly benchmarks will be triggered each Saturday at 1a.m.
@@ -36,10 +44,10 @@ jobs:
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
     outputs:
       op_flavor: ${{ steps.set_op_flavor.outputs.op_flavor }}
+      bench_type: ${{ steps.set_bench_type.outputs.bench_type }}
     steps:
       - name: Weekly benchmarks
-        if: github.event_name == 'workflow_dispatch' ||
-          github.event.schedule == '0 1 * * 6'
+        if: github.event.schedule == '0 1 * * 6'
         run: |
           echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
 
@@ -48,11 +56,31 @@ jobs:
         run: |
           echo "OP_FLAVOR=[\"default\", \"smart\", \"unchecked\", \"misc\"]" >> "${GITHUB_ENV}"
 
-      -  name: Set operation flavor output
-         id: set_op_flavor
-         run: |
+      - name: Set benchmark types
+        if: github.event_name == 'workflow_dispatch'
+        run: |
+          echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
+          if [[ "${{ inputs.bench_type }}" == "both" ]]; then
+            echo "BENCH_TYPE=[\"latency\", \"throughput\"]" >> "${GITHUB_ENV}"
+          else
+            echo "BENCH_TYPE=[\"${{ inputs.bench_type }}\"]" >> "${GITHUB_ENV}"
+          fi
+
+      - name: Default benchmark type
+        if: github.event_name != 'workflow_dispatch'
+        run: |
+          echo "BENCH_TYPE=[\"latency\"]" >> "${GITHUB_ENV}"
+
+      - name: Set operation flavor output
+        id: set_op_flavor
+        run: |
           echo "op_flavor=${{ toJSON(env.OP_FLAVOR) }}" >> "${GITHUB_OUTPUT}"
 
+      - name: Set benchmark types output
+        id: set_bench_type
+        run: |
+          echo "bench_type=${{ toJSON(env.BENCH_TYPE) }}" >> "${GITHUB_OUTPUT}"
+
   setup-instance:
     name: Setup instance (integer-benchmarks)
     needs: prepare-matrix
@@ -62,7 +90,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -85,11 +113,13 @@ jobs:
       matrix:
         command: [ integer, integer_multi_bit]
         op_flavor: ${{ fromJson(needs.prepare-matrix.outputs.op_flavor) }}
+        bench_type: ${{ fromJSON(needs.prepare-matrix.outputs.bench_type) }}
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -99,18 +129,13 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
       - name: Install rust
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: zama-ai/slab
           path: slab
@@ -123,7 +148,13 @@ jobs:
 
       - name: Run benchmarks with AVX512
         run: |
-          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_${{ matrix.command }}
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} BENCH_TYPE=${{ matrix.bench_type }} bench_${{ matrix.command }}
+
+      # Run these benchmarks only once per benchmark type
+      - name: Run compression benchmarks with AVX512
+        if: matrix.op_flavor == 'default' && matrix.command == 'integer'
+        run: |
+          make BENCH_TYPE=${{ matrix.bench_type }} bench_integer_compression
 
       - name: Parse results
         run: |
@@ -136,32 +167,24 @@ jobs:
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
           --name-suffix avx512 \
-          --throughput
+          --bench-type ${{ matrix.bench_type }}
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
         with:
-          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}_${{ matrix.bench_type }}
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Send data to Slab
         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 }}
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Integer full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -174,7 +197,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -185,7 +208,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (integer-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_shortint.yml

@@ -56,7 +56,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -79,9 +79,10 @@ jobs:
         op_flavor: ${{ fromJson(needs.prepare-matrix.outputs.op_flavor) }}
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -91,18 +92,13 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
       - name: Install rust
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: zama-ai/slab
           path: slab
@@ -124,8 +120,7 @@ jobs:
           --commit-date "${COMMIT_DATE}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
-          --name-suffix avx512 \
-          --throughput
+          --name-suffix avx512
 
       # This small benchmark needs to be executed only once.
       - name: Measure key sizes
@@ -137,11 +132,11 @@ jobs:
         if: matrix.op_flavor == 'default'
         run: |
           python3 ./ci/benchmark_parser.py tfhe/shortint_key_sizes.csv ${{ env.RESULTS_FILENAME }} \
-          --key-sizes \
+          --object-sizes \
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
         with:
           name: ${{ github.sha }}_shortint_${{ matrix.op_flavor }}
           path: ${{ env.RESULTS_FILENAME }}
@@ -149,21 +144,13 @@ jobs:
       - name: Send data to Slab
         shell: bash
         run: |
-          echo "Computing HMac on results file"
-          SIGNATURE="$(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 }}
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Shortint full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -176,7 +163,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -187,7 +174,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (shortint-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_signed_integer.yml

@@ -8,6 +8,14 @@ on:
         description: "Run all precisions"
         type: boolean
         default: false
+      bench_type:
+        description: "Benchmarks type"
+        type: choice
+        default: latency
+        options:
+          - latency
+          - throughput
+          - both
 
   schedule:
     # Weekly benchmarks will be triggered each Saturday at 1a.m.
@@ -36,10 +44,10 @@ jobs:
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
     outputs:
       op_flavor: ${{ steps.set_op_flavor.outputs.op_flavor }}
+      bench_type: ${{ steps.set_bench_type.outputs.bench_type }}
     steps:
       - name: Weekly benchmarks
-        if: github.event_name == 'workflow_dispatch' ||
-          github.event.schedule == '0 1 * * 6'
+        if: github.event.schedule == '0 1 * * 6'
         run: |
           echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
 
@@ -48,11 +56,31 @@ jobs:
         run: |
           echo "OP_FLAVOR=[\"default\", \"unchecked\"]" >> "${GITHUB_ENV}"
 
+      - name: Set benchmark types
+        if: github.event_name == 'workflow_dispatch'
+        run: |
+          echo "OP_FLAVOR=[\"default\"]" >> "${GITHUB_ENV}"
+          if [[ "${{ inputs.bench_type }}" == "both" ]]; then
+            echo "BENCH_TYPE=[\"latency\", \"throughput\"]" >> "${GITHUB_ENV}"
+          else
+            echo "BENCH_TYPE=[\"${{ inputs.bench_type }}\"]" >> "${GITHUB_ENV}"
+          fi
+
+      - name: Default benchmark type
+        if: github.event_name != 'workflow_dispatch'
+        run: |
+          echo "BENCH_TYPE=[\"latency\"]" >> "${GITHUB_ENV}"
+
       - name: Set operation flavor output
         id: set_op_flavor
         run: |
           echo "op_flavor=${{ toJSON(env.OP_FLAVOR) }}" >> "${GITHUB_OUTPUT}"
 
+      - name: Set benchmark types output
+        id: set_bench_type
+        run: |
+          echo "bench_type=${{ toJSON(env.BENCH_TYPE) }}" >> "${GITHUB_OUTPUT}"
+
   setup-instance:
     name: Setup instance (signed-integer-benchmarks)
     needs: prepare-matrix
@@ -62,7 +90,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -84,12 +112,14 @@ jobs:
       max-parallel: 1
       matrix:
         command: [ integer, integer_multi_bit ]
-        op_flavor: [ default, unchecked ]
+        op_flavor: ${{ fromJSON(needs.prepare-matrix.outputs.op_flavor) }}
+        bench_type: ${{ fromJSON(needs.prepare-matrix.outputs.bench_type) }}
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -99,18 +129,13 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
       - name: Install rust
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: zama-ai/slab
           path: slab
@@ -123,7 +148,7 @@ jobs:
 
       - name: Run benchmarks with AVX512
         run: |
-          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} bench_signed_${{ matrix.command }}
+          make BENCH_OP_FLAVOR=${{ matrix.op_flavor }} BENCH_TYPE=${{ matrix.bench_type }} bench_signed_${{ matrix.command }}
 
       - name: Parse results
         run: |
@@ -136,32 +161,24 @@ jobs:
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
           --name-suffix avx512 \
-          --throughput
+          --bench-type ${{ matrix.bench_type }}
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
         with:
-          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}
+          name: ${{ github.sha }}_${{ matrix.command }}_${{ matrix.op_flavor }}_${{ matrix.bench_type }}
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Send data to Slab
         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 }}
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Signed integer full benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -174,7 +191,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -185,7 +202,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (signed-integer-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_tfhe_fft.yml

@@ -0,0 +1,143 @@
+# Run FFT benchmarks on an AWS instance and return parsed results to Slab CI bot.
+name: FFT benchmarks
+
+env:
+  CARGO_TERM_COLOR: always
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
+  RUST_BACKTRACE: "full"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
+on:
+  workflow_dispatch:
+  push:
+    branches:
+      - "main"
+    paths:
+      - tfhe-fft/**
+      - .github/workflows/benchmark_tfhe_fft.yml
+  schedule:
+    # Job will be triggered each Thursday at 11p.m.
+    - cron: '0 23 * * 4'
+
+jobs:
+  setup-ec2:
+    name: Setup EC2 instance (fft-benchmarks)
+    runs-on: ubuntu-latest
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: aws
+          profile: bench
+
+  fft-benchmarks:
+    name: Execute FFT benchmarks in EC2
+    needs: setup-ec2
+    concurrency:
+      group: ${{ github.workflow }}_${{ github.ref }}
+      cancel-in-progress: true
+    runs-on: ${{ needs.setup-ec2.outputs.runner-name }}
+    steps:
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(date --iso-8601=seconds)";
+            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
+            echo "COMMIT_HASH=$(git describe --tags --dirty)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: actions-rs/toolchain@16499b5e05bf2e26879000db0c1d13f7e13fa3af
+        with:
+          toolchain: nightly
+          override: true
+
+      - name: Run benchmarks with AVX512
+        run: |
+          make bench_fft
+
+      - name: Parse AVX512 results
+        run: |
+          python3 ./ci/fft_benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database concrete_fft \
+          --hardware "hpc7a.96xlarge" \
+          --project-version "${{ env.COMMIT_HASH }}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${{ env.COMMIT_DATE }}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --name-suffix avx512
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        with:
+          name: ${{ github.sha }}_fft
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          echo "Computing HMac on downloaded artifact"
+          SIGNATURE="$(slab/scripts/hmac_calculator.sh ${{ env.RESULTS_FILENAME }} '${{ secrets.JOB_SECRET }}')"
+          echo "Sending results to Slab..."
+          curl -v -k \
+          -H "Content-Type: application/json" \
+          -H "X-Slab-Repository: ${{ github.repository }}" \
+          -H "X-Slab-Command: store_data_v2" \
+          -H "X-Hub-Signature-256: sha256=${SIGNATURE}" \
+          -d @${{ env.RESULTS_FILENAME }} \
+          ${{ secrets.SLAB_URL }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "tfhe-fft benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-ec2:
+    name: Teardown EC2 instance (fft-benchmarks)
+    if: ${{ always() && needs.setup-ec2.result != 'skipped' }}
+    needs: [ setup-ec2, fft-benchmarks ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-ec2.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "EC2 teardown (fft-benchmarks) failed. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_tfhe_ntt.yml

@@ -0,0 +1,143 @@
+# Run NTT benchmarks on an AWS instance and return parsed results to Slab CI bot.
+name: NTT benchmarks
+
+env:
+  CARGO_TERM_COLOR: always
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
+  RUST_BACKTRACE: "full"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
+on:
+  workflow_dispatch:
+  push:
+    branches:
+      - "main"
+    paths:
+      - tfhe-ntt/**
+      - .github/workflows/benchmark_tfhe_ntt.yml
+  schedule:
+    # Job will be triggered each Friday at 11p.m.
+    - cron: "0 23 * * 5"
+
+jobs:
+  setup-ec2:
+    name: Setup EC2 instance (ntt-benchmarks)
+    runs-on: ubuntu-latest
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: aws
+          profile: bench
+
+  ntt-benchmarks:
+    name: Execute NTT benchmarks in EC2
+    needs: setup-ec2
+    concurrency:
+      group: ${{ github.workflow }}_${{ github.ref }}
+      cancel-in-progress: true
+    runs-on: ${{ needs.setup-ec2.outputs.runner-name }}
+    steps:
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(date --iso-8601=seconds)";
+            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
+            echo "COMMIT_HASH=$(git describe --tags --dirty)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: actions-rs/toolchain@16499b5e05bf2e26879000db0c1d13f7e13fa3af
+        with:
+          toolchain: nightly
+          override: true
+
+      - name: Run benchmarks
+        run: |
+          make bench_ntt
+
+      - name: Parse results
+        run: |
+          python3 ./ci/ntt_benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database concrete_ntt \
+          --hardware "hpc7a.96xlarge" \
+          --project-version "${{ env.COMMIT_HASH }}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${{ env.COMMIT_DATE }}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --name-suffix avx512
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        with:
+          name: ${{ github.sha }}_ntt
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          echo "Computing HMac on downloaded artifact"
+          SIGNATURE="$(slab/scripts/hmac_calculator.sh ${{ env.RESULTS_FILENAME }} '${{ secrets.JOB_SECRET }}')"
+          echo "Sending results to Slab..."
+          curl -v -k \
+          -H "Content-Type: application/json" \
+          -H "X-Slab-Repository: ${{ github.repository }}" \
+          -H "X-Slab-Command: store_data_v2" \
+          -H "X-Hub-Signature-256: sha256=${SIGNATURE}" \
+          -d @${{ env.RESULTS_FILENAME }} \
+          ${{ secrets.SLAB_URL }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "tfhe-ntt benchmarks failed. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-ec2:
+    name: Teardown EC2 instance (ntt-benchmarks)
+    if: ${{ always() && needs.setup-ec2.result != 'skipped' }}
+    needs: [setup-ec2, ntt-benchmarks]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-ec2.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "EC2 teardown (ntt-benchmarks) failed. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_tfhe_zk_pok.yml

@@ -0,0 +1,172 @@
+# Run benchmarks of the tfhe-zk-pok crate on an instance and return parsed results to Slab CI bot.
+name: tfhe-zk-pok benchmarks
+
+on:
+  workflow_dispatch:
+  push:
+    branches:
+      - main
+  schedule:
+    # Weekly benchmarks will be triggered each Saturday at 3a.m.
+    - cron: '0 3 * * 6'
+env:
+  CARGO_TERM_COLOR: always
+  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
+  PARSE_INTEGER_BENCH_CSV_FILE: tfhe_rs_integer_benches_${{ github.sha }}.csv
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
+jobs:
+  should-run:
+    runs-on: ubuntu-latest
+    if: github.event_name == 'workflow_dispatch' ||
+      ((github.event_name == 'push' || github.event_name == 'schedule') && github.repository == 'zama-ai/tfhe-rs')
+    outputs:
+      zk_pok_changed: ${{ steps.changed-files.outputs.zk_pok_any_changed }}
+    steps:
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+
+      - name: Check for file changes
+        id: changed-files
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
+        with:
+          since_last_remote_commit: true
+          files_yaml: |
+            zk_pok:
+              - tfhe-zk-pok/**
+              - .github/workflows/benchmark_tfhe_zk_pok.yml
+
+  setup-instance:
+    name: Setup instance (tfhe-zk-pok-benchmarks)
+    runs-on: ubuntu-latest
+    needs: should-run
+    if: github.event_name == 'workflow_dispatch' ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
+      (github.event_name == 'push' &&
+      github.repository == 'zama-ai/tfhe-rs' &&
+      needs.should-run.outputs.zk_pok_changed == 'true')
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: aws
+          profile: bench
+
+  tfhe-zk-pok-benchmarks:
+    name: Execute tfhe-zk-pok benchmarks
+    if: needs.setup-instance.result != 'skipped'
+    needs: setup-instance
+    concurrency:
+      group: ${{ github.workflow }}_${{github.event_name}}_${{ github.ref }}${{ github.ref == 'refs/heads/main' && github.sha || '' }}
+      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    steps:
+      - name: Checkout tfhe-rs repo with tags
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Get benchmark details
+        run: |
+          {
+            echo "BENCH_DATE=$(date --iso-8601=seconds)";
+            echo "COMMIT_DATE=$(git --no-pager show -s --format=%cd --date=iso8601-strict ${{ github.sha }})";
+            echo "COMMIT_HASH=$(git describe --tags --dirty)";
+          } >> "${GITHUB_ENV}"
+
+      - name: Install rust
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
+        with:
+          toolchain: nightly
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Run benchmarks
+        run: |
+          make bench_tfhe_zk_pok
+
+      - name: Parse results
+        run: |
+          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
+          --database tfhe_rs \
+          --crate tfhe-zk-pok \
+          --hardware "hpc7a.96xlarge" \
+          --backend cpu \
+          --project-version "${{ env.COMMIT_HASH }}" \
+          --branch ${{ github.ref_name }} \
+          --commit-date "${{ env.COMMIT_DATE }}" \
+          --bench-date "${{ env.BENCH_DATE }}" \
+          --walk-subdirs \
+          --name-suffix avx512
+
+      - name: Upload parsed results artifact
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
+        with:
+          name: ${{ github.sha }}_tfhe_zk_pok
+          path: ${{ env.RESULTS_FILENAME }}
+
+      - name: Checkout Slab repo
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          repository: zama-ai/slab
+          path: slab
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Send data to Slab
+        shell: bash
+        run: |
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "tfhe-zk-pok benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (tfhe-zk-pok-benchmarks)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, tfhe-zk-pok-benchmarks ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (tfhe-zk-pok-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_wasm_client.yml

@@ -33,19 +33,20 @@ jobs:
       wasm_bench: ${{ steps.changed-files.outputs.wasm_bench_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@6b2903bdce6310cfbddd87c418f253cf29b2dec9
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
         with:
           since_last_remote_commit: true
           files_yaml: |
             wasm_bench:
               - tfhe/Cargo.toml
-              - concrete-csprng/**
+              - tfhe-csprng/**
               - tfhe-zk-pok/**
               - tfhe/src/**
               - '!tfhe/src/c_api/**'
@@ -64,7 +65,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -78,11 +79,16 @@ jobs:
     needs: setup-instance
     if: needs.setup-instance.result != 'skipped'
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    strategy:
+      max-parallel: 1
+      matrix:
+        browser: [ chrome, firefox ]
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -92,20 +98,46 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
       - name: Install rust
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: nightly
 
-      - name: Run benchmarks
+      - name: Get Node version
+        run: |
+          echo "NODE_VERSION=$(make node_version)" >> "${GITHUB_ENV}"
+
+      - name: Node cache restoration
+        id: node-cache
+        uses: actions/cache/restore@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
+        with:
+          path: |
+            ~/.nvm
+            ~/.npm
+          key: node-${{ env.NODE_VERSION }}
+
+      - name: Install Node
+        if: steps.node-cache.outputs.cache-hit != 'true'
         run: |
           make install_node
-          make bench_web_js_api_parallel_ci
+
+      - name: Node cache save
+        uses: actions/cache/save@1bd1e32a3bdc45362d1e726936510720a7c30a57 #v4.2.0
+        if: steps.node-cache.outputs.cache-hit != 'true'
+        with:
+          path: |
+            ~/.nvm
+            ~/.npm
+          key: node-${{ env.NODE_VERSION }}
+
+      - name: Install web resources
+        run: |
+          make install_${{ matrix.browser }}_browser
+          make install_${{ matrix.browser }}_web_driver
+
+      - name: Run benchmarks
+        run: |
+          make bench_web_js_api_parallel_${{ matrix.browser }}_ci
 
       - name: Parse results
         run: |
@@ -118,25 +150,29 @@ jobs:
           --commit-date "${{ env.COMMIT_DATE }}" \
           --bench-date "${{ env.BENCH_DATE }}" \
           --key-gen
+          rm tfhe/wasm_pk_gen.csv
 
+      # Run these benchmarks only once
       - name: Measure public key and ciphertext sizes in HL Api
+        if:  matrix.browser == 'chrome'
         run: |
           make measure_hlapi_compact_pk_ct_sizes
 
       - name: Parse key and ciphertext sizes results
+        if:  matrix.browser == 'chrome'
         run: |
           python3 ./ci/benchmark_parser.py tfhe/hlapi_cpk_and_cctl_sizes.csv ${{ env.RESULTS_FILENAME }} \
           --key-gen \
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
         with:
-          name: ${{ github.sha }}_wasm
+          name: ${{ github.sha }}_wasm_${{ matrix.browser }}
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: zama-ai/slab
           path: slab
@@ -145,24 +181,16 @@ jobs:
       - 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 }}
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "WASM benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "WASM benchmarks (${{ matrix.browser }}) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (wasm-client-benchmarks)
@@ -172,7 +200,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -183,7 +211,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (wasm-client-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/benchmark_zk_pke.yml

@@ -3,6 +3,16 @@ name: PKE ZK benchmarks
 
 on:
   workflow_dispatch:
+    inputs:
+      bench_type:
+        description: "Benchmarks type"
+        type: choice
+        default: latency
+        options:
+          - latency
+          - throughput
+          - both
+
   push:
     branches:
       - main
@@ -30,19 +40,21 @@ jobs:
       zk_pok_changed: ${{ steps.changed-files.outputs.zk_pok_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@6b2903bdce6310cfbddd87c418f253cf29b2dec9
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
         with:
           since_last_remote_commit: true
           files_yaml: |
             zk_pok:
               - tfhe/Cargo.toml
-              - concrete-csprng/**
+              - tfhe-csprng/**
+              - tfhe-fft/**
               - tfhe-zk-pok/**
               - tfhe/src/core_crypto/**
               - tfhe/src/shortint/**
@@ -51,10 +63,37 @@ jobs:
               - tfhe/benches/integer/zk_pke.rs
               - .github/workflows/zk_pke_benchmark.yml
 
+  prepare-matrix:
+    name: Prepare operations matrix
+    runs-on: ubuntu-latest
+    if: github.event_name != 'schedule' ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
+    outputs:
+      bench_type: ${{ steps.set_bench_type.outputs.bench_type }}
+    steps:
+      - name: Set benchmark types
+        if: github.event_name == 'workflow_dispatch'
+        run: |
+          if [[ "${{ inputs.bench_type }}" == "both" ]]; then
+            echo "BENCH_TYPE=[\"latency\", \"throughput\"]" >> "${GITHUB_ENV}"
+          else
+            echo "BENCH_TYPE=[\"${{ inputs.bench_type }}\"]" >> "${GITHUB_ENV}"
+          fi
+
+      - name: Default benchmark type
+        if: github.event_name != 'workflow_dispatch'
+        run: |
+          echo "BENCH_TYPE=[\"latency\"]" >> "${GITHUB_ENV}"
+
+      - name: Set benchmark types output
+        id: set_bench_type
+        run: |
+          echo "bench_type=${{ toJSON(env.BENCH_TYPE) }}" >> "${GITHUB_OUTPUT}"
+
   setup-instance:
     name: Setup instance (pke-zk-benchmarks)
     runs-on: ubuntu-latest
-    needs: should-run
+    needs: [ should-run, prepare-matrix ]
     if: github.event_name == 'workflow_dispatch' ||
       (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs') ||
       (github.event_name == 'push' &&
@@ -65,7 +104,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -77,16 +116,21 @@ jobs:
   pke-zk-benchmarks:
     name: Execute PKE ZK benchmarks
     if: needs.setup-instance.result != 'skipped'
-    needs: setup-instance
+    needs: [ prepare-matrix, setup-instance ]
     concurrency:
-      group: ${{ github.workflow }}_${{github.event_name}}_${{ github.ref }}
+      group: ${{ github.workflow }}_${{github.event_name}}_${{ github.ref }}${{ github.ref == 'refs/heads/main' && github.sha || '' }}
       cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    strategy:
+      max-parallel: 1
+      matrix:
+        bench_type: ${{ fromJSON(needs.prepare-matrix.outputs.bench_type) }}
     steps:
       - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Get benchmark details
         run: |
@@ -96,18 +140,13 @@ jobs:
             echo "COMMIT_HASH=$(git describe --tags --dirty)";
           } >> "${GITHUB_ENV}"
 
-      - name: Set up home
-        # "Install rust" step require root user to have a HOME directory which is not set.
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
-
       - name: Install rust
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: nightly
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: zama-ai/slab
           path: slab
@@ -115,7 +154,7 @@ jobs:
 
       - name: Run benchmarks with AVX512
         run: |
-          make bench_integer_zk
+          make BENCH_TYPE=${{ matrix.bench_type }} bench_integer_zk
 
       - name: Parse results
         run: |
@@ -129,22 +168,22 @@ jobs:
           --bench-date "${{ env.BENCH_DATE }}" \
           --walk-subdirs \
           --name-suffix avx512 \
-          --throughput
+          --bench-type ${{ matrix.bench_type }}
 
       - name: Parse CRS sizes results
         run: |
           python3 ./ci/benchmark_parser.py tfhe/pke_zk_crs_sizes.csv ${{ env.RESULTS_FILENAME }} \
-          --key-sizes \
+          --object-sizes \
           --append-results
 
       - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
+        uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882
         with:
           name: ${{ github.sha }}_integer_zk
           path: ${{ env.RESULTS_FILENAME }}
 
       - name: Checkout Slab repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: zama-ai/slab
           path: slab
@@ -153,21 +192,13 @@ jobs:
       - 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 }}
+          python3 slab/scripts/data_sender.py ${{ env.RESULTS_FILENAME }} "${{ secrets.JOB_SECRET }}" \
+          --slab-url "${{ secrets.SLAB_URL }}"
 
       - name: Slack Notification
-        if: ${{ !success() && !cancelled() }}
+        if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "PKE ZK benchmarks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -180,7 +211,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -191,7 +222,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (pke-zk-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/cargo_build.yml

@@ -1,4 +1,4 @@
-name: Cargo Build
+name: Cargo Build TFHE-rs
 
 on:
   pull_request:
@@ -19,14 +19,21 @@ jobs:
 
     strategy:
       matrix:
-        os: [large_ubuntu_16, macos-latest-large, large_windows_16_latest]
+        # GitHub macos-latest are now M1 macs, so use ours, we limit what runs so it will be fast
+        # even with a few PRs
+        os: [large_ubuntu_16, macos-latest, windows-latest]
       fail-fast: false
 
     steps:
-      - uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+
+      - name: Install latest stable
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
+        with:
+          toolchain: stable
 
       - name: Install and run newline linter checks
-        if: matrix.os == 'ubuntu-latest'
+        if: ${{ contains(matrix.os, 'ubuntu') }}
         run: |
           wget https://github.com/fernandrone/linelint/releases/download/0.0.6/linelint-linux-amd64
           echo "16b70fb7b471d6f95cbdc0b4e5dc2b0ac9e84ba9ecdc488f7bdf13df823aca4b linelint-linux-amd64" > checksum
@@ -36,27 +43,33 @@ jobs:
           make check_newline
 
       - name: Run pcc checks
+        if: ${{ contains(matrix.os, 'ubuntu') }}
         run: |
           make pcc
 
-      - name: Build concrete-csprng
+      - name: Build tfhe-csprng
+        if: ${{ contains(matrix.os, 'ubuntu') }}
         run: |
-          make build_concrete_csprng
+          make build_tfhe_csprng
 
       - name: Build Release core
+        if: ${{ contains(matrix.os, 'ubuntu') }}
         run: |
           make build_core AVX512_SUPPORT=ON
           make build_core_experimental AVX512_SUPPORT=ON
 
       - name: Build Release boolean
+        if: ${{ contains(matrix.os, 'ubuntu') }}
         run: |
           make build_boolean
 
       - name: Build Release shortint
+        if: ${{ contains(matrix.os, 'ubuntu') }}
         run: |
           make build_shortint
 
       - name: Build Release integer
+        if: ${{ contains(matrix.os, 'ubuntu') }}
         run: |
           make build_integer
 
@@ -65,10 +78,12 @@ jobs:
           make build_tfhe_full
 
       - name: Build Release c_api
+        if: ${{ contains(matrix.os, 'ubuntu') }}
         run: |
           make build_c_api
 
       - name: Build coverage tests
+        if: ${{ contains(matrix.os, 'ubuntu') }}
         run: |
           make build_tfhe_coverage
 

.github/workflows/cargo_build_tfhe_fft.yml

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

.github/workflows/cargo_build_tfhe_ntt.yml

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

.github/workflows/cargo_test_fft.yml

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

.github/workflows/cargo_test_ntt.yml

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

.github/workflows/check_commit.yml

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

.github/workflows/ci_lint.yml

@@ -13,7 +13,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
 
       - name: Get actionlint
         run: |
@@ -25,3 +25,9 @@ jobs:
       - name: Lint workflows
         run: |
           make lint_workflow
+
+      - name: Ensure SHA pinned actions
+        uses: zgosalvez/github-actions-ensure-sha-pinned-actions@5d6ac37a4cef8b8df67f482a8e384987766f0213 # v3.0.17
+        with:
+          allowlist: |
+            slsa-framework/slsa-github-generator

.github/workflows/code_coverage.yml

@@ -25,7 +25,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -44,26 +44,22 @@ jobs:
     timeout-minutes: 5760 # 4 days
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@6b2903bdce6310cfbddd87c418f253cf29b2dec9
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
         with:
           files_yaml: |
             tfhe:
               - tfhe/src/**
-            concrete_csprng:
-              - concrete-csprng/src/**
+            tfhe_csprng:
+              - tfhe-csprng/src/**
 
       - name: Generate Keys
         if: steps.changed-files.outputs.tfhe_any_changed == 'true'
@@ -87,7 +83,7 @@ jobs:
           make test_shortint_cov
 
       - name: Upload tfhe coverage to Codecov
-        uses: codecov/codecov-action@e28ff129e5465c2c0dcc6f003fc735cb6ae0c673
+        uses: codecov/codecov-action@7f8b4b4bde536c465e797be725718b88c5d95e0e
         if: steps.changed-files.outputs.tfhe_any_changed == 'true'
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
@@ -101,7 +97,7 @@ jobs:
           make test_integer_cov
 
       - name: Upload tfhe coverage to Codecov
-        uses: codecov/codecov-action@e28ff129e5465c2c0dcc6f003fc735cb6ae0c673
+        uses: codecov/codecov-action@7f8b4b4bde536c465e797be725718b88c5d95e0e
         if: steps.changed-files.outputs.tfhe_any_changed == 'true'
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
@@ -112,7 +108,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Code coverage finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -125,7 +121,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -136,7 +132,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (code-coverage) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/csprng_randomness_tests.yml

@@ -27,7 +27,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -45,16 +45,13 @@ jobs:
     runs-on: ${{ needs.setup-instance.outputs.runner-name }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
-
-      - name: Set up home
-        run: |
-          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
@@ -65,10 +62,10 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "concrete-csprng randomness check finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "tfhe-csprng randomness check finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
 
   teardown-instance:
     name: Teardown instance (csprng-randomness-tests)
@@ -78,7 +75,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -89,7 +86,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (csprng-randomness-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/data_pr_close.yml

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

.github/workflows/gpu_4090_tests.yml

@@ -34,12 +34,13 @@ jobs:
 
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
@@ -74,9 +75,9 @@ jobs:
           github_token: ${{ secrets.GITHUB_TOKEN }}
 
       - name: Slack Notification
-        if: ${{ always() }}
+        if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "CUDA RTX 4090 tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_fast_h100_tests.yml

@@ -0,0 +1,204 @@
+# 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@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Check for file changes
+        id: changed-files
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
+        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/src/integer/server_key/radix_parallel/tests_unsigned/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
+              - tfhe/src/shortint/parameters/**
+              - tfhe/src/high_level_api/**
+              - tfhe/src/c_api/**
+              - 'tfhe/docs/**.md'
+              - '.github/workflows/gpu_fast_h100_tests.yml'
+              - scripts/integer-tests.sh
+              - ci/slab.toml
+
+  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@98f0788261a7323d5d695a883e20df36591a92b7
+        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 libclang-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          persist-credentials: 'false'
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install latest stable
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
+        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_integer_compression_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' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
+          SLACK_MESSAGE: "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@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_fast_tests.yml

@@ -27,13 +27,14 @@ jobs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@6b2903bdce6310cfbddd87c418f253cf29b2dec9
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
         with:
           since_last_remote_commit: true
           files_yaml: |
@@ -42,11 +43,17 @@ jobs:
               - tfhe/build.rs
               - backends/tfhe-cuda-backend/**
               - tfhe/src/core_crypto/gpu/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
               - tfhe/src/integer/gpu/**
-              - tfhe/shortint/parameters/**
+              - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
               - 'tfhe/docs/**.md'
+              - '.github/workflows/gpu_fast_tests.yml'
+              - scripts/integer-tests.sh
+              - ci/slab.toml
 
   setup-instance:
     name: Setup instance (cuda-tests)
@@ -59,13 +66,13 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
           slab-url: ${{ secrets.SLAB_BASE_URL }}
           job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
+          backend: hyperstack
           profile: gpu-test
 
   cuda-tests-linux:
@@ -84,22 +91,35 @@ jobs:
         include:
           - os: ubuntu-22.04
             cuda: "12.2"
-            gcc: 9
+            gcc: 11 
     env:
       CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-
+      CMAKE_VERSION: 3.29.6
     steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev libclang-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
@@ -122,9 +142,14 @@ jobs:
             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_integer_compression_gpu
           make test_cuda_backend
 
       - name: Run user docs tests
@@ -139,13 +164,18 @@ jobs:
         run: |
           make test_high_level_api_gpu
 
-      - name: Slack Notification
-        if: ${{ always() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-tests-linux ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-tests-linux.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "CUDA AWS tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          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)
@@ -155,7 +185,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -166,7 +196,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_full_h100_tests.yml

@@ -0,0 +1,156 @@
+# Compile and test tfhe-cuda-backend on an H100 VM on hyperstack
+name: TFHE Cuda Backend - Full 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:
+  workflow_dispatch:
+
+jobs:
+  setup-instance:
+    name: Setup instance (cuda-h100-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@98f0788261a7323d5d695a883e20df36591a92b7
+        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: [ setup-instance ]
+    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 libclang-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          persist-credentials: 'false'
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install latest stable
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
+        with:
+          toolchain: stable
+
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
+          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "HOME=/home/ubuntu";
+          } >> "${GITHUB_ENV}"
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
+      - name: Run core crypto, integer and internal CUDA backend tests
+        run: |
+          make test_gpu
+
+      - 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: ${{ failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
+          SLACK_MESSAGE: "Full H100 tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (cuda-h100-tests)
+    needs: [ setup-instance, cuda-tests-linux ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_full_multi_gpu_tests.yml

@@ -28,13 +28,14 @@ jobs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@6b2903bdce6310cfbddd87c418f253cf29b2dec9
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
         with:
           since_last_remote_commit: true
           files_yaml: |
@@ -43,11 +44,17 @@ jobs:
               - tfhe/build.rs
               - backends/tfhe-cuda-backend/**
               - tfhe/src/core_crypto/gpu/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
               - tfhe/src/integer/gpu/**
-              - tfhe/shortint/parameters/**
+              - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
               - 'tfhe/docs/**.md'
+              - '.github/workflows/**_multi_gpu_tests.yml'
+              - scripts/integer-tests.sh
+              - ci/slab.toml
 
   setup-instance:
     name: Setup instance (cuda-tests-multi-gpu)
@@ -61,13 +68,13 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
           slab-url: ${{ secrets.SLAB_BASE_URL }}
           job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
+          backend: hyperstack
           profile: multi-gpu-test
 
   cuda-tests-linux:
@@ -86,20 +93,35 @@ jobs:
         include:
           - os: ubuntu-22.04
             cuda: "12.2"
-            gcc: 9
+            gcc: 11 
     env:
       CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-
+      CMAKE_VERSION: 3.29.6
     steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev libclang-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          persist-credentials: 'false'
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
@@ -122,30 +144,43 @@ jobs:
             echo "HOME=/home/ubuntu";
           } >> "${GITHUB_ENV}"
 
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
+      - name: Run multi-bit CUDA integer compression tests
+        run: |
+          BIG_TESTS_INSTANCE=TRUE make test_integer_compression_gpu
+
       # 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: |
-          make test_integer_multi_bit_gpu_ci
+          BIG_TESTS_INSTANCE=TRUE make test_integer_multi_bit_gpu_ci
 
       - name: Run user docs tests
         run: |
-          make test_user_doc_gpu
+          BIG_TESTS_INSTANCE=TRUE make test_user_doc_gpu
 
       - name: Test C API
         run: |
-          make test_c_api_gpu
+          BIG_TESTS_INSTANCE=TRUE make test_c_api_gpu
 
       - name: Run High Level API Tests
         run: |
-          make test_high_level_api_gpu
+          BIG_TESTS_INSTANCE=TRUE make test_high_level_api_gpu
 
-      - name: Slack Notification
-        if: ${{ always() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-tests-linux ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-tests-linux.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "CUDA AWS multi-GPU tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          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)
@@ -155,7 +190,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -166,7 +201,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-tests-multi-gpu) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_integer_long_run_tests.yml

@@ -0,0 +1,146 @@
+name: AWS Long Run Tests on GPU
+
+env:
+  CARGO_TERM_COLOR: always
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUSTFLAGS: "-C target-cpu=native"
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
+on:
+  # Allows you to run this workflow manually from the Actions tab as an alternative.
+  workflow_dispatch:
+  schedule:
+    # Weekly tests will be triggered each Friday at 1a.m.
+    - cron: '0 1 * * FRI'
+
+jobs:
+  setup-instance:
+    name: Setup instance (gpu-tests)
+    if: github.event_name != 'schedule' ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
+    runs-on: ubuntu-latest
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: hyperstack
+          profile: 2-h100
+
+  cuda-tests:
+    name: Long run GPU H100 tests
+    needs: [ setup-instance ]
+    concurrency:
+      group: ${{ github.workflow }}_${{github.event_name}}_${{ github.ref }}
+      cancel-in-progress: true
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    strategy:
+      fail-fast: false
+      # explicit include-based build matrix, of known valid options
+      matrix:
+        include:
+          - os: ubuntu-22.04
+            cuda: "12.2"
+            gcc: 11 
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
+    steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev libclang-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install latest stable
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
+        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 tests
+        run: |
+          make test_integer_long_run_gpu
+
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-tests ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-tests.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ needs.cuda-tests.result }}
+          SLACK_MESSAGE: "Integer GPU H100 long run tests finished with status: ${{ needs.cuda-tests.result }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (gpu-tests)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, cuda-tests ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (gpu-long-run-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_pcc.yml

@@ -24,7 +24,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -53,16 +53,17 @@ jobs:
 
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           persist-credentials: 'false'
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
@@ -94,9 +95,9 @@ jobs:
           make pcc_gpu
 
       - name: Slack Notification
-        if: ${{ always() }}
+        if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "CUDA AWS post-commit checks finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -109,7 +110,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -120,7 +121,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-pcc) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_signed_integer_classic_tests.yml

@@ -0,0 +1,188 @@
+# Signed integer GPU tests on an RTXA6000 VM on hyperstack with classical PBS
+name: TFHE Cuda Backend - Signed integer tests with classical PBS
+
+env:
+  CARGO_TERM_COLOR: always
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUSTFLAGS: "-C target-cpu=native"
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' }}
+
+on:
+  # Allows you to run this workflow manually from the Actions tab as an alternative.
+  workflow_dispatch:
+  pull_request:
+      types: [ labeled ]
+
+jobs:
+  should-run:
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: write
+    outputs:
+      gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
+    steps:
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Check for file changes
+        id: changed-files
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
+        with:
+          since_last_remote_commit: true
+          files_yaml: |
+            gpu:
+              - tfhe/Cargo.toml
+              - tfhe/build.rs
+              - backends/tfhe-cuda-backend/**
+              - tfhe/src/core_crypto/gpu/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
+              - tfhe/src/integer/gpu/**
+              - tfhe/src/shortint/parameters/**
+              - tfhe/src/high_level_api/**
+              - tfhe/src/c_api/**
+              - 'tfhe/docs/**.md'
+              - '.github/workflows/gpu_signed_integer_classic_tests.yml'
+              - scripts/integer-tests.sh
+              - ci/slab.toml
+
+  setup-instance:
+    name: Setup instance (cuda-signed-classic-tests)
+    needs: should-run
+    if: github.event_name != 'pull_request' ||
+      (github.event.action != 'labeled' && needs.should-run.outputs.gpu_test == 'true') ||
+      (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.gpu_test == 'true')
+    runs-on: ubuntu-latest
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: hyperstack
+          profile: gpu-test
+
+  cuda-tests-linux:
+    name: CUDA signed integer tests with classical PBS
+    needs: [ should-run, setup-instance ]
+    if: github.event_name != 'pull_request' ||
+      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
+    concurrency:
+      group: ${{ github.workflow }}_${{ github.ref }}
+      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    strategy:
+      fail-fast: false
+      # explicit include-based build matrix, of known valid options
+      matrix:
+        include:
+          - os: ubuntu-22.04
+            cuda: "12.2"
+            gcc: 11 
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
+    steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev libclang-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install latest stable
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
+        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
+
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-tests-linux ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-tests-linux.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
+          SLACK_MESSAGE: "Integer GPU signed integer tests with classical PBS finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (cuda-signed-classic-tests)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, cuda-tests-linux ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cuda-signed-classic-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_signed_integer_h100_tests.yml

@@ -1,5 +1,5 @@
-# Compile and test tfhe-cuda-backend on an H100 VM on hyperstack
-name: TFHE Cuda Backend - Full tests on H100
+# Signed integer GPU tests on an H100 VM on hyperstack
+name: TFHE Cuda Backend - Signed integer tests on H100
 
 env:
   CARGO_TERM_COLOR: always
@@ -28,13 +28,14 @@ jobs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@6b2903bdce6310cfbddd87c418f253cf29b2dec9
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
         with:
           since_last_remote_commit: true
           files_yaml: |
@@ -43,11 +44,17 @@ jobs:
               - tfhe/build.rs
               - backends/tfhe-cuda-backend/**
               - tfhe/src/core_crypto/gpu/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
               - tfhe/src/integer/gpu/**
-              - tfhe/shortint/parameters/**
+              - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
               - 'tfhe/docs/**.md'
+              - '.github/workflows/gpu_signed_integer_h100_tests.yml'
+              - scripts/integer-tests.sh
+              - ci/slab.toml
 
   setup-instance:
     name: Setup instance (cuda-h100-tests)
@@ -61,7 +68,7 @@ jobs:
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -71,7 +78,7 @@ jobs:
           profile: single-h100
 
   cuda-tests-linux:
-    name: CUDA H100 tests
+    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')
@@ -95,7 +102,7 @@ jobs:
       - name: Install dependencies
         run: |
           sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev 
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev libclang-dev
           wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
           tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
           cd cmake-${{ env.CMAKE_VERSION }}
@@ -103,15 +110,16 @@ jobs:
           make -j"$(nproc)"
           sudo make install
 
+
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
@@ -138,40 +146,19 @@ jobs:
         if: ${{ !cancelled() }}
         run: nvidia-smi
 
-      - name: Run core crypto and internal CUDA backend tests
+      - name: Run signed integer multi-bit tests
         run: |
-          make test_core_crypto_gpu
-          make test_cuda_backend
-
-      - name: Run integer tests
-        run: |
-          make test_integer_gpu_ci
-
-      - name: Run integer multi-bit tests
-        run: |
-          make test_integer_multi_bit_gpu_ci
-
-      - 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
+          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' }}
+    if: ${{ always() && needs.cuda-tests-linux.result != 'skipped' && failure() }}
     continue-on-error: true
     steps:
       - name: Send message
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
           SLACK_MESSAGE: "Integer GPU H100 tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
@@ -184,7 +171,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -195,7 +182,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_signed_integer_tests.yml

@@ -22,7 +22,6 @@ on:
     types:
       - opened
       - synchronize
-      - labeled
   schedule:
     # Nightly tests @ 1AM after each work day
     - cron: "0 1 * * MON-FRI"
@@ -36,13 +35,14 @@ jobs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@6b2903bdce6310cfbddd87c418f253cf29b2dec9
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
         with:
           since_last_remote_commit: true
           files_yaml: |
@@ -51,11 +51,17 @@ jobs:
               - tfhe/build.rs
               - backends/tfhe-cuda-backend/**
               - tfhe/src/core_crypto/gpu/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
               - tfhe/src/integer/gpu/**
-              - tfhe/shortint/parameters/**
+              - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
               - 'tfhe/docs/**.md'
+              - '.github/workflows/gpu_signed_integer_tests.yml'
+              - scripts/integer-tests.sh
+              - ci/slab.toml
 
   setup-instance:
     name: Setup instance (cuda-signed-integer-tests)
@@ -63,20 +69,19 @@ jobs:
     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') ||
-      (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.gpu_test == 'true')
+      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
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
           slab-url: ${{ secrets.SLAB_BASE_URL }}
           job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
+          backend: hyperstack
           profile: gpu-test
 
   cuda-signed-integer-tests:
@@ -95,20 +100,36 @@ jobs:
         include:
           - os: ubuntu-22.04
             cuda: "12.2"
-            gcc: 9
+            gcc: 11 
     env:
       CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-
+      CMAKE_VERSION: 3.29.6
     steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev libclang-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          persist-credentials: 'false'
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
@@ -139,22 +160,26 @@ jobs:
             echo "NIGHTLY_TESTS=TRUE";
           } >> "${GITHUB_ENV}"
 
-      - name: Run signed integer tests
-        if: github.event_name != 'pull_request' || contains(github.event.label.name, 'approved')
-        run: |
-          make test_signed_integer_gpu_ci
+      - 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
 
-      - name: Slack Notification
-        if: ${{ always() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+  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' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "CUDA AWS signed integer tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          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)
@@ -164,7 +189,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -175,7 +200,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-signed-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_unsigned_integer_classic_tests.yml

@@ -0,0 +1,188 @@
+# Test unsigned integers on an RTXA6000 VM on hyperstack with the classical PBS
+name: TFHE Cuda Backend - Unsigned integer tests with classical PBS
+
+env:
+  CARGO_TERM_COLOR: always
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUSTFLAGS: "-C target-cpu=native"
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+  IS_PULL_REQUEST: ${{ github.event_name == 'pull_request' }}
+
+on:
+  # Allows you to run this workflow manually from the Actions tab as an alternative.
+  workflow_dispatch:
+  pull_request:
+      types: [ labeled ]
+
+jobs:
+  should-run:
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: write
+    outputs:
+      gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
+    steps:
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Check for file changes
+        id: changed-files
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
+        with:
+          since_last_remote_commit: true
+          files_yaml: |
+            gpu:
+              - tfhe/Cargo.toml
+              - tfhe/build.rs
+              - backends/tfhe-cuda-backend/**
+              - tfhe/src/core_crypto/gpu/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
+              - tfhe/src/integer/gpu/**
+              - tfhe/src/shortint/parameters/**
+              - tfhe/src/high_level_api/**
+              - tfhe/src/c_api/**
+              - 'tfhe/docs/**.md'
+              - '.github/workflows/gpu_unsigned_integer_classic_tests.yml'
+              - scripts/integer-tests.sh
+              - ci/slab.toml
+
+  setup-instance:
+    name: Setup instance (cuda-unsigned-classic-tests)
+    needs: should-run
+    if: github.event_name != 'pull_request' ||
+      (github.event.action != 'labeled' && needs.should-run.outputs.gpu_test == 'true') ||
+      (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.gpu_test == 'true')
+    runs-on: ubuntu-latest
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: hyperstack
+          profile: gpu-test
+
+  cuda-tests-linux:
+    name: CUDA unsigned integer tests with classical PBS
+    needs: [ should-run, setup-instance ]
+    if: github.event_name != 'pull_request' ||
+      (github.event_name == 'pull_request' && needs.setup-instance.result != 'skipped')
+    concurrency:
+      group: ${{ github.workflow }}_${{ github.ref }}
+      cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    strategy:
+      fail-fast: false
+      # explicit include-based build matrix, of known valid options
+      matrix:
+        include:
+          - os: ubuntu-22.04
+            cuda: "12.2"
+            gcc: 11 
+    env:
+      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
+      CMAKE_VERSION: 3.29.6
+    steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev libclang-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install latest stable
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
+        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
+
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-tests-linux ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-tests-linux.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
+          SLACK_MESSAGE: "Unsigned integer GPU classic tests finished with status: ${{ needs.cuda-tests-linux.result }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (cuda-unsigned-classic-tests)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, cuda-tests-linux ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cuda-unsigned-classic-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_unsigned_integer_h100_tests.yml

@@ -0,0 +1,188 @@
+# 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@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Check for file changes
+        id: changed-files
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
+        with:
+          since_last_remote_commit: true
+          files_yaml: |
+            gpu:
+              - tfhe/Cargo.toml
+              - tfhe/build.rs
+              - backends/tfhe-cuda-backend/**
+              - tfhe/src/core_crypto/gpu/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
+              - tfhe/src/integer/gpu/**
+              - tfhe/src/shortint/parameters/**
+              - tfhe/src/high_level_api/**
+              - tfhe/src/c_api/**
+              - 'tfhe/docs/**.md'
+              - '.github/workflows/gpu_unsigned_integer_h100_tests.yml'
+              - scripts/integer-tests.sh
+              - ci/slab.toml
+
+  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@98f0788261a7323d5d695a883e20df36591a92b7
+        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 libclang-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+
+      - name: Set up home
+        run: |
+          echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
+
+      - name: Install latest stable
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
+        with:
+          toolchain: stable
+
+      - name: Export CUDA variables
+        if: ${{ !cancelled() }}
+        run: |
+          echo "CUDA_PATH=$CUDA_PATH" >> "${GITHUB_ENV}"
+          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
+          echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH" >> "${GITHUB_ENV}"
+          echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc" >> "${GITHUB_ENV}"
+
+      # Specify the correct host compilers
+      - name: Export gcc and g++ variables
+        if: ${{ !cancelled() }}
+        run: |
+          {
+            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
+            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
+            echo "HOME=/home/ubuntu";
+          } >> "${GITHUB_ENV}"
+
+      - name: Check device is detected
+        if: ${{ !cancelled() }}
+        run: nvidia-smi
+
+      - name: Run unsigned integer multi-bit tests
+        run: |
+          BIG_TESTS_INSTANCE=TRUE make test_unsigned_integer_multi_bit_gpu_ci
+
+  slack-notify:
+    name: Slack Notification
+    needs: [ setup-instance, cuda-tests-linux ]
+    runs-on: ubuntu-latest
+    if: ${{ always() && needs.cuda-tests-linux.result != 'skipped' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ needs.cuda-tests-linux.result }}
+          SLACK_MESSAGE: "Unsigned integer GPU 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@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cuda-h100-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/gpu_unsigned_integer_tests.yml

@@ -21,7 +21,6 @@ on:
     types:
       - opened
       - synchronize
-      - labeled
   schedule:
     # Nightly tests @ 1AM after each work day
     - cron: "0 1 * * MON-FRI"
@@ -35,13 +34,14 @@ jobs:
       gpu_test: ${{ env.IS_PULL_REQUEST == 'false' || steps.changed-files.outputs.gpu_any_changed }}
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Check for file changes
         id: changed-files
-        uses: tj-actions/changed-files@6b2903bdce6310cfbddd87c418f253cf29b2dec9
+        uses: tj-actions/changed-files@bab30c2299617f6615ec02a68b9a40d10bd21366
         with:
           since_last_remote_commit: true
           files_yaml: |
@@ -50,32 +50,37 @@ jobs:
               - tfhe/build.rs
               - backends/tfhe-cuda-backend/**
               - tfhe/src/core_crypto/gpu/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_unsigned/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_signed/**
+              - tfhe/src/integer/server_key/radix_parallel/tests_cases_unsigned.rs
               - tfhe/src/integer/gpu/**
-              - tfhe/shortint/parameters/**
+              - tfhe/src/shortint/parameters/**
               - tfhe/src/high_level_api/**
               - tfhe/src/c_api/**
               - 'tfhe/docs/**.md'
+              - '.github/workflows/gpu_unsigned_integer_tests.yml'
+              - scripts/integer-tests.sh
+              - ci/slab.toml
 
   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') ||
-      (github.event.action == 'labeled' && github.event.label.name == 'approved' && needs.should-run.outputs.gpu_test == 'true')
+      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
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: start
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
           slab-url: ${{ secrets.SLAB_BASE_URL }}
           job-secret: ${{ secrets.JOB_SECRET }}
-          backend: aws
+          backend: hyperstack
           profile: gpu-test
 
   cuda-unsigned-integer-tests:
@@ -94,20 +99,33 @@ jobs:
         include:
           - os: ubuntu-22.04
             cuda: "12.2"
-            gcc: 9
+            gcc: 11
     env:
       CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-
+      CMAKE_VERSION: 3.29.6
     steps:
+      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y checkinstall zlib1g-dev libssl-dev libclang-dev
+          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
+          cd cmake-${{ env.CMAKE_VERSION }}
+          ./bootstrap
+          make -j"$(nproc)"
+          sudo make install
+
+
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
@@ -138,22 +156,26 @@ jobs:
             echo "NIGHTLY_TESTS=TRUE";
           } >> "${GITHUB_ENV}"
 
-      - name: Run unsigned integer tests
-        if: github.event_name != 'pull_request' || contains(github.event.label.name, 'approved')
-        run: |
-          make test_unsigned_integer_gpu_ci
+      - 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
 
-      - name: Slack Notification
-        if: ${{ always() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+  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' && failure() }}
+    continue-on-error: true
+    steps:
+      - name: Send message
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "CUDA AWS unsigned integer tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          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)
@@ -163,7 +185,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -174,7 +196,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (cuda-unsigned-integer-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/integer_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 GPU full benchmarks
-
-on:
-  workflow_dispatch:
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 1a.m.
-    - cron: '0 1 * * 6'
-
-env:
-  CARGO_TERM_COLOR: always
-  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-
-jobs:
-  setup-instance:
-    name: Setup instance (cuda-integer-full-benchmarks)
-    runs-on: ubuntu-latest
-    if: github.event_name != 'schedule' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@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: ${{ needs.setup-instance.outputs.runner-name }}
-    timeout-minutes: 1440 # 24 hours
-    continue-on-error: true
-    strategy:
-      fail-fast: false
-      max-parallel: 1
-      matrix:
-        command: [integer, integer_multi_bit]
-        op_flavor: [default, unchecked]
-        # explicit include-based build matrix, of known valid options
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ 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@21dc36fb71dd22e3317045c0c31a3f4249868b17
-        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 ${{ env.RESULTS_FILENAME }} \
-          --database tfhe_rs \
-          --hardware "n3-H100x1" \
-          --backend gpu \
-          --project-version "${{ env.COMMIT_HASH }}" \
-          --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
-          --bench-date "${{ env.BENCH_DATE }}" \
-          --walk-subdirs \
-          --name-suffix avx512 \
-          --throughput
-
-      - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
-        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-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-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() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-integer-full-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/integer_long_run_tests.yml

@@ -0,0 +1,94 @@
+name: AWS Long Run Tests on CPU
+
+env:
+  CARGO_TERM_COLOR: always
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+  RUSTFLAGS: "-C target-cpu=native"
+  RUST_BACKTRACE: "full"
+  RUST_MIN_STACK: "8388608"
+  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
+
+on:
+  # Allows you to run this workflow manually from the Actions tab as an alternative.
+  workflow_dispatch:
+  schedule:
+    # Weekly tests will be triggered each Friday at 1a.m.
+    - cron: '0 1 * * FRI'
+
+jobs:
+  setup-instance:
+    name: Setup instance (cpu-tests)
+    if: github.event_name != 'schedule' ||
+      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
+    runs-on: ubuntu-latest
+    outputs:
+      runner-name: ${{ steps.start-instance.outputs.label }}
+    steps:
+      - name: Start instance
+        id: start-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: start
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          backend: aws
+          profile: cpu-big
+
+  cpu-tests:
+    name: Long run CPU tests
+    needs: [ setup-instance ]
+    concurrency:
+      group: ${{ github.workflow }}_${{github.event_name}}_${{ github.ref }}
+      cancel-in-progress: true
+    runs-on: ${{ needs.setup-instance.outputs.runner-name }}
+    steps:
+      - name: Checkout tfhe-rs
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          persist-credentials: 'false'
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Install latest stable
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
+        with:
+          toolchain: stable
+
+      - name: Run tests
+        run: |
+          make test_integer_long_run
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "CPU long run tests finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+
+  teardown-instance:
+    name: Teardown instance (cpu-tests)
+    if: ${{ always() && needs.setup-instance.result != 'skipped' }}
+    needs: [ setup-instance, cpu-tests ]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Stop instance
+        id: stop-instance
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
+        with:
+          mode: stop
+          github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
+          slab-url: ${{ secrets.SLAB_BASE_URL }}
+          job-secret: ${{ secrets.JOB_SECRET }}
+          label: ${{ needs.setup-instance.outputs.runner-name }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_MESSAGE: "Instance teardown (cpu-long-run-tests) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/integer_multi_bit_gpu_benchmark.yml

@@ -1,232 +0,0 @@
-# Run integer benchmarks with multi-bit cryptographic parameters on an instance and return parsed results to Slab CI bot.
-name: Integer GPU Multi-bit benchmarks
-
-on:
-  workflow_dispatch:
-    inputs:
-      all_precisions:
-        description: "Run all precisions"
-        type: boolean
-        default: false
-      fast_default:
-        description: "Run only deduplicated default operations without scalar variants"
-        type: boolean
-        default: false
-
-  schedule:
-    # Weekly benchmarks will be triggered each Saturday at 1a.m.
-    - cron: '0 1 * * 6'
-
-env:
-  CARGO_TERM_COLOR: always
-  RESULTS_FILENAME: parsed_benchmark_results_${{ github.sha }}.json
-  PARSE_INTEGER_BENCH_CSV_FILE: tfhe_rs_integer_benches_${{ github.sha }}.csv
-  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
-  RUST_BACKTRACE: "full"
-  RUST_MIN_STACK: "8388608"
-  SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
-  SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-  SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
-  SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
-  FAST_BENCH: TRUE
-  BENCH_OP_FLAVOR: default
-
-jobs:
-  setup-instance:
-    name: Setup instance (cuda-integer-multi-bit-benchmarks)
-    runs-on: ubuntu-latest
-    if: github.event_name != 'schedule' ||
-      (github.event_name == 'schedule' && github.repository == 'zama-ai/tfhe-rs')
-    outputs:
-      runner-name: ${{ steps.start-instance.outputs.label }}
-    steps:
-      - name: Start instance
-        id: start-instance
-        uses: zama-ai/slab-github-runner@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: ${{ needs.setup-instance.outputs.runner-name }}
-    timeout-minutes: 1440 # 24 hours
-    strategy:
-      fail-fast: false
-      # explicit include-based build matrix, of known valid options
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            cuda: "12.2"
-            gcc: 11
-    env:
-      CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
-      CMAKE_VERSION: 3.29.6
-    steps:
-      # Mandatory on hyperstack since a bootable volume is not re-usable yet.
-      - name: Install dependencies
-        run: |
-          sudo apt update
-          sudo apt install -y checkinstall zlib1g-dev libssl-dev
-          wget https://github.com/Kitware/CMake/releases/download/v${{ env.CMAKE_VERSION }}/cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          tar -zxvf cmake-${{ env.CMAKE_VERSION }}.tar.gz
-          cd cmake-${{ env.CMAKE_VERSION }}
-          ./bootstrap
-          make -j"$(nproc)"
-          sudo make install
-
-      - name: Checkout tfhe-rs repo with tags
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
-        with:
-          fetch-depth: 0
-
-      - name: Get benchmark details
-        run: |
-          {
-            echo "BENCH_DATE=$(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@21dc36fb71dd22e3317045c0c31a3f4249868b17
-        with:
-          toolchain: nightly
-
-      - name: Export CUDA variables
-        if: ${{ !cancelled() }}
-        run: |
-          {
-            echo "CUDA_PATH=$CUDA_PATH";
-            echo "LD_LIBRARY_PATH=$CUDA_PATH/lib:$LD_LIBRARY_PATH";
-            echo "CUDACXX=/usr/local/cuda-${{ matrix.cuda }}/bin/nvcc";
-          } >> "${GITHUB_ENV}"
-          echo "$CUDA_PATH/bin" >> "${GITHUB_PATH}"
-
-      # Specify the correct host compilers
-      - name: Export gcc and g++ variables
-        if: ${{ !cancelled() }}
-        run: |
-          {
-            echo "CC=/usr/bin/gcc-${{ matrix.gcc }}";
-            echo "CXX=/usr/bin/g++-${{ matrix.gcc }}";
-            echo "CUDAHOSTCXX=/usr/bin/g++-${{ matrix.gcc }}";
-          } >> "${GITHUB_ENV}"
-
-      - name: Should run benchmarks with all precisions
-        if: inputs.all_precisions
-        run: |
-          echo "FAST_BENCH=FALSE" >> "${GITHUB_ENV}"
-
-      - name: Should run fast subset benchmarks
-        if: inputs.fast_default
-        run: |
-          echo "BENCH_OP_FLAVOR=fast_default" >> "${GITHUB_ENV}"
-
-      - name: Check device is detected
-        if: ${{ !cancelled() }}
-        run: nvidia-smi
-
-      - name: Run multi-bit benchmarks with AVX512
-        run: |
-          make bench_unsigned_integer_multi_bit_gpu
-
-      - name: Parse benchmarks to csv
-        run: |
-          make PARSE_INTEGER_BENCH_CSV_FILE=${{ env.PARSE_INTEGER_BENCH_CSV_FILE }} \
-            parse_integer_benches
-
-      - name: Upload csv results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
-        with:
-          name: ${{ github.sha }}_csv_integer
-          path: ${{ env.PARSE_INTEGER_BENCH_CSV_FILE }}
-
-      - name: Parse results
-        run: |
-          python3 ./ci/benchmark_parser.py target/criterion ${{ env.RESULTS_FILENAME }} \
-          --database tfhe_rs \
-          --hardware "n3-H100x1" \
-          --backend gpu \
-          --project-version "${{ env.COMMIT_HASH }}" \
-          --branch ${{ github.ref_name }} \
-          --commit-date "${{ env.COMMIT_DATE }}" \
-          --bench-date "${{ env.BENCH_DATE }}" \
-          --walk-subdirs \
-          --name-suffix avx512 \
-          --throughput
-
-      - name: Upload parsed results artifact
-        uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b
-        with:
-          name: ${{ github.sha }}_integer
-          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 }}
-
-
-  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() }}
-        continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
-        env:
-          SLACK_COLOR: ${{ job.status }}
-          SLACK_MESSAGE: "Instance teardown (cuda-integer-multi-bit-benchmarks) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/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-nvlink
-
-  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@21dc36fb71dd22e3317045c0c31a3f4249868b17
-        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 ${{ env.RESULTS_FILENAME }} \
-          --database tfhe_rs \
-          --hardware "n3-H100x8-NVLink" \
-          --backend gpu \
-          --project-version "${{ 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@0b2256b8c012f0828dc542b3febcab082c67f72b
-        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-nvlink
-
-  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@21dc36fb71dd22e3317045c0c31a3f4249868b17
-        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 ${{ env.RESULTS_FILENAME }} \
-          --database tfhe_rs \
-          --hardware "n3-H100x8-NVLink" \
-          --backend gpu \
-          --project-version "${{ 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@0b2256b8c012f0828dc542b3febcab082c67f72b
-        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
@@ -27,29 +27,66 @@ concurrency:
   cancel-in-progress: true
 
 jobs:
-  cargo-builds:
+  cargo-builds-m1:
     if: ${{ (github.event_name == 'schedule' &&  github.repository == 'zama-ai/tfhe-rs') || github.event_name == 'workflow_dispatch' || contains(github.event.label.name, 'm1_test') }}
     runs-on: ["self-hosted", "m1mac"]
     # 12 hours, default is 6 hours, hopefully this is more than enough
     timeout-minutes: 720
 
     steps:
-      - uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+      - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
-          persist-credentials: 'false'
+          persist-credentials: "false"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
+      - name: Run pcc FFT checks
+        run: |
+          make pcc_fft
+
+      - name: Build FFT release
+        run: |
+          make build_fft
+
+      - name: Build FFT release no-std
+        run: |
+          make build_fft_no_std
+
+      - name: Run FFT tests
+        run: |
+          make test_fft
+          make test_fft_serde
+          make test_fft_nightly
+          make test_fft_no_std
+          make test_fft_no_std_nightly
+          # we don't run the js stuff here as it's causing issues with the M1 config
+
+      - name: Run pcc NTT checks
+        run: |
+          make pcc_ntt
+
+      - name: Build NTT release
+        run: |
+          make build_ntt
+
+      - name: Build NTT release no-std
+        run: |
+          make build_ntt_no_std
+
+      - name: Run NTT tests
+        run: |
+          make test_ntt_all
+
       - name: Run pcc checks
         run: |
           make pcc
 
-      - name: Build concrete-csprng
+      - name: Build tfhe-csprng
         run: |
-          make build_concrete_csprng
+          make build_tfhe_csprng
 
       - name: Build Release core
         run: |
@@ -75,9 +112,9 @@ jobs:
         run: |
           make build_c_api
 
-      - name: Run concrete-csprng tests
+      - name: Run tfhe-csprng tests
         run: |
-          make test_concrete_csprng
+          make test_tfhe_csprng
 
       - name: Run tfhe-zk-pok tests
         run: |
@@ -137,7 +174,7 @@ jobs:
     name: Remove m1_test label
     runs-on: ubuntu-latest
     needs:
-      - cargo-builds
+      - cargo-builds-m1
     if: ${{ always() }}
     steps:
       - uses: actions-ecosystem/action-remove-labels@2ce5d41b4b6aa8503e285553f75ed56e0a40bae0
@@ -147,13 +184,13 @@ jobs:
           github_token: ${{ secrets.GITHUB_TOKEN }}
 
       - name: Slack Notification
-        if: ${{ needs.cargo-builds.result != 'skipped' }}
+        if: ${{ needs.cargo-builds-m1.result != 'skipped' }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
-          SLACK_COLOR: ${{ needs.cargo-builds.result }}
+          SLACK_COLOR: ${{ needs.cargo-builds-m1.result }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
           SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-          SLACK_MESSAGE: "M1 tests finished with status: ${{ needs.cargo-builds.result }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "M1 tests finished with status: ${{ needs.cargo-builds-m1.result }}. (${{ env.ACTION_RUN_URL }})"
           SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
           SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/make_release.yml

@@ -30,19 +30,27 @@ env:
   NPM_TAG: ""
 
 jobs:
+  verify_tag:
+    uses: ./.github/workflows/verify_tagged_commit.yml
+    secrets:
+      RELEASE_TEAM: ${{ secrets.RELEASE_TEAM }}
+      READ_ORG_TOKEN: ${{ secrets.READ_ORG_TOKEN }}
+
   package:
     runs-on: ubuntu-latest
+    needs: verify_tag
     outputs:
       hash: ${{ steps.hash.outputs.hash }}
     steps:
       - name: Checkout
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
       - name: Prepare package
         run: |
           cargo package -p tfhe
-      - uses: actions/upload-artifact@0b2256b8c012f0828dc542b3febcab082c67f72b # v4.3.4
+      - uses: actions/upload-artifact@b4b15b8c7c6ac21ea08fcf65892d2ee8f75cf882 # v4.4.3
         with:
           name: crate
           path: target/package/*.crate
@@ -74,9 +82,10 @@ jobs:
       id-token: write
     steps:
       - name: Checkout
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
       - name: Create NPM version tag
         if: ${{ inputs.npm_latest_tag }}
         run: |
@@ -101,7 +110,7 @@ jobs:
       - name: Slack notification (hashes comparison)
         if: ${{ needs.package.outputs.hash != steps.published_hash.outputs.pub_hash }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: failure
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
@@ -146,7 +155,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/make_release_concrete_csprng.yml

@@ -1,5 +1,4 @@
-# Publish new release of tfhe-rs on various platform.
-name: Publish concrete-csprng release
+name: Publish tfhe-csprng release
 
 on:
   workflow_dispatch:
@@ -13,30 +12,38 @@ env:
   ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
 
 jobs:
+  verify_tag:
+    uses: ./.github/workflows/verify_tagged_commit.yml
+    secrets:
+      RELEASE_TEAM: ${{ secrets.RELEASE_TEAM }}
+      READ_ORG_TOKEN: ${{ secrets.READ_ORG_TOKEN }}
+
   publish_release:
-    name: Publish concrete-csprng Release
+    name: Publish tfhe-csprng Release
+    needs: verify_tag
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Publish crate.io package
         env:
           CRATES_TOKEN: ${{ secrets.CARGO_REGISTRY_TOKEN }}
           DRY_RUN: ${{ inputs.dry_run && '--dry-run' || '' }}
         run: |
-          cargo publish -p concrete-csprng --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
+          cargo publish -p tfhe-csprng --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
 
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
           SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
-          SLACK_MESSAGE: "concrete-csprng release failed: (${{ env.ACTION_RUN_URL }})"
+          SLACK_MESSAGE: "tfhe-csprng release finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
           SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
           SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/make_release_cuda.yml

@@ -21,22 +21,29 @@ env:
   SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
 
 jobs:
+  verify_tag:
+    uses: ./.github/workflows/verify_tagged_commit.yml
+    secrets:
+      RELEASE_TEAM: ${{ secrets.RELEASE_TEAM }}
+      READ_ORG_TOKEN: ${{ secrets.READ_ORG_TOKEN }}
+
   setup-instance:
     name: Setup instance (publish-cuda-release)
+    needs: verify_tag
     runs-on: ubuntu-latest
     outputs:
       runner-name: ${{ steps.start-instance.outputs.label }}
     steps:
       - name: Start instance
         id: start-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         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
+          profile: gpu-build
 
   publish-cuda-release:
     name: Publish CUDA Release
@@ -54,16 +61,17 @@ jobs:
       CUDA_PATH: /usr/local/cuda-${{ matrix.cuda }}
     steps:
       - name: Checkout
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Set up home
         run: |
           echo "HOME=/home/ubuntu" >> "${GITHUB_ENV}"
 
       - name: Install latest stable
-        uses: dtolnay/rust-toolchain@21dc36fb71dd22e3317045c0c31a3f4249868b17
+        uses: dtolnay/rust-toolchain@315e265cd78dad1e1dcf3a5074f6d6c47029d5aa
         with:
           toolchain: stable
 
@@ -99,7 +107,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "tfhe-cuda-backend release finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
@@ -112,7 +120,7 @@ jobs:
     steps:
       - name: Stop instance
         id: stop-instance
-        uses: zama-ai/slab-github-runner@447a2d0fd2d1a9d647aa0d0723a6e9255372f261
+        uses: zama-ai/slab-github-runner@98f0788261a7323d5d695a883e20df36591a92b7
         with:
           mode: stop
           github-token: ${{ secrets.SLAB_ACTION_TOKEN }}
@@ -123,7 +131,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_MESSAGE: "Instance teardown (publish-cuda-release) finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"

.github/workflows/make_release_tfhe_fft.yml

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

.github/workflows/make_release_tfhe_ntt.yml

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

.github/workflows/make_release_tfhe_versionable.yml

@@ -0,0 +1,56 @@
+name: Publish tfhe-versionable release
+
+on:
+  workflow_dispatch:
+    inputs:
+      dry_run:
+        description: "Dry-run"
+        type: boolean
+        default: true
+
+env:
+  ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
+
+jobs:
+  verify_tag:
+    uses: ./.github/workflows/verify_tagged_commit.yml
+    secrets:
+      RELEASE_TEAM: ${{ secrets.RELEASE_TEAM }}
+      READ_ORG_TOKEN: ${{ secrets.READ_ORG_TOKEN }}
+
+  publish_release:
+    name: Publish tfhe-versionable Release
+    needs: verify_tag
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
+        with:
+          fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
+
+      - name: Publish proc-macro crate
+        env:
+          CRATES_TOKEN: ${{ secrets.CARGO_REGISTRY_TOKEN }}
+          DRY_RUN: ${{ inputs.dry_run && '--dry-run' || '' }}
+        run: |
+          cargo publish -p tfhe-versionable-derive --token ${{ env.CRATES_TOKEN }} ${{ env.DRY_RUN }}
+
+      - name: Publish main crate
+        if: ${{ ! inputs.dry_run }}
+        env:
+          CRATES_TOKEN: ${{ secrets.CARGO_REGISTRY_TOKEN }}
+        run: |
+          cargo publish -p tfhe-versionable --token ${{ env.CRATES_TOKEN }}
+
+      - name: Slack Notification
+        if: ${{ failure() }}
+        continue-on-error: true
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
+        env:
+          SLACK_COLOR: ${{ job.status }}
+          SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}
+          SLACK_ICON: https://pbs.twimg.com/profile_images/1274014582265298945/OjBKP9kn_400x400.png
+          SLACK_MESSAGE: "tfhe-versionable release finished with status: ${{ job.status }}. (${{ env.ACTION_RUN_URL }})"
+          SLACK_USERNAME: ${{ secrets.BOT_USERNAME }}
+          SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}

.github/workflows/make_release_zk_pok.yml

@@ -13,14 +13,22 @@ env:
   ACTION_RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
 
 jobs:
+  verify_tag:
+    uses: ./.github/workflows/verify_tagged_commit.yml
+    secrets:
+      RELEASE_TEAM: ${{ secrets.RELEASE_TEAM }}
+      READ_ORG_TOKEN: ${{ secrets.READ_ORG_TOKEN }}
+
   publish_release:
     name: Publish tfhe-zk-pok Release
+    needs: verify_tag
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
 
       - name: Publish crate.io package
         env:
@@ -32,7 +40,7 @@ jobs:
       - name: Slack Notification
         if: ${{ failure() }}
         continue-on-error: true
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/parameters_check.yml

@@ -14,17 +14,17 @@ on:
 
 jobs:
   params-curves-security-check:
-    runs-on: ubuntu-latest
+    runs-on: large_ubuntu_16
     steps:
       - name: Checkout tfhe-rs
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
 
       - name: Checkout lattice-estimator
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           repository: malb/lattice-estimator
           path: lattice_estimator
-          ref: '53508253629d3b5d31a2ad110e85dc69391ccb95'
+          ref: 'e80ec6bbbba212428b0e92d0467c18629cf9ed67'
 
       - 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@c33737706dea87cd7784c687dadc9adf1be59990
         env:
           SLACK_COLOR: ${{ job.status }}
           SLACK_CHANNEL: ${{ secrets.SLACK_CHANNEL }}

.github/workflows/sync_on_push.yml

@@ -13,9 +13,10 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout repo
-        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332
+        uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683
         with:
           fetch-depth: 0
+          token: ${{ secrets.FHE_ACTIONS_TOKEN }}
       - name: git-sync
         uses: wei/git-sync@55c6b63b4f21607da0e9877ca9b4d11a29fc6d83
         with:

.github/workflows/verify_tagged_commit.yml

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

.gitignore

@@ -12,7 +12,8 @@ target/
 **/*.bin
 
 # Some of our bench outputs
-/tfhe/benchmarks_parameters
+/crates/tfhe/benchmarks_parameters
+/crates/tfhe-zk-pok/benchmarks_parameters
 **/*.csv
 
 # dieharder run log
@@ -25,7 +26,11 @@ dieharder_run.log
 backends/tfhe-cuda-backend/cuda/cmake-build-debug/
 
 # WASM tests
-tfhe/web_wasm_parallel_tests/server.PID
+crates/tfhe/tests/web_wasm_parallel/server.PID
+venv/
+web-test-runner/
+node_modules/
+package-lock.json
 
 # Dir used for backward compatibility test data
-tfhe/tfhe-backward-compat-data/
+crates/tfhe/tfhe-backward-compat-data/

Cargo.toml

@@ -1,22 +1,40 @@
 [workspace]
 resolver = "2"
 members = [
-    "tfhe",
-    "tfhe-zk-pok",
+    "crates/*",
     "tasks",
     "apps/trivium",
-    "concrete-csprng",
     "backends/tfhe-cuda-backend",
-    "utils/tfhe-versionable",
-    "utils/tfhe-versionable-derive",
 ]
 
 exclude = [
-    "tfhe/backward_compatibility_tests",
+    "crates/tfhe/backward_compatibility_tests",
     "utils/cargo-tfhe-lints-inner",
     "utils/cargo-tfhe-lints"
 ]
 
+[workspace.dependencies]
+aligned-vec = { version = "0.5", default-features = false }
+bytemuck = "1.14.3"
+dyn-stack = { version = "0.10", default-features = false }
+num-complex = "0.4"
+pulp = { version = "0.18.22", default-features = false }
+serde = { version = "1.0", default-features = false }
+wasm-bindgen = ">=0.2.86,<0.2.94"
+
+[workspace.package]
+version = "0.11.0"
+license = "BSD-3-Clause-Clear"
+repository = "https://github.com/zama-ai/tfhe-rs"
+documentation = "https://docs.zama.ai/tfhe-rs"
+
+[workspace.lints.rust]
+unexpected_cfgs = { level = "warn", check-cfg = [
+    'cfg(bench)',
+    'cfg(tarpaulin)',
+    'cfg(tfhe_lints)',
+] }
+
 [profile.bench]
 lto = "fat"
 

README-TFHE-rs.md

@@ -1,280 +0,0 @@
-<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>
-</p>
-
-<hr/>
-
-<p align="center">
-  <a href="https://docs.zama.ai/tfhe-rs"> 📒 Documentation</a> | <a href="https://zama.ai/community"> 💛 Community support</a> | <a href="https://github.com/zama-ai/awesome-zama"> 📚 FHE resources by Zama</a>
-</p>
-
-
-<p align="center">
-  <a href="https://github.com/zama-ai/tfhe-rs/releases"><img src="https://img.shields.io/github/v/release/zama-ai/tfhe-rs?style=flat-square"></a>
-  <a href="LICENSE"><img src="https://img.shields.io/badge/License-BSD--3--Clause--Clear-%23ffb243?style=flat-square"></a>
-  <a href="https://github.com/zama-ai/bounty-program"><img src="https://img.shields.io/badge/Contribute-Zama%20Bounty%20Program-%23ffd208?style=flat-square"></a>
-</p>
-
-## About
-
-### What is TFHE-rs
-
-**TFHE-rs** is a pure Rust implementation of TFHE for boolean and integer arithmetics over encrypted data.
-
-It includes:
-- a **Rust** API
-- a **C** API
-- and a **client-side WASM** API
-
-TFHE-rs is designed for developers and researchers who want full control over
-what they can do with TFHE, while not having to worry about the low-level
-implementation. The goal is to have a stable, simple, high-performance, and
-production-ready library for all the advanced features of TFHE.
-<br></br>
-
-### Main features
-
-- **Low-level cryptographic library** that implements Zama’s variant of TFHE, including programmable bootstrapping
-- **Implementation of the original TFHE boolean API** that can be used as a drop-in replacement for other TFHE libraries
-- **Short integer API** that enables exact, unbounded FHE integer arithmetics with up to 8 bits of message space
-- **Size-efficient public key encryption**
-- **Ciphertext and server key compression** for efficient data transfer
-- **Full Rust API, C bindings to the Rust High-Level API, and client-side Javascript API using WASM**.
-
-*Learn more about TFHE-rs features in the [documentation](https://docs.zama.ai/tfhe-rs/readme).*
-<br></br>
-
-## Table of Contents
-- **[Getting started](#getting-started)**
-   - [Cargo.toml configuration](#cargotoml-configuration)
-   - [A simple example](#a-simple-example)
-- **[Resources](#resources)**
-   - [TFHE deep dive](#tfhe-deep-dive)
-   - [Tutorials](#tutorials)
-   - [Documentation](#documentation)
-- **[Working with TFHE-rs](#working-with-tfhe-rs)**
-   - [Disclaimers](#disclaimers)
-   - [Citations](#citations)
-   - [Contributing](#contributing)
-   - [License](#license)
-- **[Support](#support)**
-<br></br>
-
-## 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`:
-
-+ For x86_64-based machines running Unix-like OSes:
-
-```toml
-tfhe = { version = "*", features = ["boolean", "shortint", "integer", "x86_64-unix"] }
-```
-
-+ For Apple Silicon or aarch64-based machines running Unix-like OSes:
-
-```toml
-tfhe = { version = "*", features = ["boolean", "shortint", "integer", "aarch64-unix"] }
-```
-
-+ For x86_64-based machines with the [`rdseed instruction`](https://en.wikipedia.org/wiki/RDRAND) running Windows:
-
-```toml
-tfhe = { version = "*", features = ["boolean", "shortint", "integer", "x86_64"] }
-```
-
-> [!Note]
-> Note: You need to use a Rust version >= 1.73 to compile TFHE-rs.
-
-> [!Note]
-> Note: aarch64-based machines are not yet supported for Windows as it's currently missing an entropy source to be able to seed the [CSPRNGs](https://en.wikipedia.org/wiki/Cryptographically_secure_pseudorandom_number_generator) used in TFHE-rs.
-
-<p align="right">
-  <a href="#about" > ↑ Back to top </a> 
-</p>
-
-### A simple example
-
-Here is a full example:
-
-``` rust
-use tfhe::prelude::*;
-use tfhe::{generate_keys, set_server_key, ConfigBuilder, FheUint32, FheUint8};
-
-fn main() -> Result<(), Box<dyn std::error::Error>> {
-    // Basic configuration to use homomorphic integers
-    let config = ConfigBuilder::default().build();
-
-    // Key generation
-    let (client_key, server_keys) = generate_keys(config);
-
-    let clear_a = 1344u32;
-    let clear_b = 5u32;
-    let clear_c = 7u8;
-
-    // Encrypting the input data using the (private) client_key
-    // FheUint32: Encrypted equivalent to u32
-    let mut encrypted_a = FheUint32::try_encrypt(clear_a, &client_key)?;
-    let encrypted_b = FheUint32::try_encrypt(clear_b, &client_key)?;
-
-    // FheUint8: Encrypted equivalent to u8
-    let encrypted_c = FheUint8::try_encrypt(clear_c, &client_key)?;
-
-    // On the server side:
-    set_server_key(server_keys);
-
-    // Clear equivalent computations: 1344 * 5 = 6720
-    let encrypted_res_mul = &encrypted_a * &encrypted_b;
-
-    // Clear equivalent computations: 6720 >> 5 = 210
-    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
-    let encrypted_res_min = &casted_a.min(&encrypted_c);
-
-    // Operation between clear and encrypted data:
-    // Clear equivalent computations: 7 & 1 = 1
-    let encrypted_res = encrypted_res_min & 1_u8;
-
-    // Decrypting on the client side:
-    let clear_res: u8 = encrypted_res.decrypt(&client_key);
-    assert_eq!(clear_res, 1_u8);
-
-    Ok(())
-}
-```
-
-To run this code, use the following command: 
-<p align="center"> <code> cargo run --release </code> </p>
-
-> [!Note]
-> Note that when running code that uses `TFHE-rs`, it is highly recommended
-to run in release mode with cargo's `--release` flag to have the best performances possible.
-
-*Find an example with more explanations in [this part of the documentation](https://docs.zama.ai/tfhe-rs/getting-started/quick_start)*
-
-<p align="right">
-  <a href="#about" > ↑ Back to top </a> 
-</p>
-
-
-
-## Resources 
-
-### TFHE deep dive
-- [TFHE Deep Dive - Part I - Ciphertext types](https://www.zama.ai/post/tfhe-deep-dive-part-1)
-- [TFHE Deep Dive - Part II - Encodings and linear leveled operations](https://www.zama.ai/post/tfhe-deep-dive-part-2)
-- [TFHE Deep Dive - Part III - Key switching and leveled multiplications](https://www.zama.ai/post/tfhe-deep-dive-part-3)
-- [TFHE Deep Dive - Part IV - Programmable Bootstrapping](https://www.zama.ai/post/tfhe-deep-dive-part-4)
-<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)
-- [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)
-
-*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>
-### Documentation
-
-Full, comprehensive documentation is available here: [https://docs.zama.ai/tfhe-rs](https://docs.zama.ai/tfhe-rs).
-<p align="right">
-  <a href="#about" > ↑ Back to top </a> 
-</p>
-
-
-## Working with TFHE-rs
-
-### Disclaimers
-
-#### Security estimation
-
-Security estimations are done using the
-[Lattice Estimator](https://github.com/malb/lattice-estimator)
-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
-
-Mitigation for side-channel attacks has not yet been implemented in TFHE-rs,
-and will be released in upcoming versions.
-<br></br>
-
-### Citations
-To cite TFHE-rs in academic papers, please use the following entry:
-
-```text
-@Misc{TFHE-rs,
-  title={{TFHE-rs: A Pure Rust Implementation of the TFHE Scheme for Boolean and Integer Arithmetics Over Encrypted Data}},
-  author={Zama},
-  year={2022},
-  note={\url{https://github.com/zama-ai/tfhe-rs}},
-}
-```
-
-### Contributing
-
-There are two ways to contribute to TFHE-rs:
-
-- [Open issues](https://github.com/zama-ai/tfhe-rs/issues/new/choose) to report bugs and typos, or to suggest new ideas
-- Request to become an official contributor by emailing [hello@zama.ai](mailto:hello@zama.ai).
-
-Becoming an approved contributor involves signing our Contributor License Agreement (CLA). Only approved contributors can send pull requests, so please make sure to get in touch before you do!
-<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.
-
-<p align="right">
-  <a href="#about" > ↑ Back to top </a> 
-</p>
-
-
-## 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>
-</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. 
-
-<p align="right">
-  <a href="#about" > ↑ Back to top </a> 
-</p>

README.md

@@ -1,114 +1,280 @@
-# Artifact: Accelerating TFHE with Sorted-Bootstrapping Techniques
+<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>
+</p>
+
+<hr/>
+
+<p align="center">
+  <a href="https://docs.zama.ai/tfhe-rs"> 📒 Documentation</a> | <a href="https://zama.ai/community"> 💛 Community support</a> | <a href="https://github.com/zama-ai/awesome-zama"> 📚 FHE resources by Zama</a>
+</p>
 
 
-## Description
+<p align="center">
+  <a href="https://github.com/zama-ai/tfhe-rs/releases"><img src="https://img.shields.io/github/v/release/zama-ai/tfhe-rs?style=flat-square"></a>
+  <a href="LICENSE"><img src="https://img.shields.io/badge/License-BSD--3--Clause--Clear-%23ffb243?style=flat-square"></a>
+  <a href="https://github.com/zama-ai/bounty-program"><img src="https://img.shields.io/badge/Contribute-Zama%20Bounty%20Program-%23ffd208?style=flat-square"></a>
+</p>
 
-In what follows, we provide instructions on how to run the benchmarks given in **Table 3** (page 29), from the paper titled **Accelerating TFHE with Sorted Bootstrapping Techniques**.
-The implementation of the techniques described in the aforementioned paper has been integrated into the **TFHE-rs** library, version `0.8.0-alpha.2`.
-The code related to the  extended PBS (EBS), the sorted PBS (SBS) and the version with the companion modulus switch are located in ```tfhe/src/core_crypto/fft_impl/fft64/crypto/bootstrap.rs```.
-All the benchmarks can be found in ```tfhe/benches/core_crypto/pbs_bench.rs```.
+## About
 
+### What is TFHE-rs
 
-## Setup and Dependencies
+**TFHE-rs** is a pure Rust implementation of TFHE for boolean and integer arithmetics over encrypted data.
 
-Tested on Linux and macOS with Rust version ≥ 1.85 (we recommend installing Rust via [rustup](https://www.rust-lang.org/tools/install), as rustup is required later by the provided Makefile).
-The complete list of dependencies and a guide on how to install TFHE-rs can be found in the online documentation [here](https://docs.zama.ai/tfhe-rs/0.8) or in the local file [here](./README-TFHE-rs.md).
+It includes:
+- a **Rust** API
+- a **C** API
+- and a **client-side WASM** API
 
-## How to run benchmarks
-The benchmarks run across all precision and all failure probabilities, progressing from lower precision (4 bits) to higher precision (9 bits) and from higher failure probabilities ($2^{-64}$) to the lower failure probabilities ($2^{-128}$). 
-All the following benchmarks are **sequential**.
-At the root of the project (i.e., in the TFHE-rs folder), enter the following commands to run the different benchmarks:
+TFHE-rs is designed for developers and researchers who want full control over
+what they can do with TFHE, while not having to worry about the low-level
+implementation. The goal is to have a stable, simple, high-performance, and
+production-ready library for all the advanced features of TFHE.
+<br></br>
 
-- ```make bench_ly```: returns the timings associated with the extended PBS (EBS[LY23] in **Table 3**).
-- ```make bench_sorted```: returns the timings associated with the sorted PBS  (SBS in **Table 3**).
-- ```make bench_cms```: returns the timings associated with the SBS with the companion modulus switch (cms) (SBS + CMS in **Table 3**).
+### Main features
 
+- **Low-level cryptographic library** that implements Zama’s variant of TFHE, including programmable bootstrapping
+- **Implementation of the original TFHE boolean API** that can be used as a drop-in replacement for other TFHE libraries
+- **Short integer API** that enables exact, unbounded FHE integer arithmetics with up to 8 bits of message space
+- **Size-efficient public key encryption**
+- **Ciphertext and server key compression** for efficient data transfer
+- **Full Rust API, C bindings to the Rust High-Level API, and client-side Javascript API using WASM**.
 
-To modify the performed benchmarks, only a few lines need to be commented and modified in ```tfhe/benches/core_crypto/pbs_bench.rs```.
-To focus only on specific benchmarks, you only need to keep the wanted precision and failure probabilities and adjust the number of tests with ```[ParametersLY23; X]```, where `X` must be equal to the number of uncommented parameters.
+*Learn more about TFHE-rs features in the [documentation](https://docs.zama.ai/tfhe-rs/readme).*
+<br></br>
 
-All parameter sets follow a syntax of the form: ```{Name}_{Precision}_{-Log2(pfail)}```. 
-For instance,  ```LY_5_64``` means that the parameter set is associated to the algorithm referred as *LY* in the paper, with *5* bits of precision for input message and $2^{-64}$ as failure probability.
+## Table of Contents
+- **[Getting started](#getting-started)**
+   - [Cargo.toml configuration](#cargotoml-configuration)
+   - [A simple example](#a-simple-example)
+- **[Resources](#resources)**
+   - [TFHE deep dive](#tfhe-deep-dive)
+   - [Tutorials](#tutorials)
+   - [Documentation](#documentation)
+- **[Working with TFHE-rs](#working-with-tfhe-rs)**
+   - [Disclaimers](#disclaimers)
+   - [Citations](#citations)
+   - [Contributing](#contributing)
+   - [License](#license)
+- **[Support](#support)**
+<br></br>
 
-For the EBS and the SBS, you need to modify lines 1502 to 1507 and lines 1612 to 1616 for the SBS with CMS.
-For instance, changing lines 1502 to 1507:
-```rust 
-const PARAM_BENCHES_LY23: [ParametersLY23; 16] = [
-  //LY_5_40, LY_6_40, LY_7_40, LY_8_40, LY_9_40
-  LY_5_64, LY_6_64, LY_7_64, LY_8_64, LY_9_64,
-  LY_5_80, LY_6_80, LY_7_81, LY_8_81, LY_9_81,
-  LY_4_128, LY_5_128, LY_6_129, LY_7_128, LY_8_128, LY_9_129,
-];
-```
-into:
-```rust 
-const PARAM_BENCHES_LY23: [ParametersLY23; 6] = [
-  //LY_5_40, LY_6_40, LY_7_40, LY_8_40, LY_9_40,
-  //LY_5_64, LY_6_64, LY_7_64, LY_8_64, LY_9_64,
-  //LY_5_80, LY_6_80, LY_7_81, LY_8_81, LY_9_81,
-  LY_4_128, LY_5_128, LY_6_129, LY_7_128, LY_8_128, LY_9_129,
-];
-```
-and using the command  ```make bench_ly``` launches only benchmarks for the EBS experiment with a failure probability of $2^{-128}$.
+## 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`:
 
-### Sample Output Structure
++ For x86_64-based machines running Unix-like OSes:
 
-A typical benchmark result looks like this:
-
-```
-KS_Extended_PBS_LY23/PRECISION_6_BITS__EXTENDED_FACTOR_2^2__PFAIL_2^-64
-time:   [123.38 ms 123.46 ms 123.55 ms]
-Found 19 outliers among 500 measurements (3.80%)
+```toml
+tfhe = { version = "*", features = ["boolean", "shortint", "integer", "x86_64-unix"] }
 ```
 
-The first line indicates the operation whose latency is measured.
++ For Apple Silicon or aarch64-based machines running Unix-like OSes:
 
-**Examples:**
+```toml
+tfhe = { version = "*", features = ["boolean", "shortint", "integer", "aarch64-unix"] }
 ```
-KS_Extended_PBS_LY23/PRECISION_6_BITS__EXTENDED_FACTOR_2^2__PFAIL_2^-64 
+
++ For x86_64-based machines with the [`rdseed instruction`](https://en.wikipedia.org/wiki/RDRAND) running Windows:
+
+```toml
+tfhe = { version = "*", features = ["boolean", "shortint", "integer", "x86_64"] }
 ```
-means that this benchmark is measuring the latency of a Keyswitch (KS) followed by an Extended Bootstrapping (EBS [LY23]) with 6 bits of message, an extended factor equal to $2^2$, and a failure probability pfail $=2^{-64}$.
 
-#### Understanding Benchmark Output (Criterion.rs)
-This project uses Criterion.rs for benchmarking. Criterion is a powerful and statistically robust benchmarking framework for Rust, and it may produce outputs that are unfamiliar at first glance. Here is a short explanation:
+> [!Note]
+> Note: You need to use a Rust version >= 1.81 to compile TFHE-rs.
 
-    time: [low est.  median  high est.]: The estimated execution time of the function.
-    change: The performance change compared to a previous run (if available).
-    outliers: Some runs deviated from the typical time. Criterion detects and accounts for these using statistical methods.
+> [!Note]
+> Note: aarch64-based machines are not yet supported for Windows as it's currently missing an entropy source to be able to seed the [CSPRNGs](https://en.wikipedia.org/wiki/Cryptographically_secure_pseudorandom_number_generator) used in TFHE-rs.
+
+<p align="right">
+  <a href="#about" > ↑ Back to top </a> 
+</p>
+
+### A simple example
+
+Here is a full example:
+
+``` rust
+use tfhe::prelude::*;
+use tfhe::{generate_keys, set_server_key, ConfigBuilder, FheUint32, FheUint8};
+
+fn main() -> Result<(), Box<dyn std::error::Error>> {
+    // Basic configuration to use homomorphic integers
+    let config = ConfigBuilder::default().build();
+
+    // Key generation
+    let (client_key, server_keys) = generate_keys(config);
+
+    let clear_a = 1344u32;
+    let clear_b = 5u32;
+    let clear_c = 7u8;
+
+    // Encrypting the input data using the (private) client_key
+    // FheUint32: Encrypted equivalent to u32
+    let mut encrypted_a = FheUint32::try_encrypt(clear_a, &client_key)?;
+    let encrypted_b = FheUint32::try_encrypt(clear_b, &client_key)?;
+
+    // FheUint8: Encrypted equivalent to u8
+    let encrypted_c = FheUint8::try_encrypt(clear_c, &client_key)?;
+
+    // On the server side:
+    set_server_key(server_keys);
+
+    // Clear equivalent computations: 1344 * 5 = 6720
+    let encrypted_res_mul = &encrypted_a * &encrypted_b;
+
+    // Clear equivalent computations: 6720 >> 5 = 210
+    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
+    let encrypted_res_min = &casted_a.min(&encrypted_c);
+
+    // Operation between clear and encrypted data:
+    // Clear equivalent computations: 7 & 1 = 1
+    let encrypted_res = encrypted_res_min & 1_u8;
+
+    // Decrypting on the client side:
+    let clear_res: u8 = encrypted_res.decrypt(&client_key);
+    assert_eq!(clear_res, 1_u8);
+
+    Ok(())
+}
+```
+
+To run this code, use the following command: 
+<p align="center"> <code> cargo run --release </code> </p>
+
+> [!Note]
+> Note that when running code that uses `TFHE-rs`, it is highly recommended
+to run in release mode with cargo's `--release` flag to have the best performances possible.
+
+*Find an example with more explanations in [this part of the documentation](https://docs.zama.ai/tfhe-rs/get-started/quick_start)*
+
+<p align="right">
+  <a href="#about" > ↑ Back to top </a> 
+</p>
 
 
-####  Common Warnings and What They Mean
-##### `Found X outliers among Y measurements`
-Criterion runs each benchmark many times (default: 100) to get statistically significant results.
-An *outlier* is a run that was significantly faster or slower than the others.
 
-- **Why does this happen?** Often, it's due to **other processes on the machine** (e.g., background services, OS interrupts, or CPU scheduling) affecting performance temporarily.
-- **Why it doesn't invalidate results:** Criterion uses statistical techniques to minimize the impact of these outliers when estimating performance.
-- **Best practice to reduce outliers:** Run the benchmarks on a **freshly rebooted machine**, with as few background processes as possible. Ideally, let the system idle for a minute after boot to stabilize before running benchmarks.
+## Resources 
 
-##### `Unable to complete 100 samples in 5.0s.`
-The benchmark took longer than the expected 5 seconds.
-This is merely a warning indicating that the full set of 100 samples could not be collected within the default 5-second measurement window.
+### TFHE deep dive
+- [TFHE Deep Dive - Part I - Ciphertext types](https://www.zama.ai/post/tfhe-deep-dive-part-1)
+- [TFHE Deep Dive - Part II - Encodings and linear leveled operations](https://www.zama.ai/post/tfhe-deep-dive-part-2)
+- [TFHE Deep Dive - Part III - Key switching and leveled multiplications](https://www.zama.ai/post/tfhe-deep-dive-part-3)
+- [TFHE Deep Dive - Part IV - Programmable Bootstrapping](https://www.zama.ai/post/tfhe-deep-dive-part-4)
+<br></br>
 
-- **No action is required**: Criterion will still proceed to run all 100 samples, and the results remain statistically valid.
-- **Why the warning appears**: It's there to inform you that benchmarking is taking longer than expected and to help you tune settings if needed.
-- **Optional**: If you're constrained by time (e.g., running in CI), you can:
-    - Reduce the sample size (e.g., to 10 or 20 samples).
-    - Or increase the measurement time using:
-      ```bash
-      cargo bench -- --measurement-time 30
-      ```
+### 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)
+- [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)
 
-## Other Experiments (See long-paper version)
-This section explains how to run more benchmarks related to the paper results appearing in the long version.
-At the root of the project, run:
+*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>
+### Documentation
 
-- ```make bench_pbs_asiacrypt```: returns the timings associated to the vanilla PBS. Can be used as another baseline comparison to see the impact of the EBS, SBS and SBS+CMS.
+Full, comprehensive documentation is available here: [https://docs.zama.ai/tfhe-rs](https://docs.zama.ai/tfhe-rs).
+<p align="right">
+  <a href="#about" > ↑ Back to top </a> 
+</p>
 
-- ```make bench_ly23_parallelized```: returns the timings associated to the parallelized version of the extended PBS (EBS) from LY.
-- ```make bench_sorted_parallelized```: returns the timings associated to the parallelized sorted PBS (SBS).
 
-These benchmarks can be used to see the impact of the sorted PBS in a parallelized context.
+## Working with TFHE-rs
 
-As with the previous benchmarks, the launched experiment can be modified to only be performed on the wanted precision and failure probabilities by changing lines 1619 to 1628 for the PBS and lines 1630 to 1663 for the parallelized version.
+### Disclaimers
+
+#### Security estimation
+
+Security estimations are done using the
+[Lattice Estimator](https://github.com/malb/lattice-estimator)
+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
+
+Mitigation for side-channel attacks has not yet been implemented in TFHE-rs,
+and will be released in upcoming versions.
+<br></br>
+
+### Citations
+To cite TFHE-rs in academic papers, please use the following entry:
+
+```text
+@Misc{TFHE-rs,
+  title={{TFHE-rs: A Pure Rust Implementation of the TFHE Scheme for Boolean and Integer Arithmetics Over Encrypted Data}},
+  author={Zama},
+  year={2022},
+  note={\url{https://github.com/zama-ai/tfhe-rs}},
+}
+```
+
+### Contributing
+
+There are two ways to contribute to TFHE-rs:
+
+- [Open issues](https://github.com/zama-ai/tfhe-rs/issues/new/choose) to report bugs and typos, or to suggest new ideas
+- Request to become an official contributor by emailing [hello@zama.ai](mailto:hello@zama.ai).
+
+Becoming an approved contributor involves signing our Contributor License Agreement (CLA). Only approved contributors can send pull requests, so please make sure to get in touch before you do!
+<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.
+
+<p align="right">
+  <a href="#about" > ↑ Back to top </a> 
+</p>
+
+
+## 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>
+</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. 
+
+<p align="right">
+  <a href="#about" > ↑ Back to top </a> 
+</p>

_typos.toml

@@ -0,0 +1,15 @@
+[default]
+extend-ignore-identifiers-re = [
+    # Related to serialized object
+    "ser",
+    "unser",
+    # Used when dumping tfhe-rs parameters set into Sage format
+    "ND.*",
+    # Related to FHE strings example handling "banana"
+    "ba",
+    "enc_ba",
+    # Example with string replacing "hello" with "herlo"
+    "herlo",
+    # Example in trivium
+    "C9217BA0D762ACA1"
+]

apps/trivium/Cargo.toml

@@ -9,11 +9,11 @@ edition = "2021"
 rayon = { version = "1.7.0"}
 
 [target.'cfg(target_arch = "x86_64")'.dependencies.tfhe]
-path = "../../tfhe"
+path = "../../crates/tfhe"
 features = [ "boolean", "shortint", "integer", "x86_64" ]
 
 [target.'cfg(target_arch = "aarch64")'.dependencies.tfhe]
-path = "../../tfhe"
+path = "../../crates/tfhe"
 features = [ "boolean", "shortint", "integer", "aarch64-unix" ]
 
 [dev-dependencies]

apps/trivium/README.md

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

apps/trivium/benches/kreyvium_shortint.rs

@@ -1,5 +1,6 @@
 use criterion::Criterion;
 use tfhe::prelude::*;
+use tfhe::shortint::parameters::PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64;
 use tfhe::shortint::prelude::*;
 use tfhe::{generate_keys, ConfigBuilder, FheUint64};
 use tfhe_trivium::{KreyviumStreamShortint, TransCiphering};
@@ -10,7 +11,8 @@ pub fn kreyvium_shortint_warmup(c: &mut Criterion) {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
+    let (client_key, server_key): (ClientKey, ServerKey) =
+        gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
@@ -60,7 +62,8 @@ pub fn kreyvium_shortint_gen(c: &mut Criterion) {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
+    let (client_key, server_key): (ClientKey, ServerKey) =
+        gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
@@ -105,7 +108,8 @@ pub fn kreyvium_shortint_trans(c: &mut Criterion) {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
+    let (client_key, server_key): (ClientKey, ServerKey) =
+        gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),

apps/trivium/benches/trivium_shortint.rs

@@ -1,5 +1,6 @@
 use criterion::Criterion;
 use tfhe::prelude::*;
+use tfhe::shortint::parameters::PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64;
 use tfhe::shortint::prelude::*;
 use tfhe::{generate_keys, ConfigBuilder, FheUint64};
 use tfhe_trivium::{TransCiphering, TriviumStreamShortint};
@@ -10,7 +11,8 @@ pub fn trivium_shortint_warmup(c: &mut Criterion) {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
+    let (client_key, server_key): (ClientKey, ServerKey) =
+        gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
@@ -60,7 +62,8 @@ pub fn trivium_shortint_gen(c: &mut Criterion) {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
+    let (client_key, server_key): (ClientKey, ServerKey) =
+        gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),
@@ -105,7 +108,8 @@ pub fn trivium_shortint_trans(c: &mut Criterion) {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
+    let (client_key, server_key): (ClientKey, ServerKey) =
+        gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),

apps/trivium/src/kreyvium/kreyvium.rs

@@ -148,10 +148,9 @@ where
 
     /// Computes one turn of the stream, updating registers and outputting the new bit.
     pub fn next_bool(&mut self) -> T {
-        match &self.fhe_key {
-            Some(sk) => set_server_key(sk.clone()),
-            None => (),
-        };
+        if let Some(sk) = &self.fhe_key {
+            set_server_key(sk.clone());
+        }
 
         let [o, a, b, c] = self.get_output_and_values(0);
 
@@ -226,18 +225,12 @@ where
     /// Computes 64 turns of the stream, outputting the 64 bits all at once in a
     /// Vec (first value is oldest, last is newest)
     pub fn next_64(&mut self) -> Vec<T> {
-        match &self.fhe_key {
-            Some(sk) => {
-                rayon::broadcast(|_| set_server_key(sk.clone()));
-            }
-            None => (),
+        if let Some(sk) = &self.fhe_key {
+            rayon::broadcast(|_| set_server_key(sk.clone()));
         }
         let mut values = self.get_64_output_and_values();
-        match &self.fhe_key {
-            Some(_) => {
-                rayon::broadcast(|_| unset_server_key());
-            }
-            None => (),
+        if self.fhe_key.is_some() {
+            rayon::broadcast(|_| unset_server_key());
         }
 
         let mut ret = Vec::<T>::with_capacity(64);

apps/trivium/src/kreyvium/kreyvium_byte.rs

@@ -237,18 +237,12 @@ where
     /// Computes 64 turns of the stream, outputting the 64 bits (in 8 bytes) all at once in a
     /// Vec (first value is oldest, last is newest)
     pub fn next_64(&mut self) -> Vec<T> {
-        match &self.fhe_key {
-            Some(sk) => {
-                rayon::broadcast(|_| set_server_key(sk.clone()));
-            }
-            None => (),
+        if let Some(sk) = &self.fhe_key {
+            rayon::broadcast(|_| set_server_key(sk.clone()));
         }
         let values = self.get_64_output_and_values();
-        match &self.fhe_key {
-            Some(_) => {
-                rayon::broadcast(|_| unset_server_key());
-            }
-            None => (),
+        if self.fhe_key.is_some() {
+            rayon::broadcast(|_| unset_server_key());
         }
 
         let mut bytes = Vec::<T>::with_capacity(8);

apps/trivium/src/kreyvium/test.rs

@@ -1,7 +1,7 @@
 use crate::{KreyviumStream, KreyviumStreamByte, KreyviumStreamShortint, TransCiphering};
 use tfhe::prelude::*;
+use tfhe::shortint::parameters::PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64;
 use tfhe::{generate_keys, ConfigBuilder, FheBool, FheUint64, FheUint8};
-
 // Values for these tests come from the github repo renaud1239/Kreyvium,
 // commit fd6828f68711276c25f55e605935028f5e843f43
 
@@ -221,7 +221,8 @@ fn kreyvium_test_shortint_long() {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
+    let (client_key, server_key): (ClientKey, ServerKey) =
+        gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),

apps/trivium/src/lib.rs

@@ -1,3 +1,5 @@
+#![allow(clippy::too_long_first_doc_paragraph)]
+
 mod static_deque;
 
 mod kreyvium;

apps/trivium/src/trivium/test.rs

@@ -1,7 +1,7 @@
 use crate::{TransCiphering, TriviumStream, TriviumStreamByte, TriviumStreamShortint};
 use tfhe::prelude::*;
+use tfhe::shortint::parameters::PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64;
 use tfhe::{generate_keys, ConfigBuilder, FheBool, FheUint64, FheUint8};
-
 // Values for these tests come from the github repo cantora/avr-crypto-lib, commit 2a5b018,
 // file testvectors/trivium-80.80.test-vectors
 
@@ -357,7 +357,8 @@ fn trivium_test_shortint_long() {
     let underlying_ck: tfhe::shortint::ClientKey = (*hl_client_key.as_ref()).clone().into();
     let underlying_sk: tfhe::shortint::ServerKey = (*hl_server_key.as_ref()).clone().into();
 
-    let (client_key, server_key): (ClientKey, ServerKey) = gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS);
+    let (client_key, server_key): (ClientKey, ServerKey) =
+        gen_keys(PARAM_MESSAGE_1_CARRY_1_KS_PBS_GAUSSIAN_2M64);
 
     let ksk = KeySwitchingKey::new(
         (&client_key, Some(&server_key)),

apps/trivium/src/trivium/trivium_bool.rs

@@ -120,10 +120,9 @@ where
 
     /// Computes one turn of the stream, updating registers and outputting the new bit.
     pub fn next_bool(&mut self) -> T {
-        match &self.fhe_key {
-            Some(sk) => set_server_key(sk.clone()),
-            None => (),
-        };
+        if let Some(sk) = &self.fhe_key {
+            set_server_key(sk.clone());
+        }
 
         let [o, a, b, c] = self.get_output_and_values(0);
 
@@ -196,18 +195,12 @@ where
     /// Computes 64 turns of the stream, outputting the 64 bits all at once in a
     /// Vec (first value is oldest, last is newest)
     pub fn next_64(&mut self) -> Vec<T> {
-        match &self.fhe_key {
-            Some(sk) => {
-                rayon::broadcast(|_| set_server_key(sk.clone()));
-            }
-            None => (),
+        if let Some(sk) = &self.fhe_key {
+            rayon::broadcast(|_| set_server_key(sk.clone()));
         }
         let mut values = self.get_64_output_and_values();
-        match &self.fhe_key {
-            Some(_) => {
-                rayon::broadcast(|_| unset_server_key());
-            }
-            None => (),
+        if self.fhe_key.is_some() {
+            rayon::broadcast(|_| unset_server_key());
         }
 
         let mut ret = Vec::<T>::with_capacity(64);

apps/trivium/src/trivium/trivium_byte.rs

@@ -187,18 +187,12 @@ where
     /// Computes 64 turns of the stream, outputting the 64 bits (in 8 bytes) all at once in a
     /// Vec (first value is oldest, last is newest)
     pub fn next_64(&mut self) -> Vec<T> {
-        match &self.fhe_key {
-            Some(sk) => {
-                rayon::broadcast(|_| set_server_key(sk.clone()));
-            }
-            None => (),
+        if let Some(sk) = &self.fhe_key {
+            rayon::broadcast(|_| set_server_key(sk.clone()));
         }
         let values = self.get_64_output_and_values();
-        match &self.fhe_key {
-            Some(_) => {
-                rayon::broadcast(|_| unset_server_key());
-            }
-            None => (),
+        if self.fhe_key.is_some() {
+            rayon::broadcast(|_| unset_server_key());
         }
 
         let mut bytes = Vec::<T>::with_capacity(8);

backends/tfhe-cuda-backend/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "tfhe-cuda-backend"
-version = "0.4.0-alpha.0"
+version = "0.6.0"
 edition = "2021"
 authors = ["Zama team"]
 license = "BSD-3-Clause-Clear"
@@ -14,3 +14,4 @@ keywords = ["fully", "homomorphic", "encryption", "fhe", "cryptography"]
 [build-dependencies]
 cmake = { version = "0.1" }
 pkg-config = { version = "0.3" }
+bindgen = "0.70.1"

backends/tfhe-cuda-backend/build.rs

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

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

@@ -67,9 +67,21 @@ endif()
 
 add_compile_definitions(CUDA_ARCH=${CUDA_ARCH})
 
+# Check if the DEBUG flag is defined
+if(CMAKE_BUILD_TYPE STREQUAL "Debug")
+  # Debug mode
+  message("Compiling in Debug mode")
+  add_definitions(-DDEBUG)
+  set(OPTIMIZATION_FLAGS "${OPTIMIZATION_FLAGS} -O0 -G -g")
+else()
+  # Release mode
+  message("Compiling in Release mode")
+  set(OPTIMIZATION_FLAGS "${OPTIMIZATION_FLAGS} -O3")
+endif()
+
 # 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 ${CMAKE_CXX_COMPILER} -O3 \
+    "${CMAKE_CUDA_FLAGS} -ccbin ${CMAKE_CXX_COMPILER} ${OPTIMIZATION_FLAGS}\
   -std=c++17 --no-exceptions  --expt-relaxed-constexpr -rdc=true \
   --use_fast_math -Xcompiler -fPIC")
 

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

@@ -1,18 +1,24 @@
 #ifndef CUDA_CIPHERTEXT_H
 #define CUDA_CIPHERTEXT_H
 
-#include <cstdint>
+#include "stdint.h"
 
 extern "C" {
 void cuda_convert_lwe_ciphertext_vector_to_gpu_64(void *stream,
                                                   uint32_t gpu_index,
-                                                  void *dest, void *src,
+                                                  void *dest, void const *src,
                                                   uint32_t number_of_cts,
                                                   uint32_t lwe_dimension);
 void cuda_convert_lwe_ciphertext_vector_to_cpu_64(void *stream,
                                                   uint32_t gpu_index,
-                                                  void *dest, void *src,
+                                                  void *dest, void const *src,
                                                   uint32_t number_of_cts,
                                                   uint32_t lwe_dimension);
-};
+
+void cuda_glwe_sample_extract_64(void *stream, uint32_t gpu_index,
+                                 void *lwe_array_out, void const *glwe_array_in,
+                                 uint32_t const *nth_array, uint32_t num_nths,
+                                 uint32_t glwe_dimension,
+                                 uint32_t polynomial_size);
+}
 #endif

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

@@ -27,28 +27,38 @@ inline void cuda_error(cudaError_t code, const char *file, int line) {
     std::abort();                                                              \
   }
 
+cudaEvent_t cuda_create_event(uint32_t gpu_index);
+
+void cuda_event_record(cudaEvent_t event, cudaStream_t stream,
+                       uint32_t gpu_index);
+void cuda_stream_wait_event(cudaStream_t stream, cudaEvent_t event,
+                            uint32_t gpu_index);
+
+void cuda_event_destroy(cudaEvent_t event, uint32_t gpu_index);
+
 cudaStream_t cuda_create_stream(uint32_t gpu_index);
 
 void cuda_destroy_stream(cudaStream_t stream, uint32_t gpu_index);
 
 void cuda_synchronize_stream(cudaStream_t stream, uint32_t gpu_index);
 
+uint32_t cuda_is_available();
+
 void *cuda_malloc(uint64_t size, uint32_t gpu_index);
 
 void *cuda_malloc_async(uint64_t size, cudaStream_t stream, uint32_t gpu_index);
 
 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);
 
-void cuda_memcpy_async_gpu_to_gpu(void *dest, void *src, uint64_t size,
+void cuda_memcpy_async_gpu_to_gpu(void *dest, void const *src, uint64_t size,
                                   cudaStream_t stream, uint32_t gpu_index);
 
+void cuda_memcpy_gpu_to_gpu(void *dest, void *src, uint64_t size,
+                            uint32_t gpu_index);
+
 void cuda_memcpy_async_to_cpu(void *dest, const void *src, uint64_t size,
                               cudaStream_t stream, uint32_t gpu_index);
 
@@ -62,9 +72,13 @@ 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);
+}
 
 int cuda_get_max_shared_memory(uint32_t gpu_index);
-}
+
+bool cuda_check_support_cooperative_groups();
+
+bool cuda_check_support_thread_block_clusters();
 
 template <typename Torus>
 void cuda_set_value_async(cudaStream_t stream, uint32_t gpu_index,

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

@@ -8,7 +8,7 @@ extern std::mutex m;
 extern bool p2p_enabled;
 
 extern "C" {
-int cuda_setup_multi_gpu();
+int32_t cuda_setup_multi_gpu();
 }
 
 // Define a variant type that can be either a vector or a single pointer

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

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

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

@@ -0,0 +1,123 @@
+#ifndef CUDA_INTEGER_COMPRESSION_UTILITIES_H
+#define CUDA_INTEGER_COMPRESSION_UTILITIES_H
+
+#include "../integer_utilities.h"
+
+template <typename Torus> struct int_compression {
+  int_radix_params compression_params;
+  uint32_t storage_log_modulus;
+  uint32_t lwe_per_glwe;
+
+  uint32_t body_count;
+
+  // Compression
+  int8_t *fp_ks_buffer;
+  Torus *tmp_lwe;
+  Torus *tmp_glwe_array_out;
+
+  int_compression(cudaStream_t const *streams, uint32_t const *gpu_indexes,
+                  uint32_t gpu_count, int_radix_params compression_params,
+                  uint32_t num_radix_blocks, uint32_t lwe_per_glwe,
+                  uint32_t storage_log_modulus, bool allocate_gpu_memory) {
+    this->compression_params = compression_params;
+    this->lwe_per_glwe = lwe_per_glwe;
+    this->storage_log_modulus = storage_log_modulus;
+    this->body_count = num_radix_blocks;
+
+    if (allocate_gpu_memory) {
+      Torus glwe_accumulator_size = (compression_params.glwe_dimension + 1) *
+                                    compression_params.polynomial_size;
+
+      tmp_lwe = (Torus *)cuda_malloc_async(
+          num_radix_blocks * (compression_params.small_lwe_dimension + 1) *
+              sizeof(Torus),
+          streams[0], gpu_indexes[0]);
+      tmp_glwe_array_out = (Torus *)cuda_malloc_async(
+          lwe_per_glwe * glwe_accumulator_size * sizeof(Torus), streams[0],
+          gpu_indexes[0]);
+
+      scratch_packing_keyswitch_lwe_list_to_glwe_64(
+          streams[0], gpu_indexes[0], &fp_ks_buffer,
+          compression_params.glwe_dimension, compression_params.polynomial_size,
+          num_radix_blocks, true);
+    }
+  }
+  void release(cudaStream_t const *streams, uint32_t const *gpu_indexes,
+               uint32_t gpu_count) {
+    cuda_drop_async(tmp_lwe, streams[0], gpu_indexes[0]);
+    cuda_drop_async(tmp_glwe_array_out, streams[0], gpu_indexes[0]);
+    cleanup_packing_keyswitch_lwe_list_to_glwe(streams[0], gpu_indexes[0],
+                                               &fp_ks_buffer);
+  }
+};
+
+template <typename Torus> struct int_decompression {
+  int_radix_params encryption_params;
+  int_radix_params compression_params;
+
+  uint32_t storage_log_modulus;
+
+  uint32_t num_radix_blocks;
+  uint32_t body_count;
+
+  Torus *tmp_extracted_glwe;
+  Torus *tmp_extracted_lwe;
+  uint32_t *tmp_indexes_array;
+
+  int_radix_lut<Torus> *carry_extract_lut;
+
+  int_decompression(cudaStream_t const *streams, uint32_t const *gpu_indexes,
+                    uint32_t gpu_count, int_radix_params encryption_params,
+                    int_radix_params compression_params,
+                    uint32_t num_radix_blocks, uint32_t body_count,
+                    uint32_t storage_log_modulus, bool allocate_gpu_memory) {
+    this->encryption_params = encryption_params;
+    this->compression_params = compression_params;
+    this->storage_log_modulus = storage_log_modulus;
+    this->num_radix_blocks = num_radix_blocks;
+    this->body_count = body_count;
+
+    if (allocate_gpu_memory) {
+      Torus glwe_accumulator_size = (compression_params.glwe_dimension + 1) *
+                                    compression_params.polynomial_size;
+      Torus lwe_accumulator_size = (compression_params.glwe_dimension *
+                                        compression_params.polynomial_size +
+                                    1);
+      carry_extract_lut = new int_radix_lut<Torus>(
+          streams, gpu_indexes, gpu_count, encryption_params, 1,
+          num_radix_blocks, allocate_gpu_memory);
+
+      tmp_extracted_glwe = (Torus *)cuda_malloc_async(
+          num_radix_blocks * glwe_accumulator_size * sizeof(Torus), streams[0],
+          gpu_indexes[0]);
+      tmp_indexes_array = (uint32_t *)cuda_malloc_async(
+          num_radix_blocks * sizeof(uint32_t), streams[0], gpu_indexes[0]);
+      tmp_extracted_lwe = (Torus *)cuda_malloc_async(
+          num_radix_blocks * lwe_accumulator_size * sizeof(Torus), streams[0],
+          gpu_indexes[0]);
+
+      // Carry extract LUT
+      auto carry_extract_f = [encryption_params](Torus x) -> Torus {
+        return x / encryption_params.message_modulus;
+      };
+
+      generate_device_accumulator<Torus>(
+          streams[0], gpu_indexes[0], carry_extract_lut->get_lut(0, 0),
+          encryption_params.glwe_dimension, encryption_params.polynomial_size,
+          encryption_params.message_modulus, encryption_params.carry_modulus,
+          carry_extract_f);
+
+      carry_extract_lut->broadcast_lut(streams, gpu_indexes, 0);
+    }
+  }
+  void release(cudaStream_t const *streams, uint32_t const *gpu_indexes,
+               uint32_t gpu_count) {
+    cuda_drop_async(tmp_extracted_glwe, streams[0], gpu_indexes[0]);
+    cuda_drop_async(tmp_extracted_lwe, streams[0], gpu_indexes[0]);
+    cuda_drop_async(tmp_indexes_array, streams[0], gpu_indexes[0]);
+
+    carry_extract_lut->release(streams, gpu_indexes, gpu_count);
+    delete carry_extract_lut;
+  }
+};
+#endif

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

@@ -0,0 +1,444 @@
+#ifndef CUDA_INTEGER_H
+#define CUDA_INTEGER_H
+
+#include "../pbs/pbs_enums.h"
+#include <stdint.h>
+
+enum OUTPUT_CARRY { NONE = 0, GENERATED = 1, PROPAGATED = 2 };
+enum SHIFT_OR_ROTATE_TYPE {
+  LEFT_SHIFT = 0,
+  RIGHT_SHIFT = 1,
+  LEFT_ROTATE = 2,
+  RIGHT_ROTATE = 3
+};
+enum BITOP_TYPE {
+  BITAND = 0,
+  BITOR = 1,
+  BITXOR = 2,
+  SCALAR_BITAND = 3,
+  SCALAR_BITOR = 4,
+  SCALAR_BITXOR = 5,
+};
+
+enum COMPARISON_TYPE {
+  EQ = 0,
+  NE = 1,
+  GT = 2,
+  GE = 3,
+  LT = 4,
+  LE = 5,
+  MAX = 6,
+  MIN = 7,
+};
+
+enum CMP_ORDERING { IS_INFERIOR = 0, IS_EQUAL = 1, IS_SUPERIOR = 2 };
+
+enum SIGNED_OPERATION { ADDITION = 1, SUBTRACTION = -1 };
+
+enum outputFlag { FLAG_NONE = 0, FLAG_OVERFLOW = 1, FLAG_CARRY = 2 };
+
+extern "C" {
+void scratch_cuda_apply_univariate_lut_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, void const *input_lut, uint32_t lwe_dimension,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t ks_level,
+    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t grouping_factor, uint32_t input_lwe_ciphertext_count,
+    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
+    bool allocate_gpu_memory);
+
+void cuda_apply_univariate_lut_kb_64(void *const *streams,
+                                     uint32_t const *gpu_indexes,
+                                     uint32_t gpu_count, void *output_radix_lwe,
+                                     void const *input_radix_lwe,
+                                     int8_t *mem_ptr, void *const *ksks,
+                                     void *const *bsks, uint32_t num_blocks);
+
+void cleanup_cuda_apply_univariate_lut_kb_64(void *const *streams,
+                                             uint32_t const *gpu_indexes,
+                                             uint32_t gpu_count,
+                                             int8_t **mem_ptr_void);
+
+void scratch_cuda_apply_bivariate_lut_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, void const *input_lut, uint32_t lwe_dimension,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t ks_level,
+    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t grouping_factor, uint32_t input_lwe_ciphertext_count,
+    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
+    bool allocate_gpu_memory);
+
+void cuda_apply_bivariate_lut_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *output_radix_lwe, void const *input_radix_lwe_1,
+    void const *input_radix_lwe_2, int8_t *mem_ptr, void *const *ksks,
+    void *const *bsks, uint32_t num_blocks, uint32_t shift);
+
+void cleanup_cuda_apply_bivariate_lut_kb_64(void *const *streams,
+                                            uint32_t const *gpu_indexes,
+                                            uint32_t gpu_count,
+                                            int8_t **mem_ptr_void);
+
+void cuda_apply_many_univariate_lut_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *output_radix_lwe, void const *input_radix_lwe, int8_t *mem_ptr,
+    void *const *ksks, void *const *bsks, uint32_t num_blocks,
+    uint32_t num_luts, uint32_t lut_stride);
+
+void scratch_cuda_full_propagation_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, uint32_t lwe_dimension, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t ks_level, uint32_t ks_base_log,
+    uint32_t pbs_level, uint32_t pbs_base_log, uint32_t grouping_factor,
+    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
+    bool allocate_gpu_memory);
+
+void cuda_full_propagation_64_inplace(void *const *streams,
+                                      uint32_t const *gpu_indexes,
+                                      uint32_t gpu_count, void *input_blocks,
+                                      int8_t *mem_ptr, void *const *ksks,
+                                      void *const *bsks, uint32_t num_blocks);
+
+void cleanup_cuda_full_propagation(void *const *streams,
+                                   uint32_t const *gpu_indexes,
+                                   uint32_t gpu_count, int8_t **mem_ptr_void);
+
+void scratch_cuda_integer_mult_radix_ciphertext_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, bool const is_boolean_left, bool const is_boolean_right,
+    uint32_t message_modulus, uint32_t carry_modulus, uint32_t glwe_dimension,
+    uint32_t lwe_dimension, uint32_t polynomial_size, uint32_t pbs_base_log,
+    uint32_t pbs_level, uint32_t ks_base_log, uint32_t ks_level,
+    uint32_t grouping_factor, uint32_t num_blocks, PBS_TYPE pbs_type,
+    bool allocate_gpu_memory);
+
+void cuda_integer_mult_radix_ciphertext_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *radix_lwe_out, void const *radix_lwe_left, bool const is_bool_left,
+    void const *radix_lwe_right, bool const is_bool_right, void *const *bsks,
+    void *const *ksks, int8_t *mem_ptr, uint32_t polynomial_size,
+    uint32_t num_blocks);
+
+void cleanup_cuda_integer_mult(void *const *streams,
+                               uint32_t const *gpu_indexes, uint32_t gpu_count,
+                               int8_t **mem_ptr_void);
+
+void cuda_negate_integer_radix_ciphertext_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array_out, void const *lwe_array_in, uint32_t lwe_dimension,
+    uint32_t lwe_ciphertext_count, uint32_t message_modulus,
+    uint32_t carry_modulus);
+
+void cuda_scalar_addition_integer_radix_ciphertext_64_inplace(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array, void const *scalar_input, uint32_t lwe_dimension,
+    uint32_t lwe_ciphertext_count, uint32_t message_modulus,
+    uint32_t carry_modulus);
+
+void scratch_cuda_integer_radix_logical_scalar_shift_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
+    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
+    uint32_t carry_modulus, PBS_TYPE pbs_type, SHIFT_OR_ROTATE_TYPE shift_type,
+    bool allocate_gpu_memory);
+
+void cuda_integer_radix_logical_scalar_shift_kb_64_inplace(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array, uint32_t shift, int8_t *mem_ptr, void *const *bsks,
+    void *const *ksks, uint32_t num_blocks);
+
+void scratch_cuda_integer_radix_arithmetic_scalar_shift_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
+    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
+    uint32_t carry_modulus, PBS_TYPE pbs_type, SHIFT_OR_ROTATE_TYPE shift_type,
+    bool allocate_gpu_memory);
+
+void cuda_integer_radix_arithmetic_scalar_shift_kb_64_inplace(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array, uint32_t shift, int8_t *mem_ptr, void *const *bsks,
+    void *const *ksks, uint32_t num_blocks);
+
+void cleanup_cuda_integer_radix_logical_scalar_shift(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr_void);
+
+void cleanup_cuda_integer_radix_arithmetic_scalar_shift(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr_void);
+
+void scratch_cuda_integer_radix_shift_and_rotate_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
+    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
+    uint32_t carry_modulus, PBS_TYPE pbs_type, SHIFT_OR_ROTATE_TYPE shift_type,
+    bool is_signed, bool allocate_gpu_memory);
+
+void cuda_integer_radix_shift_and_rotate_kb_64_inplace(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array, void const *lwe_shift, int8_t *mem_ptr, void *const *bsks,
+    void *const *ksks, uint32_t num_blocks);
+
+void cleanup_cuda_integer_radix_shift_and_rotate(void *const *streams,
+                                                 uint32_t const *gpu_indexes,
+                                                 uint32_t gpu_count,
+                                                 int8_t **mem_ptr_void);
+
+void scratch_cuda_integer_radix_comparison_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
+    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t grouping_factor, uint32_t lwe_ciphertext_count,
+    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
+    COMPARISON_TYPE op_type, bool is_signed, bool allocate_gpu_memory);
+
+void cuda_comparison_integer_radix_ciphertext_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array_out, void const *lwe_array_1, void const *lwe_array_2,
+    int8_t *mem_ptr, void *const *bsks, void *const *ksks,
+    uint32_t lwe_ciphertext_count);
+
+void cuda_scalar_comparison_integer_radix_ciphertext_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array_out, void const *lwe_array_in, void const *scalar_blocks,
+    int8_t *mem_ptr, void *const *bsks, void *const *ksks,
+    uint32_t lwe_ciphertext_count, uint32_t num_scalar_blocks);
+
+void cleanup_cuda_integer_comparison(void *const *streams,
+                                     uint32_t const *gpu_indexes,
+                                     uint32_t gpu_count, int8_t **mem_ptr_void);
+
+void scratch_cuda_integer_radix_bitop_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
+    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t grouping_factor, uint32_t lwe_ciphertext_count,
+    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
+    BITOP_TYPE op_type, bool allocate_gpu_memory);
+
+void cuda_bitop_integer_radix_ciphertext_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array_out, void const *lwe_array_1, void const *lwe_array_2,
+    int8_t *mem_ptr, void *const *bsks, void *const *ksks,
+    uint32_t lwe_ciphertext_count);
+
+void cuda_scalar_bitop_integer_radix_ciphertext_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array_out, void const *lwe_array_input, void const *clear_blocks,
+    uint32_t num_clear_blocks, int8_t *mem_ptr, void *const *bsks,
+    void *const *ksks, uint32_t lwe_ciphertext_count, BITOP_TYPE op);
+
+void cleanup_cuda_integer_bitop(void *const *streams,
+                                uint32_t const *gpu_indexes, uint32_t gpu_count,
+                                int8_t **mem_ptr_void);
+
+void scratch_cuda_integer_radix_cmux_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
+    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t grouping_factor, uint32_t lwe_ciphertext_count,
+    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
+    bool allocate_gpu_memory);
+
+void cuda_cmux_integer_radix_ciphertext_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array_out, void const *lwe_condition, void const *lwe_array_true,
+    void const *lwe_array_false, int8_t *mem_ptr, void *const *bsks,
+    void *const *ksks, uint32_t lwe_ciphertext_count);
+
+void cleanup_cuda_integer_radix_cmux(void *const *streams,
+                                     uint32_t const *gpu_indexes,
+                                     uint32_t gpu_count, int8_t **mem_ptr_void);
+
+void scratch_cuda_integer_radix_scalar_rotate_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
+    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
+    uint32_t carry_modulus, PBS_TYPE pbs_type, SHIFT_OR_ROTATE_TYPE shift_type,
+    bool allocate_gpu_memory);
+
+void cuda_integer_radix_scalar_rotate_kb_64_inplace(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array, uint32_t n, int8_t *mem_ptr, void *const *bsks,
+    void *const *ksks, uint32_t num_blocks);
+
+void cleanup_cuda_integer_radix_scalar_rotate(void *const *streams,
+                                              uint32_t const *gpu_indexes,
+                                              uint32_t gpu_count,
+                                              int8_t **mem_ptr_void);
+
+void scratch_cuda_propagate_single_carry_kb_64_inplace(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
+    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
+    uint32_t carry_modulus, PBS_TYPE pbs_type, uint32_t requested_flag,
+    uint32_t uses_carry, bool allocate_gpu_memory);
+
+void scratch_cuda_add_and_propagate_single_carry_kb_64_inplace(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
+    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
+    uint32_t carry_modulus, PBS_TYPE pbs_type, uint32_t requested_flag,
+    uint32_t uses_carry, bool allocate_gpu_memory);
+
+void cuda_propagate_single_carry_kb_64_inplace(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array, void *carry_out, const void *carry_in, int8_t *mem_ptr,
+    void *const *bsks, void *const *ksks, uint32_t num_blocks,
+    uint32_t requested_flag, uint32_t uses_carry);
+
+void cuda_add_and_propagate_single_carry_kb_64_inplace(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lhs_array, const void *rhs_array, void *carry_out,
+    const void *carry_in, int8_t *mem_ptr, void *const *bsks, void *const *ksks,
+    uint32_t num_blocks, uint32_t requested_flag, uint32_t uses_carry);
+
+void cleanup_cuda_propagate_single_carry(void *const *streams,
+                                         uint32_t const *gpu_indexes,
+                                         uint32_t gpu_count,
+                                         int8_t **mem_ptr_void);
+
+void cleanup_cuda_add_and_propagate_single_carry(void *const *streams,
+                                                 uint32_t const *gpu_indexes,
+                                                 uint32_t gpu_count,
+                                                 int8_t **mem_ptr_void);
+
+void scratch_cuda_integer_overflowing_sub_kb_64_inplace(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t big_lwe_dimension, uint32_t small_lwe_dimension, uint32_t ks_level,
+    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t grouping_factor, uint32_t num_blocks, uint32_t message_modulus,
+    uint32_t carry_modulus, PBS_TYPE pbs_type, uint32_t compute_overflow,
+    bool allocate_gpu_memory);
+
+void cuda_integer_overflowing_sub_kb_64_inplace(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lhs_array, const void *rhs_array, void *overflow_block,
+    const void *input_borrow, int8_t *mem_ptr, void *const *bsks,
+    void *const *ksks, uint32_t num_blocks, uint32_t compute_overflow,
+    uint32_t uses_input_borrow);
+
+void cleanup_cuda_integer_overflowing_sub(void *const *streams,
+                                          uint32_t const *gpu_indexes,
+                                          uint32_t gpu_count,
+                                          int8_t **mem_ptr_void);
+
+void scratch_cuda_integer_radix_partial_sum_ciphertexts_vec_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t lwe_dimension, uint32_t ks_level, uint32_t ks_base_log,
+    uint32_t pbs_level, uint32_t pbs_base_log, uint32_t grouping_factor,
+    uint32_t num_blocks_in_radix, uint32_t max_num_radix_in_vec,
+    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
+    bool allocate_gpu_memory);
+
+void cuda_integer_radix_partial_sum_ciphertexts_vec_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *radix_lwe_out, void *radix_lwe_vec, uint32_t num_radix_in_vec,
+    int8_t *mem_ptr, void *const *bsks, void *const *ksks,
+    uint32_t num_blocks_in_radix);
+
+void cleanup_cuda_integer_radix_partial_sum_ciphertexts_vec(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr_void);
+
+void scratch_cuda_integer_scalar_mul_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, uint32_t glwe_dimension, uint32_t polynomial_size,
+    uint32_t lwe_dimension, uint32_t ks_level, uint32_t ks_base_log,
+    uint32_t pbs_level, uint32_t pbs_base_log, uint32_t grouping_factor,
+    uint32_t num_blocks, uint32_t message_modulus, uint32_t carry_modulus,
+    PBS_TYPE pbs_type, bool allocate_gpu_memory);
+
+void cuda_scalar_multiplication_integer_radix_ciphertext_64_inplace(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *lwe_array, uint64_t const *decomposed_scalar,
+    uint64_t const *has_at_least_one_set, int8_t *mem_ptr, void *const *bsks,
+    void *const *ksks, uint32_t lwe_dimension, uint32_t polynomial_size,
+    uint32_t message_modulus, uint32_t num_blocks, uint32_t num_scalars);
+
+void cleanup_cuda_integer_radix_scalar_mul(void *const *streams,
+                                           uint32_t const *gpu_indexes,
+                                           uint32_t gpu_count,
+                                           int8_t **mem_ptr_void);
+
+void scratch_cuda_integer_div_rem_radix_ciphertext_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    bool is_signed, int8_t **mem_ptr, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t big_lwe_dimension,
+    uint32_t small_lwe_dimension, uint32_t ks_level, uint32_t ks_base_log,
+    uint32_t pbs_level, uint32_t pbs_base_log, uint32_t grouping_factor,
+    uint32_t num_blocks, uint32_t message_modulus, uint32_t carry_modulus,
+    PBS_TYPE pbs_type, bool allocate_gpu_memory);
+
+void cuda_integer_div_rem_radix_ciphertext_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *quotient, void *remainder, void const *numerator, void const *divisor,
+    bool is_signed, int8_t *mem_ptr, void *const *bsks, void *const *ksks,
+    uint32_t num_blocks_in_radix);
+
+void cleanup_cuda_integer_div_rem(void *const *streams,
+                                  uint32_t const *gpu_indexes,
+                                  uint32_t gpu_count, int8_t **mem_ptr_void);
+
+void scratch_cuda_integer_compute_prefix_sum_hillis_steele_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, void const *input_lut, uint32_t lwe_dimension,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t ks_level,
+    uint32_t ks_base_log, uint32_t pbs_level, uint32_t pbs_base_log,
+    uint32_t grouping_factor, uint32_t num_radix_blocks,
+    uint32_t message_modulus, uint32_t carry_modulus, PBS_TYPE pbs_type,
+    bool allocate_gpu_memory);
+
+void cuda_integer_compute_prefix_sum_hillis_steele_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *output_radix_lwe, void *generates_or_propagates, int8_t *mem_ptr,
+    void *const *ksks, void *const *bsks, uint32_t num_blocks, uint32_t shift);
+
+void cleanup_cuda_integer_compute_prefix_sum_hillis_steele_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr_void);
+
+void cuda_integer_reverse_blocks_64_inplace(void *const *streams,
+                                            uint32_t const *gpu_indexes,
+                                            uint32_t gpu_count, void *lwe_array,
+                                            uint32_t num_blocks,
+                                            uint32_t lwe_size);
+
+void scratch_cuda_integer_abs_inplace_radix_ciphertext_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    int8_t **mem_ptr, bool is_signed, uint32_t glwe_dimension,
+    uint32_t polynomial_size, uint32_t big_lwe_dimension,
+    uint32_t small_lwe_dimension, uint32_t ks_level, uint32_t ks_base_log,
+    uint32_t pbs_level, uint32_t pbs_base_log, uint32_t grouping_factor,
+    uint32_t num_blocks, uint32_t message_modulus, uint32_t carry_modulus,
+    PBS_TYPE pbs_type, bool allocate_gpu_memory);
+
+void cuda_integer_abs_inplace_radix_ciphertext_kb_64(
+    void *const *streams, uint32_t const *gpu_indexes, uint32_t gpu_count,
+    void *ct, int8_t *mem_ptr, bool is_signed, void *const *bsks,
+    void *const *ksks, uint32_t num_blocks);
+
+void cleanup_cuda_integer_abs_inplace(void *const *streams,
+                                      uint32_t const *gpu_indexes,
+                                      uint32_t gpu_count,
+                                      int8_t **mem_ptr_void);
+
+} // extern C
+#endif // CUDA_INTEGER_H

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

@@ -1,21 +1,39 @@
 #ifndef CNCRT_KS_H_
 #define CNCRT_KS_H_
 
-#include <cstdint>
+#include <stdint.h>
 
 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);
+    void const *lwe_output_indexes, void const *lwe_array_in,
+    void const *lwe_input_indexes, void const *ksk, uint32_t lwe_dimension_in,
+    uint32_t lwe_dimension_out, uint32_t base_log, uint32_t level_count,
+    uint32_t num_samples);
 
 void 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);
+    void const *lwe_output_indexes, void const *lwe_array_in,
+    void const *lwe_input_indexes, void const *ksk, uint32_t lwe_dimension_in,
+    uint32_t lwe_dimension_out, uint32_t base_log, uint32_t level_count,
+    uint32_t num_samples);
+
+void scratch_packing_keyswitch_lwe_list_to_glwe_64(
+    void *stream, uint32_t gpu_index, int8_t **fp_ks_buffer,
+    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t num_lwes,
+    bool allocate_gpu_memory);
+
+void cuda_packing_keyswitch_lwe_list_to_glwe_64(
+    void *stream, uint32_t gpu_index, void *glwe_array_out,
+    void const *lwe_array_in, void const *fp_ksk_array, int8_t *fp_ks_buffer,
+    uint32_t input_lwe_dimension, uint32_t output_glwe_dimension,
+    uint32_t output_polynomial_size, uint32_t base_log, uint32_t level_count,
+    uint32_t num_lwes);
+
+void cleanup_packing_keyswitch_lwe_list_to_glwe(void *stream,
+                                                uint32_t gpu_index,
+                                                int8_t **fp_ks_buffer);
 }
 
 #endif // CNCRT_KS_H_

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

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

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

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

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

@@ -1,164 +1,112 @@
-#ifndef CUDA_MULTI_BIT_H
-#define CUDA_MULTI_BIT_H
+#ifndef CUDA_MULTI_BIT_UTILITIES_H
+#define CUDA_MULTI_BIT_UTILITIES_H
 
-#include "programmable_bootstrap.h"
-#include <cstdint>
-
-extern "C" {
-
-bool has_support_to_cuda_programmable_bootstrap_cg_multi_bit(
-    uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t level_count,
-    uint32_t num_samples, uint32_t max_shared_memory);
-
-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, uint32_t max_shared_memory,
-    bool allocate_gpu_memory, uint32_t chunk_size = 0);
-
-void cuda_multi_bit_programmable_bootstrap_lwe_ciphertext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void *lwe_output_indexes, void *lut_vector, void *lut_vector_indexes,
-    void *lwe_array_in, void *lwe_input_indexes, void *bootstrapping_key,
-    int8_t *buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t grouping_factor, uint32_t base_log,
-    uint32_t level_count, uint32_t num_samples, uint32_t num_luts,
-    uint32_t lwe_idx, uint32_t max_shared_memory, uint32_t lwe_chunk_size = 0);
-
-void cleanup_cuda_multi_bit_programmable_bootstrap(void *stream,
-                                                   uint32_t gpu_index,
-                                                   int8_t **pbs_buffer);
-}
+#include "pbs_utilities.h"
 
 template <typename Torus>
-__host__ bool
-supports_distributed_shared_memory_on_multibit_programmable_bootstrap(
-    uint32_t polynomial_size, uint32_t max_shared_memory);
+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, uint32_t max_shared_memory);
+    uint32_t level_count);
 
 #if CUDA_ARCH >= 900
-template <typename Torus, typename STorus>
+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, uint32_t max_shared_memory,
-    bool allocate_gpu_memory, uint32_t lwe_chunk_size);
+    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_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,
+    Torus const *lwe_output_indexes, Torus const *lut_vector,
+    Torus const *lut_vector_indexes, Torus const *lwe_array_in,
+    Torus const *lwe_input_indexes, Torus const *bootstrapping_key,
     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);
+    uint32_t num_many_lut, uint32_t lut_stride);
 #endif
 
-template <typename Torus, typename STorus>
-void scratch_cuda_cg_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, uint32_t max_shared_memory,
-    bool allocate_gpu_memory, uint32_t lwe_chunk_size = 0);
-
-template <typename Torus, typename STorus>
+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, uint32_t max_shared_memory,
-    bool allocate_gpu_memory, uint32_t lwe_chunk_size = 0);
+    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,
+    Torus const *lwe_output_indexes, Torus const *lut_vector,
+    Torus const *lut_vector_indexes, Torus const *lwe_array_in,
+    Torus const *lwe_input_indexes, Torus const *bootstrapping_key,
     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);
+    uint32_t num_many_lut, uint32_t lut_stride);
 
-template <typename Torus, typename STorus>
+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, uint32_t max_shared_memory,
-    bool allocate_gpu_memory, uint32_t lwe_chunk_size = 0);
+    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_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,
+    Torus const *lwe_output_indexes, Torus const *lut_vector,
+    Torus const *lut_vector_indexes, Torus const *lwe_array_in,
+    Torus const *lwe_input_indexes, Torus const *bootstrapping_key,
     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);
+    uint32_t num_many_lut, uint32_t lut_stride);
 
 template <typename Torus>
-__host__ __device__ uint64_t
-get_buffer_size_full_sm_multibit_programmable_bootstrap_keybundle(
+uint64_t get_buffer_size_full_sm_multibit_programmable_bootstrap_keybundle(
     uint32_t polynomial_size);
 template <typename Torus>
-__host__ __device__ uint64_t
-get_buffer_size_full_sm_multibit_programmable_bootstrap_step_one(
+uint64_t get_buffer_size_full_sm_multibit_programmable_bootstrap_step_one(
     uint32_t polynomial_size);
 template <typename Torus>
-__host__ __device__ uint64_t
-get_buffer_size_full_sm_multibit_programmable_bootstrap_step_two(
+uint64_t get_buffer_size_full_sm_multibit_programmable_bootstrap_step_two(
     uint32_t polynomial_size);
 template <typename Torus>
-__host__ __device__ uint64_t
-get_buffer_size_partial_sm_multibit_programmable_bootstrap_step_one(
+uint64_t get_buffer_size_partial_sm_multibit_programmable_bootstrap_step_one(
     uint32_t polynomial_size);
 template <typename Torus>
-__host__ __device__ uint64_t
-get_buffer_size_full_sm_cg_multibit_programmable_bootstrap(
+uint64_t get_buffer_size_full_sm_cg_multibit_programmable_bootstrap(
     uint32_t polynomial_size);
 template <typename Torus>
-__host__ __device__ uint64_t
-get_buffer_size_partial_sm_cg_multibit_programmable_bootstrap(
+uint64_t get_buffer_size_partial_sm_cg_multibit_programmable_bootstrap(
     uint32_t polynomial_size);
 template <typename Torus>
-__host__ __device__ uint64_t
-get_buffer_size_sm_dsm_plus_tbc_multibit_programmable_bootstrap(
+uint64_t get_buffer_size_sm_dsm_plus_tbc_multibit_programmable_bootstrap(
     uint32_t polynomial_size);
 template <typename Torus>
-__host__ __device__ uint64_t
-get_buffer_size_partial_sm_tbc_multibit_programmable_bootstrap(
+uint64_t get_buffer_size_partial_sm_tbc_multibit_programmable_bootstrap(
     uint32_t polynomial_size);
 template <typename Torus>
-__host__ __device__ uint64_t
-get_buffer_size_full_sm_tbc_multibit_programmable_bootstrap(
+uint64_t get_buffer_size_full_sm_tbc_multibit_programmable_bootstrap(
     uint32_t polynomial_size);
 
+template <typename Torus, class params>
+uint32_t get_lwe_chunk_size(uint32_t gpu_index, uint32_t max_num_pbs,
+                            uint32_t polynomial_size);
+
 template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::MULTI_BIT> {
   int8_t *d_mem_keybundle = NULL;
   int8_t *d_mem_acc_step_one = NULL;
   int8_t *d_mem_acc_step_two = NULL;
   int8_t *d_mem_acc_cg = NULL;
   int8_t *d_mem_acc_tbc = NULL;
-
+  uint32_t lwe_chunk_size;
   double2 *keybundle_fft;
   Torus *global_accumulator;
-  double2 *global_accumulator_fft;
+  double2 *global_join_buffer;
 
   PBS_VARIANT pbs_variant;
 
@@ -167,6 +115,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::MULTI_BIT> {
              uint32_t input_lwe_ciphertext_count, uint32_t lwe_chunk_size,
              PBS_VARIANT pbs_variant, bool allocate_gpu_memory) {
     this->pbs_variant = pbs_variant;
+    this->lwe_chunk_size = lwe_chunk_size;
     auto max_shared_memory = cuda_get_max_shared_memory(gpu_index);
 
     // default
@@ -276,10 +225,12 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::MULTI_BIT> {
           num_blocks_keybundle * (polynomial_size / 2) * sizeof(double2),
           stream, gpu_index);
       global_accumulator = (Torus *)cuda_malloc_async(
-          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),
+          input_lwe_ciphertext_count * (glwe_dimension + 1) * polynomial_size *
+              sizeof(Torus),
+          stream, gpu_index);
+      global_join_buffer = (double2 *)cuda_malloc_async(
+          level_count * (glwe_dimension + 1) * input_lwe_ciphertext_count *
+              (polynomial_size / 2) * sizeof(double2),
           stream, gpu_index);
     }
   }
@@ -311,13 +262,8 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::MULTI_BIT> {
 
     cuda_drop_async(keybundle_fft, stream, gpu_index);
     cuda_drop_async(global_accumulator, stream, gpu_index);
-    cuda_drop_async(global_accumulator_fft, stream, gpu_index);
+    cuda_drop_async(global_join_buffer, stream, gpu_index);
   }
 };
 
-template <typename Torus, class params>
-__host__ uint32_t get_lwe_chunk_size(uint32_t gpu_index, uint32_t max_num_pbs,
-                                     uint32_t polynomial_size,
-                                     uint32_t max_shared_memory);
-
-#endif // CUDA_MULTI_BIT_H
+#endif // CUDA_MULTI_BIT_UTILITIES_H

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

@@ -1,127 +1,32 @@
-#ifndef CUDA_BOOTSTRAP_H
-#define CUDA_BOOTSTRAP_H
+#ifndef CUDA_BOOTSTRAP_UTILITIES_H
+#define CUDA_BOOTSTRAP_UTILITIES_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, uint32_t max_shared_memory,
-    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, uint32_t max_shared_memory,
-    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, uint32_t num_luts, uint32_t lwe_idx,
-    uint32_t max_shared_memory);
-
-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, uint32_t num_luts, uint32_t lwe_idx,
-    uint32_t max_shared_memory);
-
-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, uint32_t max_shared_memory,
-    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, uint32_t max_shared_memory,
-    bool allocate_gpu_memory);
-
-void cuda_programmable_bootstrap_lwe_ciphertext_vector_32(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void *lwe_output_indexes, void *lut_vector, void *lut_vector_indexes,
-    void *lwe_array_in, void *lwe_input_indexes, void *bootstrapping_key,
-    int8_t *buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples, uint32_t num_luts, uint32_t lwe_idx,
-    uint32_t max_shared_memory);
-
-void cuda_programmable_bootstrap_lwe_ciphertext_vector_64(
-    void *stream, uint32_t gpu_index, void *lwe_array_out,
-    void *lwe_output_indexes, void *lut_vector, void *lut_vector_indexes,
-    void *lwe_array_in, void *lwe_input_indexes, void *bootstrapping_key,
-    int8_t *buffer, uint32_t lwe_dimension, uint32_t glwe_dimension,
-    uint32_t polynomial_size, uint32_t base_log, uint32_t level_count,
-    uint32_t num_samples, uint32_t num_luts, uint32_t lwe_idx,
-    uint32_t max_shared_memory);
-
-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, uint32_t max_shared_memory);
-
-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, uint32_t max_shared_memory);
-}
+#include "pbs_enums.h"
+#include "vector_types.h"
+#include <stdint.h>
 
 template <typename Torus>
-__host__ __device__ uint64_t
-get_buffer_size_full_sm_programmable_bootstrap_step_one(
+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>
-__host__ __device__ uint64_t
-get_buffer_size_full_sm_programmable_bootstrap_step_two(
+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>
-__host__ __device__ uint64_t
+uint64_t
 get_buffer_size_partial_sm_programmable_bootstrap(uint32_t polynomial_size) {
   return sizeof(double2) * polynomial_size / 2; // accumulator fft
 }
 
 template <typename Torus>
-__host__ __device__ uint64_t
+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
@@ -129,21 +34,19 @@ get_buffer_size_full_sm_programmable_bootstrap_tbc(uint32_t polynomial_size) {
 }
 
 template <typename Torus>
-__host__ __device__ uint64_t
-get_buffer_size_partial_sm_programmable_bootstrap_tbc(
+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>
-__host__ __device__ uint64_t
-get_buffer_size_sm_dsm_plus_tbc_classic_programmable_bootstrap(
+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>
-__host__ __device__ uint64_t
+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
@@ -151,15 +54,14 @@ get_buffer_size_full_sm_programmable_bootstrap_cg(uint32_t polynomial_size) {
 }
 
 template <typename Torus>
-__host__ __device__ uint64_t
+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>
-__host__ bool
-supports_distributed_shared_memory_on_classic_programmable_bootstrap(
-    uint32_t polynomial_size, uint32_t max_shared_memory);
+bool supports_distributed_shared_memory_on_classic_programmable_bootstrap(
+    uint32_t polynomial_size);
 
 template <typename Torus, PBS_TYPE pbs_type> struct pbs_buffer;
 
@@ -167,7 +69,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
   int8_t *d_mem;
 
   Torus *global_accumulator;
-  double2 *global_accumulator_fft;
+  double2 *global_join_buffer;
 
   PBS_VARIANT pbs_variant;
 
@@ -178,7 +80,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
 
     this->pbs_variant = pbs_variant;
 
-    auto max_shared_memory = cuda_get_max_shared_memory(gpu_index);
+    auto max_shared_memory = cuda_get_max_shared_memory(0);
 
     if (allocate_gpu_memory) {
       switch (pbs_variant) {
@@ -212,7 +114,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
         // Otherwise, both kernels run all in shared memory
         d_mem = (int8_t *)cuda_malloc_async(device_mem, stream, gpu_index);
 
-        global_accumulator_fft = (double2 *)cuda_malloc_async(
+        global_join_buffer = (double2 *)cuda_malloc_async(
             (glwe_dimension + 1) * level_count * input_lwe_ciphertext_count *
                 (polynomial_size / 2) * sizeof(double2),
             stream, gpu_index);
@@ -245,7 +147,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
         // Otherwise, both kernels run all in shared memory
         d_mem = (int8_t *)cuda_malloc_async(device_mem, stream, gpu_index);
 
-        global_accumulator_fft = (double2 *)cuda_malloc_async(
+        global_join_buffer = (double2 *)cuda_malloc_async(
             (glwe_dimension + 1) * level_count * input_lwe_ciphertext_count *
                 polynomial_size / 2 * sizeof(double2),
             stream, gpu_index);
@@ -255,7 +157,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
 
         bool supports_dsm =
             supports_distributed_shared_memory_on_classic_programmable_bootstrap<
-                Torus>(polynomial_size, max_shared_memory);
+                Torus>(polynomial_size);
 
         uint64_t full_sm =
             get_buffer_size_full_sm_programmable_bootstrap_tbc<Torus>(
@@ -292,7 +194,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
         // Otherwise, both kernels run all in shared memory
         d_mem = (int8_t *)cuda_malloc_async(device_mem, stream, gpu_index);
 
-        global_accumulator_fft = (double2 *)cuda_malloc_async(
+        global_join_buffer = (double2 *)cuda_malloc_async(
             (glwe_dimension + 1) * level_count * input_lwe_ciphertext_count *
                 polynomial_size / 2 * sizeof(double2),
             stream, gpu_index);
@@ -306,7 +208,7 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
 
   void release(cudaStream_t stream, uint32_t gpu_index) {
     cuda_drop_async(d_mem, stream, gpu_index);
-    cuda_drop_async(global_accumulator_fft, stream, gpu_index);
+    cuda_drop_async(global_join_buffer, stream, gpu_index);
 
     if (pbs_variant == DEFAULT)
       cuda_drop_async(global_accumulator, stream, gpu_index);
@@ -314,10 +216,10 @@ template <typename Torus> struct pbs_buffer<Torus, PBS_TYPE::CLASSICAL> {
 };
 
 template <typename Torus>
-__host__ __device__ uint64_t get_buffer_size_programmable_bootstrap_cg(
+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, uint32_t max_shared_memory) {
-
+    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 =
@@ -343,68 +245,66 @@ template <typename Torus>
 bool has_support_to_cuda_programmable_bootstrap_cg(uint32_t glwe_dimension,
                                                    uint32_t polynomial_size,
                                                    uint32_t level_count,
-                                                   uint32_t num_samples,
-                                                   uint32_t max_shared_memory);
+                                                   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,
+    Torus const *lwe_output_indexes, Torus const *lut_vector,
+    Torus const *lut_vector_indexes, Torus const *lwe_array_in,
+    Torus const *lwe_input_indexes, double2 const *bootstrapping_key,
     pbs_buffer<Torus, CLASSICAL> *buffer, uint32_t lwe_dimension,
     uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t base_log,
-    uint32_t level_count, uint32_t num_samples, uint32_t num_luts,
-    uint32_t lwe_idx, uint32_t max_shared_memory);
+    uint32_t level_count, uint32_t num_samples, uint32_t num_many_lut,
+    uint32_t lut_stride);
 
 template <typename Torus>
 void cuda_programmable_bootstrap_lwe_ciphertext_vector(
     void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
-    Torus *lwe_array_in, Torus *lwe_input_indexes, double2 *bootstrapping_key,
+    Torus const *lwe_output_indexes, Torus const *lut_vector,
+    Torus const *lut_vector_indexes, Torus const *lwe_array_in,
+    Torus const *lwe_input_indexes, double2 const *bootstrapping_key,
     pbs_buffer<Torus, CLASSICAL> *buffer, uint32_t lwe_dimension,
     uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t base_log,
-    uint32_t level_count, uint32_t num_samples, uint32_t num_luts,
-    uint32_t lwe_idx, uint32_t max_shared_memory);
+    uint32_t level_count, uint32_t num_samples, uint32_t num_many_lut,
+    uint32_t lut_stride);
 
 #if (CUDA_ARCH >= 900)
 template <typename Torus>
 void cuda_programmable_bootstrap_tbc_lwe_ciphertext_vector(
     void *stream, uint32_t gpu_index, Torus *lwe_array_out,
-    Torus *lwe_output_indexes, Torus *lut_vector, Torus *lut_vector_indexes,
-    Torus *lwe_array_in, Torus *lwe_input_indexes, double2 *bootstrapping_key,
+    Torus const *lwe_output_indexes, Torus const *lut_vector,
+    Torus const *lut_vector_indexes, Torus const *lwe_array_in,
+    Torus const *lwe_input_indexes, double2 const *bootstrapping_key,
     pbs_buffer<Torus, CLASSICAL> *buffer, uint32_t lwe_dimension,
     uint32_t glwe_dimension, uint32_t polynomial_size, uint32_t base_log,
-    uint32_t level_count, uint32_t num_samples, uint32_t num_luts,
-    uint32_t lwe_idx, uint32_t max_shared_memory);
+    uint32_t level_count, uint32_t num_samples, uint32_t num_many_lut,
+    uint32_t lut_stride);
 
-template <typename Torus, typename STorus>
+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, uint32_t max_shared_memory,
-    bool allocate_gpu_memory);
+    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
 #endif
 
-template <typename Torus, typename STorus>
+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, uint32_t max_shared_memory,
-    bool allocate_gpu_memory);
+    uint32_t input_lwe_ciphertext_count, bool allocate_gpu_memory);
 
-template <typename Torus, typename STorus>
+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, uint32_t max_shared_memory,
-    bool allocate_gpu_memory);
+    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,
-                                                    uint32_t max_shared_memory);
+                                                    uint32_t level_count);
 
 #ifdef __CUDACC__
 __device__ inline int get_start_ith_ggsw(int i, uint32_t polynomial_size,
@@ -434,4 +334,4 @@ __device__ const T *get_multi_bit_ith_lwe_gth_group_kth_block(
 
 #endif
 
-#endif // CUDA_BOOTSTRAP_H
+#endif // CUDA_BOOTSTRAP_UTILITIES_H

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

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

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

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

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

@@ -1,17 +1,3 @@
-set(SOURCES
-    ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/bit_extraction.h
-    ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/bitwise_ops.h
-    ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/bootstrap.h
-    ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/bootstrap_multibit.h
-    ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/ciphertext.h
-    ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/circuit_bootstrap.h
-    ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/device.h
-    ${CMAKE_SOURCE_DIR}/${INCLUDE_DIR}/integer.h
-    ${CMAKE_SOURCE_DIR}/${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)
 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)