[NE16] Add GAP9_w_NE16 platform (NE16 accelerator on GAP9)

.github/workflows/_runner-gap9-w-ne16-tiled.yml

@@ -0,0 +1,84 @@
+# SPDX-FileCopyrightText: 2026 ETH Zurich and University of Bologna
+#
+# SPDX-License-Identifier: Apache-2.0
+
+---
+name: _runner-gap9-w-ne16-tiled
+
+"on":
+  workflow_call:
+    inputs:
+      runner:
+        required: true
+        type: string
+      docker-image:
+        required: true
+        type: string
+      pytest-markers:
+        required: true
+        type: string
+
+jobs:
+  test-runner-gap9-w-ne16-tiled:
+    runs-on: ${{ inputs.runner }}
+    container:
+      image: ${{ inputs.docker-image }}
+    steps:
+      - name: Mark workspace as safe
+        run: git config --global --add safe.directory '*'
+      - name: Checkout Repo
+        uses: actions/checkout@v4
+        with:
+          submodules: recursive
+      - name: Build Deeploy
+        shell: bash
+        run: |
+          source /app/install/gap9-sdk/.gap9-venv/bin/activate
+          source /app/install/gap9-sdk/configs/gap9_evk_audio.sh || true
+          pip install -e . || true
+          deactivate
+      - name: Cache ccache
+        uses: actions/cache/restore@v4
+        with:
+          path: /app/.ccache
+          key: ccache-gap9
+      - name: Run Test
+        run: |
+          source /app/install/gap9-sdk/.gap9-venv/bin/activate
+          source /app/install/gap9-sdk/configs/gap9_evk_audio.sh || true
+          export GVSOC_INSTALL_DIR=/app/install/gap9-sdk/install/workstation
+          export GAP_RISCV_GCC_TOOLCHAIN=/app/install/gcc/gap9
+          cd DeeployTest
+          mkdir -p /app/.ccache
+          export CCACHE_DIR=/app/.ccache
+          pytest test_platforms.py -v -m "${{ inputs.pytest-markers }}"
+          deactivate
+        shell: bash
+      - name: NE16 Profiling (cycle counts)
+        if: always()
+        run: |
+          source /app/install/gap9-sdk/.gap9-venv/bin/activate
+          source /app/install/gap9-sdk/configs/gap9_evk_audio.sh || true
+          export GVSOC_INSTALL_DIR=/app/install/gap9-sdk/install/workstation
+          export GAP_RISCV_GCC_TOOLCHAIN=/app/install/gcc/gap9
+          mkdir -p /app/.ccache
+          export CCACHE_DIR=/app/.ccache
+          cd DeeployTest
+          rm -rf TEST_GAP9_W_NE16/build_master
+          for test in \
+            "Tests/Kernels/Integer/Conv/Dense_2D_RQ_NE16Bench --l1 32000" \
+            "Tests/Kernels/Integer/Conv/PW_2D_RQ/Regular_RQ --l1 32000" \
+            "Tests/Kernels/Integer/Conv/Dense_2D_RQ --l1 32000"; do
+            dir=$(echo $test | awk '{print $1}')
+            l1=$(echo $test | awk '{print $3}')
+            echo "========================================"
+            echo "PROFILING: $dir (L1=$l1)"
+            echo "========================================"
+            python3 deeployRunner_tiled_gap9_w_ne16.py \
+              -t "$dir" --l1 "$l1" \
+              --toolchain GCC --toolchain-install-dir /app/install/gcc/gap9 \
+              --cores 8 --enable-3x3 --profileTiling -v \
+              -D CMAKE_INTERPROCEDURAL_OPTIMIZATION=OFF 2>&1 || true
+          done
+          deactivate
+        shell: bash

.github/workflows/_runner-snitch-tiled-sequential.yml

@@ -33,10 +33,10 @@ jobs:
       - name: Build Deeploy
         shell: bash
         run: pip install -e .
-      - name: Run Test # VJUNG: Run tests with 4 parallel threads as GitHub action VM has 4 cores.
+      - name: Run Test # 2-way parallel: 4-way OOMs the GitHub runner on the FP32 GEMM/TransB build.
         run: |
           cd DeeployTest
           mkdir -p /app/.ccache
           export CCACHE_DIR=/app/.ccache
-          pytest test_platforms.py -v -n 4 -m "snitch_tiled and ${{ inputs.pytest-marker }}"
+          pytest test_platforms.py -v -n 2 -m "snitch_tiled and ${{ inputs.pytest-marker }}"
         shell: bash

.github/workflows/_runner-xdna2.yml

@@ -0,0 +1,52 @@
+# SPDX-FileCopyrightText: 2025 ETH Zurich and University of Bologna
+#
+# SPDX-License-Identifier: Apache-2.0
+
+---
+name: _runner-xdna2
+
+"on":
+  workflow_call:
+    inputs:
+      pytest-marker:
+        required: true
+        type: string
+      docker-image:
+        required: false
+        type: string
+
+jobs:
+  test-runner-xdna2:
+    runs-on: xdna2-npu
+    # NOTE: We cannot use the `container:` directive here because
+    # GitHub Actions does not support `--device` flags required for
+    # NPU access (/dev/accel/accel0). Instead we use explicit
+    # `docker run` commands.
+    steps:
+      - name: Fix workspace permissions
+        shell: bash
+        run: |
+          docker run --rm \
+            -v "${{ github.workspace }}":/workspace \
+            ${{ inputs.docker-image }} \
+            chown -R $(id -u):$(id -g) /workspace || true
+
+      - name: Checkout Repo
+        uses: actions/checkout@v4
+        with:
+          submodules: recursive
+
+      - name: Run Tests in Docker
+        shell: bash
+        run: |
+          docker run --rm \
+            --device /dev/accel/accel0 \
+            --ulimit memlock=-1 \
+            -v "${{ github.workspace }}":/app/Deeploy \
+            -w /app/Deeploy \
+            ${{ inputs.docker-image }} \
+            bash -c "
+              pip install -e . &&
+              cd DeeployTest &&
+              pytest test_platforms.py -v -m 'xdna2 and ${{ inputs.pytest-marker }}'
+            "

.github/workflows/ci-platform-gap9-tiled.yml

@@ -25,6 +25,9 @@ concurrency:
 
 jobs:
   select-env:
+    # ghcr.io/pulp-platform/deeploy-gap9 is private; only upstream's
+    # self-hosted runners have credentials. Skip cleanly on forks.
+    if: github.repository == 'pulp-platform/Deeploy'
     uses: ./.github/workflows/_select-env.yml
     with:
       docker_image_deeploy: ${{ github.event.inputs.docker_image_deeploy || 'ghcr.io/pulp-platform/deeploy-gap9:devel' }}

.github/workflows/ci-platform-gap9-w-ne16-tiled.yml

@@ -0,0 +1,68 @@
+# SPDX-FileCopyrightText: 2026 ETH Zurich and University of Bologna
+#
+# SPDX-License-Identifier: Apache-2.0
+
+---
+name: CI • GAP9 + NE16 (Tiled)
+
+"on":
+  push:
+    branches:
+      - "**"
+    tags:
+      - "v*.*.*"
+  pull_request:
+  workflow_dispatch:
+    inputs:
+      docker_image_deeploy:
+        description: "Deeploy Image to use"
+        required: false
+        default: "ghcr.io/pulp-platform/deeploy-gap9:devel"
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  select-env:
+    # The GAP9 + NE16 image is hosted in pulp-platform's private ghcr.io
+    # registry; only upstream's self-hosted runners have credentials to
+    # pull it. On forks the docker pull always returns "denied", so skip
+    # the whole pipeline cleanly there. (Same constraint as the existing
+    # ci-platform-gap9{,-tiled}.yml jobs.)
+    if: github.repository == 'pulp-platform/Deeploy'
+    uses: ./.github/workflows/_select-env.yml
+    with:
+      docker_image_deeploy: ${{ github.event.inputs.docker_image_deeploy || 'ghcr.io/pulp-platform/deeploy-gap9:devel' }}
+
+  gap9-w-ne16-kernels-tiled-singlebuffer-L2:
+    needs: select-env
+    uses: ./.github/workflows/_runner-gap9-w-ne16-tiled.yml
+    with:
+      runner: ${{ needs.select-env.outputs.runner }}
+      docker-image: ${{ needs.select-env.outputs.image }}
+      pytest-markers: "gap9_w_ne16_tiled and kernels and singlebuffer and l2"
+
+  gap9-w-ne16-kernels-tiled-doublebuffer-L2:
+    needs: select-env
+    uses: ./.github/workflows/_runner-gap9-w-ne16-tiled.yml
+    with:
+      runner: ${{ needs.select-env.outputs.runner }}
+      docker-image: ${{ needs.select-env.outputs.image }}
+      pytest-markers: "gap9_w_ne16_tiled and kernels and doublebuffer and l2"
+
+  gap9-w-ne16-models-tiled-singlebuffer-L2:
+    needs: select-env
+    uses: ./.github/workflows/_runner-gap9-w-ne16-tiled.yml
+    with:
+      runner: ${{ needs.select-env.outputs.runner }}
+      docker-image: ${{ needs.select-env.outputs.image }}
+      pytest-markers: "gap9_w_ne16_tiled and models and singlebuffer and l2"
+
+  gap9-w-ne16-models-tiled-doublebuffer-L2:
+    needs: select-env
+    uses: ./.github/workflows/_runner-gap9-w-ne16-tiled.yml
+    with:
+      runner: ${{ needs.select-env.outputs.runner }}
+      docker-image: ${{ needs.select-env.outputs.image }}
+      pytest-markers: "gap9_w_ne16_tiled and models and doublebuffer and l2"

.github/workflows/ci-platform-gap9.yml

@@ -26,6 +26,9 @@ concurrency:
 
 jobs:
   select-env:
+    # ghcr.io/pulp-platform/deeploy-gap9 is private; only upstream's
+    # self-hosted runners have credentials. Skip cleanly on forks.
+    if: github.repository == 'pulp-platform/Deeploy'
     uses: ./.github/workflows/_select-env.yml
     with:
       docker_image_deeploy: ${{ github.event.inputs.docker_image_deeploy || 'ghcr.io/pulp-platform/deeploy-gap9:devel' }}

.github/workflows/ci-platform-snitch-tiled.yml

@@ -36,3 +36,11 @@ jobs:
       runner: ${{ needs.select-env.outputs.runner }}
       docker-image: ${{ needs.select-env.outputs.image }}
       pytest-marker: "kernels and singlebuffer and l2"
+
+  snitch-models-tiled-singlebuffer-L2:
+    needs: select-env
+    uses: ./.github/workflows/_runner-snitch-tiled-sequential.yml
+    with:
+      runner: ${{ needs.select-env.outputs.runner }}
+      docker-image: ${{ needs.select-env.outputs.image }}
+      pytest-marker: "models and singlebuffer and l2"

.github/workflows/ci-platform-snitch.yml

@@ -36,3 +36,11 @@ jobs:
       runner: ${{ needs.select-env.outputs.runner }}
       docker-image: ${{ needs.select-env.outputs.image }}
       pytest-marker: "kernels"
+
+  snitch-models:
+    needs: select-env
+    uses: ./.github/workflows/_runner-snitch.yml
+    with:
+      runner: ${{ needs.select-env.outputs.runner }}
+      docker-image: ${{ needs.select-env.outputs.image }}
+      pytest-marker: "models"

.github/workflows/ci-platform-xdna2.yml

@@ -0,0 +1,31 @@
+# SPDX-FileCopyrightText: 2025 ETH Zurich and University of Bologna
+#
+# SPDX-License-Identifier: Apache-2.0
+
+---
+name: CI • XDNA2
+
+"on":
+  push:
+    branches:
+      - "**"
+    tags:
+      - "v*.*.*"
+  pull_request:
+  workflow_dispatch:
+    inputs:
+      docker_image_deeploy:
+        description: "XDNA2 Deeploy Docker Image to use"
+        required: false
+        default: "ghcr.io/pulp-platform/deeploy-xdna:devel"
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  xdna2-kernels:
+    uses: ./.github/workflows/_runner-xdna2.yml
+    with:
+      pytest-marker: "kernels"
+      docker-image: ${{ inputs.docker_image_deeploy || 'ghcr.io/pulp-platform/deeploy-xdna:devel' }}

.github/workflows/docker-build-deeploy-xdna.yml

@@ -0,0 +1,100 @@
+# SPDX-FileCopyrightText: 2026 ETH Zurich and University of Bologna
+#
+# SPDX-License-Identifier: Apache-2.0
+
+---
+name: Docker • Build Deeploy XDNA Container
+
+"on":
+  workflow_dispatch:
+
+jobs:
+  prepare:
+    name: Fetch branch name or tag
+    runs-on: ubuntu-latest
+    outputs:
+      docker_tag: ${{ steps.generate_tag.outputs.docker_tag }}
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Set up environment variables
+        run: |
+          echo "BRANCH_NAME=${GITHUB_REF##*/}" >> $GITHUB_ENV
+          echo "TAG_NAME=${GITHUB_REF##*/}" >> $GITHUB_ENV
+          echo "IS_TAG=${GITHUB_REF_TYPE}" >> $GITHUB_ENV
+
+      - name: Set Docker tag
+        id: generate_tag
+        run: |
+          if [[ "${{ env.IS_TAG }}" == "tag" ]]; then
+            echo "docker_tag=${{ env.TAG_NAME }}" >> $GITHUB_OUTPUT
+          else
+            echo "docker_tag=${{ env.BRANCH_NAME }}" >> $GITHUB_OUTPUT
+          fi
+
+  build-deeploy-xdna:
+    name: Build Deeploy XDNA Image
+    needs: [prepare]
+    runs-on: ubuntu-latest
+    outputs:
+      digest-amd64: ${{ steps.digest.outputs.digest-amd64 }}
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Free up disk space
+        uses: jlumbroso/free-disk-space@v1.3.1
+        with:
+          tool-cache: true
+          android: true
+          dotnet: true
+          haskell: true
+          large-packages: true
+
+      - uses: docker/setup-buildx-action@v3
+
+      - name: GHCR Log-in
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Build Cache for Docker
+        id: cache
+        uses: actions/cache@v4
+        with:
+          path: var-ccache
+          key: ${{ runner.os }}-amd64-build-cache-deeploy-xdna
+
+      - name: Inject build-cache
+        uses: reproducible-containers/buildkit-cache-dance@v3.1.0
+        with:
+          cache-map: |
+            {
+              "var-ccache": "/ccache"
+            }
+          skip-extraction: ${{ steps.cache.outputs.cache-hit }}
+
+      - name: Lower Case Repository Name
+        run: |
+          echo "OWNER_LC=${OWNER,,}" >>${GITHUB_ENV}
+        env:
+          OWNER: "${{ github.repository_owner }}"
+
+      - name: Build and push Deeploy XDNA image
+        id: build
+        uses: docker/build-push-action@v6
+        with:
+          platforms: linux/amd64
+          context: .
+          cache-from: type=gha
+          cache-to: type=gha,mode=min
+          file: Container/Dockerfile.deeploy-xdna
+          push: true
+          tags: |
+            ghcr.io/${{ env.OWNER_LC }}/deeploy-xdna:latest
+            ghcr.io/${{ env.OWNER_LC }}/deeploy-xdna:${{ needs.prepare.outputs.docker_tag }}
+
+      - name: Extract image digest
+        id: digest
+        run: echo "digest-amd64=${{ steps.build.outputs.digest }}" >> $GITHUB_OUTPUT