Add arm64 pytorch nightly entries and fix incorrect comment

PyTorch ships linux-aarch64 GPU wheels (cu126, cu130). Added 4 arm64
pytorch entries to the nightly matrix with GPU l4, and a corresponding
test-pytorch-linux-aarch64 job in ci-nightly.yml.

Removed the incorrect comment claiming PyTorch doesn't ship arm64 wheels.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: leofang

.github/workflows/ci-nightly.yml

@@ -90,6 +90,24 @@ jobs:
       test-mode: nightly-pytorch
       matrix_filter: 'map(select(.MODE == "nightly-pytorch"))'
 
+  test-pytorch-linux-aarch64:
+    name: "Nightly PyTorch (linux-aarch64)"
+    if: ${{ github.repository_owner == 'nvidia' }}
+    needs: find-wheels
+    permissions:
+      contents: read
+      actions: read
+    secrets: inherit
+    uses: ./.github/workflows/test-wheel-linux.yml
+    with:
+      build-type: nightly
+      host-platform: linux-aarch64
+      build-ctk-ver: ${{ needs.find-wheels.outputs.CUDA_BUILD_VER }}
+      run-id: ${{ needs.find-wheels.outputs.RUN_ID }}
+      sha: ${{ needs.find-wheels.outputs.HEAD_SHA }}
+      test-mode: nightly-pytorch
+      matrix_filter: 'map(select(.MODE == "nightly-pytorch"))'
+
   test-pytorch-windows:
     name: "Nightly PyTorch (win-64)"
     if: ${{ github.repository_owner == 'nvidia' }}
@@ -172,6 +190,7 @@ jobs:
     runs-on: ubuntu-latest
     needs:
       - test-pytorch-linux
+      - test-pytorch-linux-aarch64
       - test-pytorch-windows
       - test-numba-cuda-linux-64
       - test-numba-cuda-linux-aarch64
@@ -186,6 +205,8 @@ jobs:
           # default behaviour.
           if ${{ needs.test-pytorch-linux.result == 'cancelled' ||
                  needs.test-pytorch-linux.result == 'failure' ||
+                 needs.test-pytorch-linux-aarch64.result == 'cancelled' ||
+                 needs.test-pytorch-linux-aarch64.result == 'failure' ||
                  needs.test-pytorch-windows.result == 'cancelled' ||
                  needs.test-pytorch-windows.result == 'failure' ||
                  needs.test-numba-cuda-linux-64.result == 'cancelled' ||

ci/test-matrix.yml

@@ -63,11 +63,15 @@ linux:
     - { ARCH: 'amd64', PY_VER: '3.11',  CUDA_VER: '12.9.1', LOCAL_CTK: '0', GPU: 't4',         GPU_COUNT: '1', DRIVER: 'latest', FLAVOR: 'wsl' }
     - { ARCH: 'amd64', PY_VER: '3.12',  CUDA_VER: '13.2.1', LOCAL_CTK: '0', GPU: 'rtx4090',    GPU_COUNT: '1', DRIVER: 'latest', FLAVOR: 'wsl' }
   nightly:
-    # nightly-pytorch (amd64 only — PyTorch does not ship arm64 GPU wheels)
-    - { MODE: 'nightly-pytorch',    ARCH: 'amd64', PY_VER: '3.12', CUDA_VER: '12.6.3', LOCAL_CTK: '0', GPU: 'l4', GPU_COUNT: '1', DRIVER: 'latest', TORCH_VER: '2.11.0',   TORCH_CUDA: 'cu126' }
-    - { MODE: 'nightly-pytorch',    ARCH: 'amd64', PY_VER: '3.12', CUDA_VER: '13.0.2', LOCAL_CTK: '0', GPU: 'l4', GPU_COUNT: '1', DRIVER: 'latest', TORCH_VER: '2.11.0',   TORCH_CUDA: 'cu130' }
+    # nightly-pytorch
+    - { MODE: 'nightly-pytorch',    ARCH: 'amd64', PY_VER: '3.12', CUDA_VER: '12.6.3', LOCAL_CTK: '0', GPU: 'l4', GPU_COUNT: '1', DRIVER: 'latest', TORCH_VER: '2.11.0', TORCH_CUDA: 'cu126' }
+    - { MODE: 'nightly-pytorch',    ARCH: 'amd64', PY_VER: '3.12', CUDA_VER: '13.0.2', LOCAL_CTK: '0', GPU: 'l4', GPU_COUNT: '1', DRIVER: 'latest', TORCH_VER: '2.11.0', TORCH_CUDA: 'cu130' }
     - { MODE: 'nightly-pytorch',    ARCH: 'amd64', PY_VER: '3.12', CUDA_VER: '12.6.3', LOCAL_CTK: '0', GPU: 'l4', GPU_COUNT: '1', DRIVER: 'latest', TORCH_VER: '2.9.1',  TORCH_CUDA: 'cu126' }
     - { MODE: 'nightly-pytorch',    ARCH: 'amd64', PY_VER: '3.12', CUDA_VER: '13.0.2', LOCAL_CTK: '0', GPU: 'l4', GPU_COUNT: '1', DRIVER: 'latest', TORCH_VER: '2.9.1',  TORCH_CUDA: 'cu130' }
+    - { MODE: 'nightly-pytorch',    ARCH: 'arm64', PY_VER: '3.12', CUDA_VER: '12.6.3', LOCAL_CTK: '0', GPU: 'l4', GPU_COUNT: '1', DRIVER: 'latest', TORCH_VER: '2.11.0', TORCH_CUDA: 'cu126' }
+    - { MODE: 'nightly-pytorch',    ARCH: 'arm64', PY_VER: '3.12', CUDA_VER: '13.0.2', LOCAL_CTK: '0', GPU: 'l4', GPU_COUNT: '1', DRIVER: 'latest', TORCH_VER: '2.11.0', TORCH_CUDA: 'cu130' }
+    - { MODE: 'nightly-pytorch',    ARCH: 'arm64', PY_VER: '3.12', CUDA_VER: '12.6.3', LOCAL_CTK: '0', GPU: 'l4', GPU_COUNT: '1', DRIVER: 'latest', TORCH_VER: '2.9.1',  TORCH_CUDA: 'cu126' }
+    - { MODE: 'nightly-pytorch',    ARCH: 'arm64', PY_VER: '3.12', CUDA_VER: '13.0.2', LOCAL_CTK: '0', GPU: 'l4', GPU_COUNT: '1', DRIVER: 'latest', TORCH_VER: '2.9.1',  TORCH_CUDA: 'cu130' }
     # nightly-numba-cuda
     - { MODE: 'nightly-numba-cuda', ARCH: 'amd64', PY_VER: '3.12', CUDA_VER: '12.9.1', LOCAL_CTK: '0', GPU: 'l4', GPU_COUNT: '1', DRIVER: 'latest' }
     - { MODE: 'nightly-numba-cuda', ARCH: 'amd64', PY_VER: '3.12', CUDA_VER: '13.2.1', LOCAL_CTK: '0', GPU: 'l4', GPU_COUNT: '1', DRIVER: 'latest' }