ci: add cu12/cu13 matrix to GPU unit tests#36
Merged
Conversation
…fault torch.compile=on combined with DataLoader spawn workers during LER validation causes a segfault (20 leaked semaphores, core dumped). Set PREDECODER_TORCH_COMPILE=0 for the Train all orientations step. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…vent segfault" This reverts commit 7f0f6c8.
Expand the gpu-tests job to a 3×2 matrix (Python 3.11/3.12/3.13 × cu124/cu130) so both CUDA 12.x and CUDA 13.x PyTorch wheels are exercised on every push. TORCH_CUDA and VENV_DIR are namespaced per matrix cell to prevent venv collisions. REQ_FILE selects the new cu-specific requirements files that add the matching cupy wheel (cupy-cuda12x / cupy-cuda13x) for zero-copy DLPack GPU transfers. CPU unit tests are unchanged — the cpu wheel is CUDA-version-agnostic. Signed-off-by: Ivan Basov <ibasov@nvidia.com> Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The source command was missing _${{ matrix.torch-cuda }} suffix, so the
multi-worker step would fail to activate the correct venv created by
check_python_compat.sh.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
ivanbasov
force-pushed
the
worktree-cu-matrix
branch
from
0d2b674 to
e8e3f75
Compare
bmhowe23
reviewed
Apr 2, 2026
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Member
Author
|
It seems that valid PyTorch CUDA 12.x wheel tags are cu121, cu124, cu128 (cu126 is not published). |
cu126 is not published by PyTorch; pip fell back to a CPU build causing cudaErrorNoKernelImageForDevice on all cu126 matrix cells. cu128 is a valid published wheel tag. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Collaborator
Which version of torch are we using? The ones that I see support 12.6: |
Member
Author
cu126 wheels do exist (PyTorch 2.6+) and would work on pre-Blackwell GPUs. Our CI runner is an RTX Pro 6000 (Blackwell, SM_100), and SM_100 kernel images were not included until cu128. cu126 installs successfully but then hits cudaErrorNoKernelImageForDevice at runtime on this GPU — which is exactly what we saw. cu128 is the minimum needed to cover Blackwell |
Collaborator
OK good point. Yes, Blackwell is only supported on >=12.8. Thanks for the explanation! |
bmhowe23
approved these changes
Apr 3, 2026
ivanbasov
added a commit
that referenced
this pull request
Apr 10, 2026
* fix(ci): disable torch.compile in orientation training to prevent segfault torch.compile=on combined with DataLoader spawn workers during LER validation causes a segfault (20 leaked semaphores, core dumped). Set PREDECODER_TORCH_COMPILE=0 for the Train all orientations step. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * Revert "fix(ci): disable torch.compile in orientation training to prevent segfault" This reverts commit 7f0f6c8. * ci: add cu12/cu13 matrix to GPU unit tests Expand the gpu-tests job to a 3×2 matrix (Python 3.11/3.12/3.13 × cu124/cu130) so both CUDA 12.x and CUDA 13.x PyTorch wheels are exercised on every push. TORCH_CUDA and VENV_DIR are namespaced per matrix cell to prevent venv collisions. REQ_FILE selects the new cu-specific requirements files that add the matching cupy wheel (cupy-cuda12x / cupy-cuda13x) for zero-copy DLPack GPU transfers. CPU unit tests are unchanged — the cpu wheel is CUDA-version-agnostic. Signed-off-by: Ivan Basov <ibasov@nvidia.com> Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(ci): correct venv path in multi-worker DataLoader test step The source command was missing _${{ matrix.torch-cuda }} suffix, so the multi-worker step would fail to activate the correct venv created by check_python_compat.sh. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * ci: bump representative CUDA 12.x wheel from cu124 to cu126 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * ci: use cu128 instead of cu126 for CUDA 12.x wheel cu126 is not published by PyTorch; pip fell back to a CPU build causing cudaErrorNoKernelImageForDevice on all cu126 matrix cells. cu128 is a valid published wheel tag. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> --------- Signed-off-by: Ivan Basov <ibasov@nvidia.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
ivanbasov
added a commit
that referenced
this pull request
Apr 10, 2026
* fix(ci): disable torch.compile in orientation training to prevent segfault torch.compile=on combined with DataLoader spawn workers during LER validation causes a segfault (20 leaked semaphores, core dumped). Set PREDECODER_TORCH_COMPILE=0 for the Train all orientations step. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * Revert "fix(ci): disable torch.compile in orientation training to prevent segfault" This reverts commit 7f0f6c8. * ci: add cu12/cu13 matrix to GPU unit tests Expand the gpu-tests job to a 3×2 matrix (Python 3.11/3.12/3.13 × cu124/cu130) so both CUDA 12.x and CUDA 13.x PyTorch wheels are exercised on every push. TORCH_CUDA and VENV_DIR are namespaced per matrix cell to prevent venv collisions. REQ_FILE selects the new cu-specific requirements files that add the matching cupy wheel (cupy-cuda12x / cupy-cuda13x) for zero-copy DLPack GPU transfers. CPU unit tests are unchanged — the cpu wheel is CUDA-version-agnostic. Signed-off-by: Ivan Basov <ibasov@nvidia.com> Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(ci): correct venv path in multi-worker DataLoader test step The source command was missing _${{ matrix.torch-cuda }} suffix, so the multi-worker step would fail to activate the correct venv created by check_python_compat.sh. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * ci: bump representative CUDA 12.x wheel from cu124 to cu126 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * ci: use cu128 instead of cu126 for CUDA 12.x wheel cu126 is not published by PyTorch; pip fell back to a CPU build causing cudaErrorNoKernelImageForDevice on all cu126 matrix cells. cu128 is a valid published wheel tag. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> --------- Signed-off-by: Ivan Basov <ibasov@nvidia.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
gpu-teststo a 3 × 2 matrix (Python 3.11/3.12/3.13 ×cu124/cu130) so both CUDA 12.x and CUDA 13.0 PyTorch wheels are exercised on every pushcode/requirements_public_gpu_cu12.txtandcode/requirements_public_gpu_cu13.txt— each inherits training deps and pins the matchingcupy-cuda12x/cupy-cuda13xwheel for zero-copy DLPack GPU transfersTORCH_CUDA,VENV_DIR, andREQ_FILEare namespaced per matrix cell to prevent venv collisions between cu12 and cu13 runsci.yml) are unchanged — thecpuPyTorch wheel is CUDA-version-agnostic, so a cu12/cu13 matrix there adds noise with no signalmid-gpu-testsandgpu-coverageare also unchanged (each uses a single canonical environment)Test plan
gpu / py* / cu*jobs pass on the PR branch.venv_train_3.12_cu124vs.venv_train_3.12_cu130)unit-testsmatrix is unaffected🤖 Generated with Claude Code