CUDA aarch64 12.6 and 12.8 builds fix triton constraints#165013
CUDA aarch64 12.6 and 12.8 builds fix triton constraints#165013atalman wants to merge 1 commit intopytorch:mainfrom
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/165013
Note: Links to docs will display an error until the docs builds have been completed. ⏳ No Failures, 15 PendingAs of commit e51ef81 with merge base 4308b8a ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
| if [[ "$DESIRED_CUDA" == "cu129" ]] || [[ "$DESIRED_CUDA" == "cu130" ]]; then | ||
| TRITON_CONSTRAINT="platform_system == 'Linux'" | ||
| fi | ||
| TRITON_CONSTRAINT="platform_system == 'Linux'" |
There was a problem hiding this comment.
Wouldn't that add triton constrain on CPU builds as well?
There was a problem hiding this comment.
No it changes platform_system == "Linux" and platform_machine == "x86_64" -> platform_system == "Linux"
nWEIdia
left a comment
nWEIdia
left a comment
There was a problem hiding this comment.
Good catch! With this CUDA aarch64 is truely in parity with CUDA x86_64.
|
@pytorchmergebot merge -f "lint is green " |
Merge startedYour change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes). Please use Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
|
@pytorchbot cherry-pick --onto release/2.9 --fixes "Critical CI fix" -c critical |
Since we have introduced CUDA aarch64 builds for all cuda versions we need to remove this constraint. This was missed by #162364 Proper constraint on triton should be: ``` Requires-Dist: triton==3.5.0; platform_system == "Linux" ``` not: ``` Requires-Dist: triton==3.5.0; platform_system == "Linux" and platform_machine == "x86_64" ``` Pull Request resolved: #165013 Approved by: https://github.com/Camyll, https://github.com/nWEIdia, https://github.com/tinglvv (cherry picked from commit 81dbeb0)
Cherry picking #165013The cherry pick PR is at #165022 and it is linked with issue Critical CI fix. The following tracker issues are updated: Details for Dev Infra teamRaised by workflow job |
CUDA aarch64 12.6 and 12.8 builds fix triton constraints (#165013) Since we have introduced CUDA aarch64 builds for all cuda versions we need to remove this constraint. This was missed by #162364 Proper constraint on triton should be: ``` Requires-Dist: triton==3.5.0; platform_system == "Linux" ``` not: ``` Requires-Dist: triton==3.5.0; platform_system == "Linux" and platform_machine == "x86_64" ``` Pull Request resolved: #165013 Approved by: https://github.com/Camyll, https://github.com/nWEIdia, https://github.com/tinglvv (cherry picked from commit 81dbeb0) Co-authored-by: atalman <atalman@fb.com>
) Since we have introduced CUDA aarch64 builds for all cuda versions we need to remove this constraint. This was missed by pytorch#162364 Proper constraint on triton should be: ``` Requires-Dist: triton==3.5.0; platform_system == "Linux" ``` not: ``` Requires-Dist: triton==3.5.0; platform_system == "Linux" and platform_machine == "x86_64" ``` Pull Request resolved: pytorch#165013 Approved by: https://github.com/Camyll, https://github.com/nWEIdia, https://github.com/tinglvv
7 participants
Since we have introduced CUDA aarch64 builds for all cuda versions we need to remove this constraint.
This was missed by #162364
Proper constraint on triton should be:
not: