[Triton] [Inductor] Pruned failed compilations from Autotuning candidates#162673
Closed
njriasan wants to merge 1 commit intopytorch:mainfrom
Closed
[Triton] [Inductor] Pruned failed compilations from Autotuning candidates#162673njriasan wants to merge 1 commit intopytorch:mainfrom
njriasan wants to merge 1 commit intopytorch:mainfrom
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/162673
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (2 Unrelated Failures)As of commit 28907b4 with merge base f654cff ( BROKEN TRUNK - The following jobs failed but were present on the merge base:👉 Rebase onto the `viable/strict` branch to avoid these failures
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D82172207 |
njriasan
force-pushed
the
export-D82172207
branch
from
fd45ee6 to
aa49bfd
Compare
njriasan
added a commit
to njriasan/pytorch
that referenced
this pull request
Sep 11, 2025
…ates (pytorch#162673) Summary: When exahaustively autotuning a new template you may hit situations that lead to compilation failures. This template will still attempt to autotune because nothing was marking this as failed and in my experiments lead to a crash/segfault if I didn't set `TORCHINDUCTOR_AUTOTUNE_IN_SUBPROC=1`. To help eliminate this issue this PR marks any template that fails to compile as "failed" and then removes all of the failed templates from the choice candidates. In the case where it would have just failed to compile twice, this should at least reduce compilation time. Test Plan: Tested locally when experminenting with the new blackwell templates and a Triton version that contains a bug related to `num_warps < 4`. Rollback Plan: Differential Revision: D82172207
…ates (pytorch#162673) Summary: When exahaustively autotuning a new template you may hit situations that lead to compilation failures. This template will still attempt to autotune because nothing was marking this as failed and in my experiments lead to a crash/segfault if I didn't set `TORCHINDUCTOR_AUTOTUNE_IN_SUBPROC=1`. To help eliminate this issue this PR marks any template that fails to compile as "failed" and then removes all of the failed templates from the choice candidates. In the case where it would have just failed to compile twice, this should at least reduce compilation time. Test Plan: Tested locally when experminenting with the new blackwell templates and a Triton version that contains a bug related to `num_warps < 4`. Rollback Plan: Differential Revision: D82172207
njriasan
force-pushed
the
export-D82172207
branch
from
aa49bfd to
28907b4
Compare
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D82172207 |
1 similar comment
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D82172207 |
PaulZhang12
approved these changes
Sep 11, 2025
Contributor
PaulZhang12
left a comment
PaulZhang12
left a comment
There was a problem hiding this comment.
This is an awesome fix! Thank you, cc @eellison for any thoughts
Contributor
Author
|
@pytorchbot merge |
Collaborator
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
mlazos
reviewed
Sep 11, 2025
| else "max_autotune_conv_backends" | ||
| ) | ||
| raise NoValidChoicesError( | ||
| return NoValidChoicesError( |
Contributor
There was a problem hiding this comment.
It might be good to modify this to indicate the reason for no valid choices, I think this could be a good idea (e.g. no compileable choices vs no choices at the beginning)
I had some plans to update this since this is the most common error I've seen with users by far. They usually end up adding aten, but it would be useful to know why.
Contributor
Author
There was a problem hiding this comment.
Happy to submit a followup. Thanks for the suggestion.
mlazos
approved these changes
Sep 11, 2025
markc-614
pushed a commit
to markc-614/pytorch
that referenced
this pull request
Sep 17, 2025
…ates (pytorch#162673) Summary: When exahaustively autotuning a new template you may hit situations that lead to compilation failures. This template will still attempt to autotune because nothing was marking this as failed and in my experiments lead to a crash/segfault if I didn't set `TORCHINDUCTOR_AUTOTUNE_IN_SUBPROC=1`. To help eliminate this issue this PR marks any template that fails to compile as "failed" and then removes all of the failed templates from the choice candidates. In the case where it would have just failed to compile twice, this should at least reduce compilation time. Test Plan: Tested locally when experminenting with the new blackwell templates and a Triton version that contains a bug related to `num_warps < 4`. Rollback Plan: Differential Revision: D82172207 Pull Request resolved: pytorch#162673 Approved by: https://github.com/PaulZhang12, https://github.com/mlazos
mansiag05
pushed a commit
to mansiag05/pytorch
that referenced
this pull request
Sep 22, 2025
…ates (pytorch#162673) Summary: When exahaustively autotuning a new template you may hit situations that lead to compilation failures. This template will still attempt to autotune because nothing was marking this as failed and in my experiments lead to a crash/segfault if I didn't set `TORCHINDUCTOR_AUTOTUNE_IN_SUBPROC=1`. To help eliminate this issue this PR marks any template that fails to compile as "failed" and then removes all of the failed templates from the choice candidates. In the case where it would have just failed to compile twice, this should at least reduce compilation time. Test Plan: Tested locally when experminenting with the new blackwell templates and a Triton version that contains a bug related to `num_warps < 4`. Rollback Plan: Differential Revision: D82172207 Pull Request resolved: pytorch#162673 Approved by: https://github.com/PaulZhang12, https://github.com/mlazos
cleonard530
pushed a commit
to cleonard530/pytorch
that referenced
this pull request
Sep 22, 2025
…ates (pytorch#162673) Summary: When exahaustively autotuning a new template you may hit situations that lead to compilation failures. This template will still attempt to autotune because nothing was marking this as failed and in my experiments lead to a crash/segfault if I didn't set `TORCHINDUCTOR_AUTOTUNE_IN_SUBPROC=1`. To help eliminate this issue this PR marks any template that fails to compile as "failed" and then removes all of the failed templates from the choice candidates. In the case where it would have just failed to compile twice, this should at least reduce compilation time. Test Plan: Tested locally when experminenting with the new blackwell templates and a Triton version that contains a bug related to `num_warps < 4`. Rollback Plan: Differential Revision: D82172207 Pull Request resolved: pytorch#162673 Approved by: https://github.com/PaulZhang12, https://github.com/mlazos
dsashidh
pushed a commit
to dsashidh/pytorch
that referenced
this pull request
Sep 26, 2025
…ates (pytorch#162673) Summary: When exahaustively autotuning a new template you may hit situations that lead to compilation failures. This template will still attempt to autotune because nothing was marking this as failed and in my experiments lead to a crash/segfault if I didn't set `TORCHINDUCTOR_AUTOTUNE_IN_SUBPROC=1`. To help eliminate this issue this PR marks any template that fails to compile as "failed" and then removes all of the failed templates from the choice candidates. In the case where it would have just failed to compile twice, this should at least reduce compilation time. Test Plan: Tested locally when experminenting with the new blackwell templates and a Triton version that contains a bug related to `num_warps < 4`. Rollback Plan: Differential Revision: D82172207 Pull Request resolved: pytorch#162673 Approved by: https://github.com/PaulZhang12, https://github.com/mlazos
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
5 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:
When exahaustively autotuning a new template you may hit situations that lead to compilation failures. This template will still attempt to autotune because nothing was marking this as failed and in my experiments lead to a crash/segfault if I didn't set
TORCHINDUCTOR_AUTOTUNE_IN_SUBPROC=1.To help eliminate this issue this PR marks any template that fails to compile as "failed" and then removes all of the failed templates from the choice candidates. In the case where it would have just failed to compile twice, this should at least reduce compilation time.
Test Plan:
Tested locally when experminenting with the new blackwell templates and a Triton version that contains a bug related to
num_warps < 4.Rollback Plan:
Differential Revision: D82172207
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @chenyang78 @kadeng @muchulee8 @amjames @chauhang @aakhundov @coconutruben @mlazos