[inductor] add subsystem to pattern matcher#163922
Closed
avikchaudhuri wants to merge 1 commit intopytorch:mainfrom
Closed
[inductor] add subsystem to pattern matcher#163922avikchaudhuri wants to merge 1 commit intopytorch:mainfrom
avikchaudhuri wants to merge 1 commit intopytorch:mainfrom
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/163922
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 44a4c64 with merge base d7491fb ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
Contributor
|
@avikchaudhuri has exported this pull request. If you are a Meta employee, you can view the originating diff in D83306676. |
eellison
reviewed
Sep 26, 2025
Contributor
eellison
left a comment
eellison
left a comment
There was a problem hiding this comment.
nice ! but lets wait for test failures to be fixed..
| @@ -1,5 +1,6 @@ | |||
| # mypy: allow-untyped-defs | |||
| import copy | |||
| import functools | |||
Contributor
There was a problem hiding this comment.
pre_grad_passes doesnt exist yet in bisector - mind adding it ?
Contributor
Author
There was a problem hiding this comment.
So basically add a line here?
Contributor
Author
There was a problem hiding this comment.
Hmm, I'll do this in a separate PR.
qq:
Looks like the bisector doesn't pick up something like:
with GraphTransformObserver(gm, "pre_grad_custom_pass"):
config.pre_grad_custom_pass(gm.graph)
so instead I have to do:
GraphTransformObserver(gm, "pre_grad_custom_pass").apply_graph_pass(
config.pre_grad_custom_pass
)
KP? @eellison
avikchaudhuri
force-pushed
the
export-D83306676
branch
from
c8c702c to
4e442e6
Compare
pytorch-bot bot
pushed a commit
that referenced
this pull request
Sep 26, 2025
Summary: Running a toy example through `torch.compile(fullgraph=True, backend="inductor")` with default inductor config, we tried to see what passes are run in each of pre-grad, joint-graph, and post-grad phases by printing out the subsystem in `GraphTransformObserver`. However the subsystem showed up as None in a bunch of transforms that were run in each of those phases, so this PR adds some additional annotations. Note that these annotations are probably not a complete set, since other transforms may run based on changes to the config that are not covered here. Hopefully this doesn't change behavior. However, I did notice that bisecting relies on disabling various phases, which means that while before some passes would *not* be disabled (because their subsystem was `None`, now they would). Test Plan: existing tests + manual test described in summary Differential Revision: D83306676
Contributor
|
@avikchaudhuri has exported this pull request. If you are a Meta employee, you can view the originating diff in D83306676. |
Summary: Pull Request resolved: pytorch#163922 Running a toy example through `torch.compile(fullgraph=True, backend="inductor")` with default inductor config, we tried to see what passes are run in each of pre-grad, joint-graph, and post-grad phases by printing out the subsystem in `GraphTransformObserver`. However the subsystem showed up as None in a bunch of transforms that were run in each of those phases, so this PR adds some additional annotations. Note that these annotations are probably not a complete set, since other transforms may run based on changes to the config that are not covered here. Hopefully this doesn't change behavior. However, I did notice that bisecting relies on disabling various phases, which means that while before some passes would *not* be disabled (because their subsystem was `None`, now they would). Test Plan: existing tests + manual test described in summary Differential Revision: D83306676
Contributor
|
@avikchaudhuri has exported this pull request. If you are a Meta employee, you can view the originating diff in D83306676. |
avikchaudhuri
force-pushed
the
export-D83306676
branch
from
4e442e6 to
44a4c64
Compare
jansel
approved these changes
Sep 27, 2025
Contributor
Author
|
@pytorchbot merge |
Contributor
|
@pytorchbot merge (Initiating merge automatically since Phabricator Diff has merged) |
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 |
jainapurva
pushed a commit
that referenced
this pull request
Sep 29, 2025
Summary: Running a toy example through `torch.compile(fullgraph=True, backend="inductor")` with default inductor config, I tried to see what passes are run in each of pre-grad, joint-graph, and post-grad phases by printing out the subsystem in `GraphTransformObserver`. However the subsystem showed up as None in a bunch of transforms that were run in each of those phases, so this PR adds some additional annotations. Note that these annotations are probably not a complete set, since other transforms may run based on changes to the config that are not covered here. Hopefully this doesn't change behavior. However, I did notice that bisecting relies on disabling various phases, which means that while before some passes would *not* be disabled (because their subsystem was `None`), now they would. Test Plan: existing tests + manual test described in summary Differential Revision: D83306676 Pull Request resolved: #163922 Approved by: https://github.com/jansel
maggiemoss
pushed a commit
to maggiemoss/pytorch
that referenced
this pull request
Sep 29, 2025
Summary: Running a toy example through `torch.compile(fullgraph=True, backend="inductor")` with default inductor config, I tried to see what passes are run in each of pre-grad, joint-graph, and post-grad phases by printing out the subsystem in `GraphTransformObserver`. However the subsystem showed up as None in a bunch of transforms that were run in each of those phases, so this PR adds some additional annotations. Note that these annotations are probably not a complete set, since other transforms may run based on changes to the config that are not covered here. Hopefully this doesn't change behavior. However, I did notice that bisecting relies on disabling various phases, which means that while before some passes would *not* be disabled (because their subsystem was `None`), now they would. Test Plan: existing tests + manual test described in summary Differential Revision: D83306676 Pull Request resolved: pytorch#163922 Approved by: https://github.com/jansel
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:
Running a toy example through
torch.compile(fullgraph=True, backend="inductor")with default inductor config, I tried to see what passes are run in each of pre-grad, joint-graph, and post-grad phases by printing out the subsystem inGraphTransformObserver. However the subsystem showed up as None in a bunch of transforms that were run in each of those phases, so this PR adds some additional annotations.Note that these annotations are probably not a complete set, since other transforms may run based on changes to the config that are not covered here.
Hopefully this doesn't change behavior. However, I did notice that bisecting relies on disabling various phases, which means that while before some passes would not be disabled (because their subsystem was
None), now they would.Test Plan: existing tests + manual test described in summary
Differential Revision: D83306676
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @chenyang78 @kadeng @muchulee8 @amjames @chauhang @aakhundov @coconutruben