[inductor] enable bf32 for mkldnn linear pointwise/binary in inductor #127294
Conversation
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
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/127294
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 New Failure, 2 Unrelated FailuresAs of commit ac9ed89 with merge base 4015166 ( NEW FAILURE - The following job has failed:
FLAKY - The following job failed but was likely due to flakiness present on trunk:
UNSTABLE - The following job is marked as unstable, possibly due to flakiness on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This was referenced May 28, 2024
jgong5
reviewed
May 29, 2024
Comment on lines
786
to
790
| if not hasattr(add_node.args[1], "meta"): | ||
| # May add an "int" | ||
| # We meet this issue while enabling bf32 for test_linear_unary | ||
| # the hardsimoid case will add "3" here and cannot check meta | ||
| # TODO: Further investigate on this issue |
There was a problem hiding this comment.
Is it for the bias node or something else? Can you double check? Why didn't we get the problem in the past?
There was a problem hiding this comment.
It should be another input node for "add".
We did not meet this in the past because we only enabled and tested lp for mkldnn fusion. For lp case, there will be a "to_dtype" inserted by autocast and will not match this pattern.
I am trying to find a case to satisfy this pattern by raising a runtime error if this function will return True.
#127597
jgong5
approved these changes
Jun 7, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
When
torch.backends.mkldnn.matmul.fp32_precision=='bf16', we also enabled mkldnn linear in inductor path and allow to run with bf16 computation data type.Testplan:
Stack from ghstack (oldest at bottom):
cc @jgong5 @mingfeima @XiaobingSuper @sanchitintel @ashokei @jingxu10 @voznesenskym @penguinwu @EikanWang @Guobing-Chen @blzheng @wenzhe-nrv @jiayisunx @peterbell10 @ipiszy @yf225 @chenyang78 @kadeng @muchulee8 @ColinPeppler @amjames @desertfire @chauhang