[Executorch][Portable] Dont upcast to double for sigmoid #6892
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
Upcasting to double for compute precision may not be aten compliant. Reason for internal test change: Apparently running on broadwell CPU vs test runner with Cooper lake gives different results for this change. Without this change: Both broadwell and Cooper lake will produce "Once upon a time, there was a little" With this change: Broadwell still produces "Once upon a time, there was a little", while Cooperlake produces "Once upon a time, there was a girl". So one possibility is that that some XNNPACK kernel for Cooper lake is produces slightly different numerical result that propagates through. Still landing this change since upcasting to double for compute, does not seem necessary. Differential Revision: [D65928920](https://our.internmc.facebook.com/intern/diff/D65928920/) [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/6892
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ✅ You can merge normally! (1 Unrelated Failure)As of commit 8eabd3e with merge base dc41596 ( FLAKY - The following job failed but was likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
kimishpatel
added a commit
that referenced
this pull request
Nov 15, 2024
Upcasting to double for compute precision may not be aten compliant. Reason for internal test change: Apparently running on broadwell CPU vs test runner with Cooper lake gives different results for this change. Without this change: Both broadwell and Cooper lake will produce "Once upon a time, there was a little" With this change: Broadwell still produces "Once upon a time, there was a little", while Cooperlake produces "Once upon a time, there was a girl". So one possibility is that that some XNNPACK kernel for Cooper lake is produces slightly different numerical result that propagates through. Still landing this change since upcasting to double for compute, does not seem necessary. Differential Revision: [D65928920](https://our.internmc.facebook.com/intern/diff/D65928920/) ghstack-source-id: 253799652 Pull Request resolved: #6892
|
This pull request was exported from Phabricator. Differential Revision: D65928920 |
manuelcandales
approved these changes
Nov 15, 2024
Upcasting to double for compute precision may not be aten compliant. Reason for internal test change: Apparently running on broadwell CPU vs test runner with Cooper lake gives different results for this change. Without this change: Both broadwell and Cooper lake will produce "Once upon a time, there was a little" With this change: Broadwell still produces "Once upon a time, there was a little", while Cooperlake produces "Once upon a time, there was a girl". So one possibility is that that some XNNPACK kernel for Cooper lake is produces slightly different numerical result that propagates through. Still landing this change since upcasting to double for compute, does not seem necessary. Differential Revision: [D65928920](https://our.internmc.facebook.com/intern/diff/D65928920/) [ghstack-poisoned]
kimishpatel
added a commit
that referenced
this pull request
Nov 15, 2024
Pull Request resolved: #6892 Upcasting to double for compute precision may not be aten compliant. Reason for internal test change: Apparently running on broadwell CPU vs test runner with Cooper lake gives different results for this change. Without this change: Both broadwell and Cooper lake will produce "Once upon a time, there was a little" With this change: Broadwell still produces "Once upon a time, there was a little", while Cooperlake produces "Once upon a time, there was a girl". So one possibility is that that some XNNPACK kernel for Cooper lake is produces slightly different numerical result that propagates through. Still landing this change since upcasting to double for compute, does not seem necessary. ghstack-source-id: 253832495 @exported-using-ghexport Differential Revision: [D65928920](https://our.internmc.facebook.com/intern/diff/D65928920/)
|
This pull request was exported from Phabricator. Differential Revision: D65928920 |
manuelcandales
approved these changes
Nov 15, 2024
facebook-github-bot
merged commit 6968d01
into
gh/kimishpatel/142/base
38 of 41 checks passed
facebook-github-bot
had a problem deploying
to
cherry-pick-bot
GitHub Actions
Failure
kirklandsign
added a commit
that referenced
this pull request
Nov 18, 2024
Pull Request resolved: #6892 Upcasting to double for compute precision may not be aten compliant. Reason for internal test change: Apparently running on broadwell CPU vs test runner with Cooper lake gives different results for this change. Without this change: Both broadwell and Cooper lake will produce "Once upon a time, there was a little" With this change: Broadwell still produces "Once upon a time, there was a little", while Cooperlake produces "Once upon a time, there was a girl". So one possibility is that that some XNNPACK kernel for Cooper lake is produces slightly different numerical result that propagates through. Still landing this change since upcasting to double for compute, does not seem necessary. ghstack-source-id: 253832495 @exported-using-ghexport Differential Revision: [D65928920](https://our.internmc.facebook.com/intern/diff/D65928920/) Co-authored-by: Kimish Patel <[email protected]>
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.
Stack from ghstack (oldest at bottom):
Upcasting to double for compute precision may not be aten compliant.
Reason for internal test change:
Apparently running on broadwell CPU vs test runner with Cooper lake gives
different results for this change.
Without this change:
Both broadwell and Cooper lake will produce "Once upon a time, there was a
little"
With this change:
Broadwell still produces "Once upon a time, there was a little", while
Cooperlake produces "Once upon a time, there was a girl".
So one possibility is that that some XNNPACK kernel for Cooper lake is produces
slightly different numerical result that propagates through.
Still landing this change since upcasting to double for compute, does not seem
necessary.
Differential Revision: D65928920