You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
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
When fp8 dtype is involved, Inductor may set min_elem_per_thread to be a positive value. This will force increasing XBLOCK even for a small xnumel (e.g. 1). Inductor will report an error later when sanity check the triton config.
The simple fix here is to just not let XBLOCK to be larger than xnumel.
Is this the right way to solve it or just a patch?
The stacked PR fails some internal tests: https://www.internalfb.com/diff/D64796587 . Those tests exists in OSS but we don't run it in OSS CI since we don't have h100...
This PR is a patch for quick fix those test failures.
A fundamental fix may need us to auditing all the scaling rules for BLOCK size and all kind of overriding (like min_elem_per_thread).
You need to provide a reason for using force merge, in the format @pytorchbot merge -f 'Explanation'.
The explanation needs to be clear on why this is needed. Here are some good examples:
Bypass checks due to unrelated upstream failures from ...
This is a minor fix to ..., which shouldn't break anything
Your change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes). Please use -f as last resort and instead consider -i/--ignore-current to continue the merge ignoring current failures. This will allow currently pending tests to finish and report signal before the merge.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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):
When fp8 dtype is involved, Inductor may set min_elem_per_thread to be a positive value. This will force increasing XBLOCK even for a small xnumel (e.g. 1). Inductor will report an error later when sanity check the triton config.
The simple fix here is to just not let XBLOCK to be larger than xnumel.
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @yf225 @chenyang78 @kadeng @muchulee8 @ColinPeppler @amjames @desertfire @chauhang @aakhundov