Fix GEMMK=1 kernel #543
Merged
Fix GEMMK=1 kernel #543
Conversation
This file contains 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
This was referenced May 27, 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Outdated suggestions cannot be applied.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.
This solves a bug that was originally introduced in #274 (which introduced the GEMMK=1 kernel), reported in #340, but then incorrectly fixed in #341. It resurfaced again in #542 recently as more tuning results are coming in, in particular for Intel devices that seem to favour the GEMMK=1 kernel in certain cases.
The issue only shows up when MWG != NWG, and even then only in certain situations.
It is very likely also solving the issue with the Intel Arc A770 reported in #533, since that device also uses tuning parameters that were leading to this bug. Furthermore the logs from the CLBlast tests in that issue also correspond with what I reproduced locally. Therefore this PR also re-adds those tuning parameters.