Fix Numeric Error in SBGEMM Kernel for NEOVERSEV2 when DYNAMIC_ARCH=1
#5129
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
… to `NEOVERSEV1` SBGEMM kernel Signed-off-by: Ye Tao <ye.tao@arm.com>
|
Ah, did not think of that in time - we'd need to make the V2 a full-fledged TARGET for this to work (parameter assignment and all) which we had initially wanted to avoid in order to keep the total number of DYNAMIC_ARCH configurations (and library size) small |
# for free
to join this conversation on GitHub.
Already have an account?
# 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Outdated suggestions cannot be applied.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.
In our previous PR (#5108), we introduced an SBGEMM kernel optimized for NEOVERSEV1 (SVE-256 bit). However, when running on NEOVERSEV2 (SVE-128 bit) under the dynamic architecture setting, the SBGEMM kernel incorrectly falls back to the NEOVERSEV1 implementation due to this line of code. This fallback behavior leads to numeric errors.
This PR addresses the issue by explicitly adding NEOVERSEV2 to DYNAMIC_ARCH, ensuring that the correct SBGEMM kernel is selected for NEOVERSEV2 instead of defaulting to NEOVERSEV1.