Updated SGEMV ramps. #5551
Open
Updated SGEMV ramps. #5551
+14
−5
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
Contributor
|
cc: @martin-frbg @Mousius |
aditew01
reviewed
Nov 25, 2025
| : (MN < 1050625L) ? MIN(ncpu, 40) | ||
| : ncpu; | ||
| #else | ||
| return (MN < 25600L) ? 1 |
Contributor
There was a problem hiding this comment.
do we want to guard it for NEOVERSEV2 / NEOVERSEN2?
Author
There was a problem hiding this comment.
Discussed offline, happy that this is done by the calling function. No changes needed.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
Significant performance improvements are gained by the proposed changes on both c7g (NEOVERSEV1) and c8g (NEOVERSEV2) instances. To reproduce these values you need to run with OMP_ADAPTIVE=1. The plots below show the average time taken for 10000 iterations of sgemv operations on increasing square matrix/vector sizes, from 2x2 through to 1024x1024. The x axis reaches 2046 as we first run sgemv without transposition, then with. I've also include plots of the ratio, relative to the original stats (lower is better). These generate the following stats:
Geometric mean for c7g_sgemv.txt: 0.890437968914142
Geometric mean for c8g_sgemv.txt: 0.7884951978206536