Fix stats #231
Merged
Fix stats #231
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
LAhmos
approved these changes
Jul 5, 2023
|
this PR can't be merged until we add the traces for the new benchmark to |
JRPan
added a commit
to JRPan/accel-sim-framework
that referenced
this pull request
Sep 3, 2024
* fix shared memory and l1d cache bankconflict stat and add shmem benchmark * this change requires dev branch of app collection --------- Co-authored-by: JRPan <[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Outdated suggestions cannot be applied.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.
Adds microbenchmark / correlation data for shared memory bank conflicts counter. Relies on accel-sim/gpgpu-sim_distribution#58 and accel-sim/gpu-app-collection#16 part of #226
For the ubench with bank conflicts:
HW: 260593 bank conflicts, Simulator: 253952 (97%)
For the ubench without bank conflicts:
HW: 2484, Simulator 0
It's odd that for some reason nsight compute shows bank conflicts on the example that has none.