Skip to content

CMS implementation for 192x256x64TN#2623

Merged
jfactory07 merged 8 commits into
hipblaslt_common_cms_devfrom
users/jzhou/cms-192x256x64TN
Nov 18, 2025
Merged

CMS implementation for 192x256x64TN#2623
jfactory07 merged 8 commits into
hipblaslt_common_cms_devfrom
users/jzhou/cms-192x256x64TN

Conversation

@jfactory07
Copy link
Copy Markdown
Contributor

Motivation

CMS implementation for 192x256x64TN

Test Result

test for 3072x4096x8192
got 7% uplift

Submission Checklist

talumbau and others added 5 commits November 11, 2025 08:08
## Motivation

We are about to land a significant number of custom schedules for
various tile sizes and transpose orientations. By using the convention
"a schedule is described in a function" we can reasonably scale the code
out, instead of a chain of if/elif blocks that spans hundreds or
thousands of lines. Initially we keep the existing schedules in
functions inside CustomSchedule.py, but we can refactor as needed.

## Technical Details


## Test Plan

TENSILELITE_CLIENT_ARGS="--build-type Debug --gpu-targets gfx950
--clean" tox -e py39 -- Tensile/Tests -k custom_mainloop_scheduling.yaml
## Test Result

<!-- Briefly summarize test outcomes. -->
@msujon-AMD
Copy link
Copy Markdown
Collaborator

Pls resolve the conflicts and upload test logfiles in our directory.

@jfactory07
Copy link
Copy Markdown
Contributor Author

Pls resolve the conflicts and upload test logfiles in our directory.

Done.

@msujon-AMD
Copy link
Copy Markdown
Collaborator

CMS test seems failing, would you pls investigate it?

@jfactory07
Copy link
Copy Markdown
Contributor Author

CMS test seems failing, would you pls investigate it?

@msujon-AMD pass now.

@jfactory07 jfactory07 merged commit 4e1c92c into hipblaslt_common_cms_dev Nov 18, 2025
27 checks passed
@jfactory07 jfactory07 deleted the users/jzhou/cms-192x256x64TN branch November 18, 2025 01:16
smalekta added a commit that referenced this pull request Nov 18, 2025
jfactory07 added a commit that referenced this pull request Nov 19, 2025
jfactory07 added a commit that referenced this pull request Nov 19, 2025
jfactory07 added a commit that referenced this pull request Nov 19, 2025
jfactory07 added a commit that referenced this pull request Nov 19, 2025
@jfactory07 jfactory07 restored the users/jzhou/cms-192x256x64TN branch November 20, 2025 05:30
minsukim-amd pushed a commit that referenced this pull request Nov 25, 2025
CMS implementation for 192x256x64TN

test for 3072x4096x8192
got 7% uplift

- [ ] Look over the contributing guidelines at
https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.

---------

Co-authored-by: T.J. Alumbaugh <talumbau@users.noreply.github.com>
minsukim-amd pushed a commit that referenced this pull request Nov 25, 2025
b-shi pushed a commit that referenced this pull request Dec 12, 2025
## Motivation

CMS implementation for 192x256x64TN

## Test Result

test for 3072x4096x8192
got 7% uplift

## Submission Checklist

- [ ] Look over the contributing guidelines at
https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.

---------

Co-authored-by: T.J. Alumbaugh <talumbau@users.noreply.github.com>
b-shi pushed a commit that referenced this pull request Dec 12, 2025
ammallya pushed a commit that referenced this pull request Feb 3, 2026
This reverts commit 19caeff.

[ROCm/composable_kernel commit: 833ae1d]
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants