Skip to content

Conversation

@k50112113
Copy link
Contributor

This PR includes the following changes:

  1. gemm_a8w8_blockscale
  • added splitk version
  • tunned config for specific shapes
  1. gemm_a8w8_per_token_scale
  • added per token scale with no splitk
  • added splitk version
  • tunned config for specific shapes

@k50112113 k50112113 requested review from azaidy and rahulbatra85 July 23, 2025 16:23
@k50112113 k50112113 force-pushed the shaoclee/triton_gemm_a8w8_dev branch from 7b86b6a to 0fdb5b0 Compare July 23, 2025 21:40
Copy link
Contributor

@rahulbatra85 rahulbatra85 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please see comments

@k50112113 k50112113 force-pushed the shaoclee/triton_gemm_a8w8_dev branch from db474b8 to 2c9c94e Compare July 25, 2025 14:11
@rahulbatra85 rahulbatra85 merged commit 6b92d30 into main Jul 25, 2025
13 checks passed
@rahulbatra85 rahulbatra85 deleted the shaoclee/triton_gemm_a8w8_dev branch July 25, 2025 19:25
cagrikymk pushed a commit that referenced this pull request Jul 30, 2025
* tune gemm_a8w8_blockscale and gemm_a8w8_per_token_scale

* tune

* fix typo

* configs for MI300

* reformat

* black reformat
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants