Skip to content

opt: optimize musa_applyadagrad_op#262

Open
awexxxx wants to merge 2 commits into
MooreThreads:mainfrom
awexxxx:opt/applyadagrad
Open

opt: optimize musa_applyadagrad_op#262
awexxxx wants to merge 2 commits into
MooreThreads:mainfrom
awexxxx:opt/applyadagrad

Conversation

@awexxxx
Copy link
Copy Markdown
Contributor

@awexxxx awexxxx commented May 20, 2026

  • 优化掉假融合问题,launch kerner时间缩短3.78倍
  • 新kernel的运行时间是原始mudnn总和的1/1.98; 峰值带宽提升2倍

awexxxx added 2 commits May 21, 2026 10:47
- 优化掉假融合问题,launch kerner时间缩短3.78倍
- 新kernel的运行时间是原始mudnn总和的1/1.98; 峰值带宽提升2倍
@awexxxx awexxxx force-pushed the opt/applyadagrad branch from 22eda87 to 0807332 Compare May 21, 2026 02:48
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant