Skip to content

Conversation

@jeremylt
Copy link
Member

Two changes here

  • Use memset when zeroing a vec on the device
  • Swap from the early return of some kernels to only taking action if the index is before the end of the vec (more normal pattern online in examples)

I'd be curious how much this helps the benchmarking @zatkins-dev

@jeremylt jeremylt merged commit 1731780 into main Feb 12, 2025
28 of 29 checks passed
@jeremylt jeremylt deleted the jeremy/vec-set branch February 12, 2025 19:59
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.

2 participants