Skip to content

Conversation

ibahmed-oai
Copy link
Contributor

Adding scatter and gather kernel to map token to experts and perform the inverse op.
Also added moe matmul kernels. These kernels speed up prefill.
These kernels are not used by the metal backend yet. In the next PR, they weill be

@davecummings davecummings merged commit 7e31d93 into openai:main Sep 18, 2025
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.

2 participants