New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Port Optimized Legacy All-Gather (Height Dim, Interleaved, Tile, Multi-Tile High Tensor) To All-Gather Async #18738

Open

SeanNijjar opened this issue Mar 6, 2025 · 0 comments

Assignees

Labels

host performance optimization llm_t3000 P1

Contributor

SeanNijjar commented Mar 6, 2025 •

edited

Loading

An additional configuration requiring support for height concat dim in all-gather

Known Config(s):

(Mixtral) [1,1,32,4096] [1,1,256,4096] 2 bfp8 Interleaved, L1, Tile

Related to #18730 - do after that one

SeanNijjar added host performance optimization llm_t3000 P1 labels

SeanNijjar self-assigned this

SeanNijjar mentioned this issue

T3K Tensor Parallel Models (CCL) Port to Fabric Master Issue #18709

Open

7 tasks

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