Skip to content

[vLLM][moe] Route blockwise FP8 fused MoE to Triton#7569

Draft
sarah12121212 wants to merge 2 commits into
moe2-workspacemanager-xpu-initfrom
moe2-fp8-moe
Draft

[vLLM][moe] Route blockwise FP8 fused MoE to Triton#7569
sarah12121212 wants to merge 2 commits into
moe2-workspacemanager-xpu-initfrom
moe2-fp8-moe

Conversation

@sarah12121212

@sarah12121212 sarah12121212 commented Jul 23, 2026

Copy link
Copy Markdown
Contributor

Summary

  • For blockwise quantization key pair activation_key == kFp8Dynamic128Sym and weight_key == kFp8Static128BlockSym route fused MoE through Triton instead of the native Xe2 grouped-GEMM path, which accepts only 1D per-expert scales and rejects multidimensional scale grids.

  • Unskip the 44 affected Xe2 tests.

Blocked by #7462 (it needs to be merged first)
Relates to issue #7574

Testing

vLLM MoE suite on B580 (tested locally):

2,987 tests
794 passed
2,193 skipped
0 failures

vLLM MoE suite on PVC max 1100:

@sarah12121212
sarah12121212 changed the base branch from main to moe2-workspacemanager-xpu-init July 24, 2026 14:20
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