xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-07-10 01:27:11 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Perf] Enable separate shared_experts stream only for CUDA (#30085)

Browse Source 
Signed-off-by: Alexander Matveev <amatveev@redhat.com>
This commit is contained in:
Alexander Matveev 2025-12-04 19:03:17 -05:00 committed by GitHub
parent 690cc3ef20
commit 4470ee2f90
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 2 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
vllm/model_executor/layers/fused_moe/layer.py
View File

@ -863,7 +863,8 @@ class FusedMoE(CustomOp):
use_chunked_impl: bool, use_chunked_impl: bool,
) -> tuple[bool, torch.Tensor | None]: ) -> tuple[bool, torch.Tensor | None]:
use_shared_experts_stream = ( use_shared_experts_stream = (
has_separate_shared_experts current_platform.is_cuda()
and has_separate_shared_experts
and not use_chunked_impl and not use_chunked_impl
and self.shared_experts_stream is not None and self.shared_experts_stream is not None
and ( and (