biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Bugfix][MTP][Sparse MLA] Allow sparse MLA with MTP to run with FULL cudagraphs (#34457)

Browse Source 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
This commit is contained in:
Matthew Bonanni
2026-02-17 14:01:27 -05:00
committed by GitHub
parent 1e4a084c8e
commit dc5fa77a4e
2 changed files with 17 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
docs/design/cuda_graphs.md
View File

@@ -182,6 +182,7 @@ The following table lists backends that support full CUDA Graphs at the time of
| FlashInfer | `UNIFORM_SINGLE_TOKEN_DECODE` | Will be set to `UNIFORM_BATCH` when using TRTLLM attention on Blackwell |
| FlashMLA | `UNIFORM_BATCH` | |
| FlashInferMLA | `UNIFORM_BATCH` | |
| FlashInferMLASparse | `UNIFORM_BATCH` | |
| AITER MLA | `UNIFORM_SINGLE_TOKEN_DECODE` | |
| CUTLASS MLA | `UNIFORM_SINGLE_TOKEN_DECODE` | |
| Mamba attention| `UNIFORM_SINGLE_TOKEN_DECODE` | |