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

[ROCM] Enable aiter attn backend for qwen3-next model (#32492)

Browse Source 
Signed-off-by: jennyyyyzhen <yzhen@hmc.edu>
This commit is contained in:
jennyyyyzhen
2026-01-31 01:03:57 -08:00
committed by GitHub
parent f68e3ea4e1
commit 527bcd14d4
2 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
docs/design/attention_backends.md
View File

@@ -168,7 +168,7 @@ Priority is **1 = highest** (tried first).
| `FLASH_ATTN` | FA3* | fp16, bf16 | `auto`, `bfloat16`, `fp8`, `fp8_e4m3`, `fp8_e5m2` | %16 | Any | ✅ | ❌ | All | 9.x |
| `FLASH_ATTN_DIFFKV` | | fp16, bf16 | `auto` | Any | Any | ❌ | ❌ | Decoder | Any |
| `FLEX_ATTENTION` | | fp16, bf16, fp32 | `auto`, `bfloat16` | Any | Any | ❌ | ✅ | Decoder, Encoder Only | Any |
| `ROCM_AITER_FA` | | fp16, bf16 | `auto` | %16 | 64, 128, 256 | ❌ | ❌ | Decoder | N/A |
| `ROCM_AITER_FA` | | fp16, bf16 | `auto` | 16, 32 | 64, 128, 256 | ❌ | ❌ | Decoder | N/A |
| `ROCM_AITER_UNIFIED_ATTN` | | fp16, bf16 | `auto` | Any | Any | ❌ | ❌ | Decoder | N/A |
| `ROCM_ATTN` | | fp16, bf16, fp32 | `auto` | 16, 32, 544 | 32, 64, 96, 128, 160, 192, 224, 256 | ❌ | ❌ | Decoder | N/A |
| `TREE_ATTN` | | fp16, bf16 | `auto` | %16 | 32, 64, 96, 128, 160, 192, 224, 256 | ❌ | ❌ | Decoder | Any |

2
vllm/v1/attention/backends/rocm_aiter_fa.py
View File

@@ -683,7 +683,7 @@ class AiterFlashAttentionBackend(AttentionBackend):
@staticmethod
def get_supported_kernel_block_sizes() -> list[int | MultipleOf]:
return [MultipleOf(16)]
return [16, 32]
@classmethod
def get_supported_head_sizes(cls) -> list[int]: