iHore/vLLM
8
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

[ROCm] Enable gluon paged MQA logits on gfx950 (MI355X) (#42062)

Browse Source 
Signed-off-by: Frida Andersson <fanderss@amd.com>
This commit is contained in:
frida-andersson
2026-05-14 15:39:26 +00:00
committed by GitHub
parent f60c6b33a5
commit f07b1da797
1 changed files with 3 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
vllm/v1/attention/ops/rocm_aiter_mla_sparse.py
+3 -2
View File
@@ -16,9 +16,10 @@ from vllm.v1.attention.backends.mla.indexer import DeepseekV32IndexerMetadata
from vllm.v1.attention.ops.common import pack_seq_triton, unpack_seq_triton
if current_platform.is_rocm():
from vllm.platforms.rocm import _ON_GFX942
from vllm.platforms.rocm import _ON_GFX942, _ON_GFX950
else:
_ON_GFX942 = False
_ON_GFX950 = False
@triton.jit
@@ -385,7 +386,7 @@ def rocm_fp8_paged_mqa_logits(
aiter_paged_mqa_logits_module = paged_mqa_logits_module()
if aiter_paged_mqa_logits_module is not None:
if _ON_GFX942:
if _ON_GFX942 or _ON_GFX950:
deepgemm_fp8_paged_mqa_logits = (
aiter_paged_mqa_logits_module.deepgemm_fp8_paged_mqa_logits
)