xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-23 14:15:01 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Misc] fix tests failure by using current_platform (#25825)

Browse Source 
Signed-off-by: Juechen Liu <jueliu@meta.com>
Signed-off-by: yewentao256 <zhyanwentao@126.com>
This commit is contained in:
Juechen Liu 2025-09-28 21:18:57 -07:00 committed by yewentao256
parent cf0a7912ca
commit eb447aff56
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/attention/ops/triton_reshape_and_cache_flash.py
View File

@ -137,7 +137,7 @@ def triton_reshape_and_cache_flash(
# heuristics instead of autotuning
TILE_SIZE = min(2048, triton.next_power_of_2(n))
if torch.version.hip or torch.version.xpu:
if current_platform.is_rocm() or current_platform.is_xpu():
num_stages = 4
num_warps = 8
else: # cuda