Skip to content

GitLab

Commit 8f8d56c3 authored by zhuwenwen's avatar zhuwenwen
Browse files

set TRITON_HIP_USE_NEW_STREAM_PIPELINE

parent 54944679
Show whitespace changes
Inline Side-by-side
Showing with 2 additions and 0 deletions
vllm/attention/ops/triton_decode_attention.py
View file @ 8f8d56c3
...@@ -28,6 +28,7 @@ Memory-efficient attention for decoding. ...@@ -28,6 +28,7 @@ Memory-efficient attention for decoding.
It supports page size >= 1. It supports page size >= 1.
""" """
import os
import logging import logging
import triton import triton
...@@ -36,6 +37,7 @@ import triton.language as tl ...@@ -36,6 +37,7 @@ import triton.language as tl
from vllm.platforms import current_platform from vllm.platforms import current_platform
is_hip_ = current_platform.is_rocm() is_hip_ = current_platform.is_rocm()
os.environ["TRITON_HIP_USE_NEW_STREAM_PIPELINE"] = f"0"
logger = logging.getLogger(__name__) logger = logging.getLogger(__name__)
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment