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

[Hardware][Powerpc] Fix VLLM_CPU_OMP_THREADS_BIND="auto" low CPU utilization for Power (#27734)

Browse Source 
Signed-off-by: Akash Kaothalkar <akash.kaothalkar@ibm.com>
Co-authored-by: Akash Kaothalkar <akash.kaothalkar@ibm.com>
This commit is contained in:
Akash kaothalkar
2025-10-31 13:15:26 +05:30
committed by GitHub
parent b2e65cb4a7
commit 36960501d3
1 changed files with 2 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
vllm/platforms/cpu.py
View File

@@ -316,7 +316,8 @@ class CpuPlatform(Platform):
if ( if (
platform.system() == "Linux" platform.system() == "Linux"
and Platform.get_cpu_architecture() == CpuArchEnum.ARM and Platform.get_cpu_architecture()
in (CpuArchEnum.ARM, CpuArchEnum.POWERPC)
and not ("libomp" in ld_preload_str or "libgomp" in ld_preload_str) and not ("libomp" in ld_preload_str or "libgomp" in ld_preload_str)
): ):
# We need to LD_PRELOAD PyTorch's libgomp, otherwise only # We need to LD_PRELOAD PyTorch's libgomp, otherwise only