Automatically bind CPU OMP Threads of a rank to CPU ids of a NUMA node. (#17930)

Signed-off-by: Tsai, Louie <louie.tsai@intel.com>
Co-authored-by: Li, Jiang <bigpyj64@gmail.com>
This commit is contained in:
Louie Tsai
2025-06-10 06:22:05 +00:00
committed by GitHub
co-authored by Li, Jiang <bigpyj64@gmail.com>
parent 32b3946bb4
commit 9368cc90b2
6 changed files with 134 additions and 8 deletions
+3
View File
@@ -208,6 +208,9 @@ class CpuPlatform(Platform):
# Disable torch async compiling which won't work with daemonic processes
os.environ["TORCHINDUCTOR_COMPILE_THREADS"] = "1"
# Share the cpusets list among ranks by spawning process instead
os.environ["VLLM_WORKER_MULTIPROC_METHOD"] = "spawn"
# Intel OpenMP setting
ld_prealod_str = os.getenv("LD_PRELOAD", "")
if "libiomp5.so" in ld_prealod_str: