@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only / commits / 418cb3b93f
Fork 0

A high-throughput and memory-efficient inference and serving engine for LLMs

Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git

Synced Nov 16, 05:45 UTC

[Bugfix][Hardware][CPU] Fix intel-omp version to avoid segfault (#10700)

Signed-off-by: jiang1.li <jiang1.li@intel.com>

Li, Jiang committed Nov 27, 2024 at 19:55 UTC 418cb3b93fbf85f0735b5c0ed3f62d4b36808968
1 file changed +1 -1
Dockerfile.cpu
+1 -1
@@ -16,7 +16,7 @@ RUN --mount=type=cache,target=/var/cache/apt \
16 # intel-openmp provides additional performance improvement vs. openmp
17 # tcmalloc provides better memory allocation efficiency, e.g, holding memory in caches to speed up access of commonly-used objects.
18 RUN --mount=type=cache,target=/root/.cache/pip \
19 - pip install intel-openmp
19 + pip install intel-openmp==2025.0.1
20
21 ENV LD_PRELOAD="/usr/lib/x86_64-linux-gnu/libtcmalloc_minimal.so.4:/usr/local/lib/libiomp5.so"
22