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 8, 16:30 UTC
Update DeepGEMM version pin in Dockerfile to match #32479 (#33935)
Signed-off-by: Zifei Tong <zifeitong@gmail.com> Signed-off-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com> Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
zifeitong committed
Feb 7, 2026 at 05:30 UTC
52181baaea01f8cc64e2ffde7761a3ff3b7e4d3e
3 files changed
+3
-3
docker/Dockerfile
+1
-1
@@ -320,7 +320,7 @@ WORKDIR /workspace
320
321
# Build DeepGEMM wheel
322
# Default moved here from tools/install_deepgemm.sh for centralized version management
323
-ARG DEEPGEMM_GIT_REF=594953acce41793ae00a1233eb516044d604bcb6
323
+ARG DEEPGEMM_GIT_REF=477618cd51baffca09c4b0b87e97c03fe827ef03
324
COPY tools/install_deepgemm.sh /tmp/install_deepgemm.sh
325
RUN --mount=type=cache,target=/root/.cache/uv \
326
mkdir -p /tmp/deepgemm/dist && \
docker/versions.json
+1
-1
@@ -50,7 +50,7 @@
50
"default": "cuda"
51
},
52
"DEEPGEMM_GIT_REF": {
53
- "default": "594953acce41793ae00a1233eb516044d604bcb6"
53
+ "default": "477618cd51baffca09c4b0b87e97c03fe827ef03"
54
},
55
"PPLX_COMMIT_HASH": {
56
"default": "12cecfd"
tools/install_deepgemm.sh
+1
-1
@@ -6,7 +6,7 @@ set -e
6
7
# Default values
8
DEEPGEMM_GIT_REPO="https://github.com/deepseek-ai/DeepGEMM.git"
9
-DEEPGEMM_GIT_REF="0f5f2662027f0db05d4e3f6a94e56e2d8fc45c51"
9
+DEEPGEMM_GIT_REF="477618cd51baffca09c4b0b87e97c03fe827ef03"
10
WHEEL_DIR=""
11
12
# Parse command line arguments