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 10, 04:00 UTC
[ROCM] Add ROCm image build to release pipeline (#31995)
Signed-off-by: Doug Lehr <douglehr@amd.com> Co-authored-by: Doug Lehr <douglehr@amd.com>
Douglas Lehr committed
Jan 15, 2026 at 05:01 UTC
c5891b54306669db6e78e7df32250b6eee7d34b5
3 files changed
+29
.buildkite/release-pipeline.yaml
+18
@@ -169,6 +169,24 @@ steps:
169
env:
170
DOCKER_BUILDKIT: "1"
171
172
+ - block: "Build ROCm release image"
173
+ key: block-rocm-release-image-build
174
+ depends_on: ~
175
+
176
+ - label: "Build release image (ROCm)"
177
+ depends_on: block-rocm-release-image-build
178
+ id: build-release-image-rocm
179
+ agents:
180
+ queue: cpu_queue_postmerge
181
+ commands:
182
+ - "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
183
+ # Build base image first
184
+ - "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --tag rocm/vllm-dev:base-$BUILDKITE_COMMIT --target final --progress plain -f docker/Dockerfile.rocm_base ."
185
+ # Build vLLM ROCm image using the base
186
+ - "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg BASE_IMAGE=rocm/vllm-dev:base-$BUILDKITE_COMMIT --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-rocm --target vllm-openai --progress plain -f docker/Dockerfile.rocm ."
187
+ - "docker push public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-rocm"
188
+
189
+
190
- label: "Build and publish nightly multi-arch image to DockerHub"
191
depends_on:
192
- create-multi-arch-manifest
.buildkite/scripts/annotate-release.sh
+7
@@ -32,6 +32,7 @@ To download and upload the image:
32
\`\`\`
33
docker pull public.ecr.aws/q9t5s3a7/vllm-release-repo:${BUILDKITE_COMMIT}-x86_64
34
docker pull public.ecr.aws/q9t5s3a7/vllm-release-repo:${BUILDKITE_COMMIT}-aarch64
35
+docker pull public.ecr.aws/q9t5s3a7/vllm-release-repo:${BUILDKITE_COMMIT}-rocm
36
37
docker tag public.ecr.aws/q9t5s3a7/vllm-release-repo:${BUILDKITE_COMMIT}-x86_64 vllm/vllm-openai:x86_64
38
docker tag vllm/vllm-openai:x86_64 vllm/vllm-openai:latest-x86_64
@@ -45,6 +46,12 @@ docker tag vllm/vllm-openai:aarch64 vllm/vllm-openai:v${RELEASE_VERSION}-aarch64
46
docker push vllm/vllm-openai:latest-aarch64
47
docker push vllm/vllm-openai:v${RELEASE_VERSION}-aarch64
48
49
+docker tag public.ecr.aws/q9t5s3a7/vllm-release-repo:${BUILDKITE_COMMIT}-rocm vllm/vllm-openai:rocm
50
+docker tag vllm/vllm-openai:rocm vllm/vllm-openai:latest-rocm
51
+docker tag vllm/vllm-openai:rocm vllm/vllm-openai:v${RELEASE_VERSION}-rocm
52
+docker push vllm/vllm-openai:latest-rocm
53
+docker push vllm/vllm-openai:v${RELEASE_VERSION}-rocm
54
+
55
docker manifest rm vllm/vllm-openai:latest
56
docker manifest create vllm/vllm-openai:latest vllm/vllm-openai:latest-x86_64 vllm/vllm-openai:latest-aarch64
57
docker manifest create vllm/vllm-openai:v${RELEASE_VERSION} vllm/vllm-openai:v${RELEASE_VERSION}-x86_64 vllm/vllm-openai:v${RELEASE_VERSION}-aarch64
docker/Dockerfile.rocm
+4
@@ -235,3 +235,7 @@ ENV KINETO_CONFIG="${COMMON_WORKDIR}/libkineto.conf"
235
RUN echo "VLLM_BASE_IMAGE=${BASE_IMAGE}" >> ${COMMON_WORKDIR}/versions.txt
236
237
CMD ["/bin/bash"]
238
+
239
+#Set entrypoint for vllm-openai official images
240
+FROM final As vllm-openai
241
+ENTRYPOINT ["vllm", "serve"]