diff --git a/buildkite/test-template-ci.j2 b/buildkite/test-template-ci.j2 index 8547ddde..7cae4e16 100644 --- a/buildkite/test-template-ci.j2 +++ b/buildkite/test-template-ci.j2 @@ -1,11 +1,13 @@ {% set docker_image = "public.ecr.aws/q9t5s3a7/vllm-ci-test-repo:$BUILDKITE_COMMIT" %} {% set docker_image_torch_nightly = "public.ecr.aws/q9t5s3a7/vllm-ci-test-repo:$BUILDKITE_COMMIT-torch-nightly" %} {% set docker_image_cu118 = "public.ecr.aws/q9t5s3a7/vllm-ci-test-repo:$BUILDKITE_COMMIT-cu118" %} +{% set docker_image_cpu = "public.ecr.aws/q9t5s3a7/vllm-ci-test-repo:$BUILDKITE_COMMIT-cpu" %} {% if branch == "main" %} {% set docker_image = "public.ecr.aws/q9t5s3a7/vllm-ci-postmerge-repo:$BUILDKITE_COMMIT" %} {% set docker_image_latest = "public.ecr.aws/q9t5s3a7/vllm-ci-postmerge-repo:latest" %} {% set docker_image_torch_nightly = "public.ecr.aws/q9t5s3a7/vllm-ci-postmerge-repo:$BUILDKITE_COMMIT-torch-nightly" %} {% set docker_image_cu118 = "public.ecr.aws/q9t5s3a7/vllm-ci-postmerge-repo:$BUILDKITE_COMMIT-cu118" %} +{% set docker_image_cpu = "public.ecr.aws/q9t5s3a7/vllm-ci-postmerge-repo:$BUILDKITE_COMMIT-cpu" %} {% endif %} {% set docker_image_amd = "rocm/vllm-ci:$BUILDKITE_COMMIT" %} {% set default_working_dir = "/vllm-workspace/tests" %} @@ -182,6 +184,36 @@ steps: - exit_status: -10 # Agent was lost limit: 2 + - label: ":docker: build image CPU" + key: image-build-cpu + depends_on: ~ + agents: + {% if branch == "main" %} + queue: cpu_queue_postmerge + {% else %} + queue: cpu_queue_premerge + {% endif %} + commands: + - "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7" + - | + #!/bin/bash + if [[ -z $(docker manifest inspect {{ docker_image_cpu }}) ]]; then + echo "Image not found, proceeding with build..." + else + echo "Image found" + exit 0 + fi + - "docker build --file docker/Dockerfile.cpu --build-arg max_jobs=16 --build-arg buildkite_commit=$BUILDKITE_COMMIT --tag {{ docker_image_cpu }} --target vllm-test --progress plain ." + - "docker push {{ docker_image_cpu }}" + env: + DOCKER_BUILDKIT: "1" + retry: + automatic: + - exit_status: -1 # Agent was lost + limit: 2 + - exit_status: -10 # Agent was lost + limit: 2 + {% for step in steps %} {% if step.fast_check_only != true %} @@ -285,7 +317,7 @@ steps: {% else %} {% set ns.blocked = 0 %} {% endif %} - + {% if ns.blocked == 1 or (step.optional and nightly != "1") %} - block: "Run Torch Nightly {{ step.label }}" depends_on: image-build-torch-nightly