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 6, 03:15 UTC
[release] Move agent queue to Release cluster queues (#37783)
Signed-off-by: khluu <khluu000@gmail.com>
Kevin H. Luu committed
Mar 23, 2026 at 20:36 UTC
7281199a8c36445caa6e10d57fd7650441986a59
1 file changed
+24
-24
.buildkite/release-pipeline.yaml
+24
-24
@@ -12,7 +12,7 @@ steps:
12
depends_on: ~
13
id: build-wheel-arm64-cuda-12-9
14
agents:
15
- queue: arm64_cpu_queue_postmerge
15
+ queue: arm64_cpu_queue_release
16
commands:
17
# #NOTE: torch_cuda_arch_list is derived from upstream PyTorch build files here:
18
# https://github.com/pytorch/pytorch/blob/main/.ci/aarch64_linux/aarch64_ci_build.sh#L7
@@ -27,7 +27,7 @@ steps:
27
depends_on: ~
28
id: build-wheel-arm64-cuda-13-0
29
agents:
30
- queue: arm64_cpu_queue_postmerge
30
+ queue: arm64_cpu_queue_release
31
commands:
32
# #NOTE: torch_cuda_arch_list is derived from upstream PyTorch build files here:
33
# https://github.com/pytorch/pytorch/blob/main/.ci/aarch64_linux/aarch64_ci_build.sh#L7
@@ -42,7 +42,7 @@ steps:
42
depends_on: ~
43
id: build-wheel-arm64-cpu
44
agents:
45
- queue: arm64_cpu_queue_postmerge
45
+ queue: arm64_cpu_queue_release
46
commands:
47
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg GIT_REPO_CHECK=1 --build-arg VLLM_BUILD_ACL=ON --tag vllm-ci:build-image --target vllm-build --progress plain -f docker/Dockerfile.cpu ."
48
- "mkdir artifacts"
@@ -55,7 +55,7 @@ steps:
55
depends_on: ~
56
id: build-wheel-x86-cuda-12-9
57
agents:
58
- queue: cpu_queue_postmerge
58
+ queue: cpu_queue_release
59
commands:
60
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.9.1 --tag vllm-ci:build-image --target build --progress plain -f docker/Dockerfile ."
61
- "mkdir artifacts"
@@ -68,7 +68,7 @@ steps:
68
depends_on: ~
69
id: build-wheel-x86-cuda-13-0
70
agents:
71
- queue: cpu_queue_postmerge
71
+ queue: cpu_queue_release
72
commands:
73
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=13.0.1 --build-arg BUILD_BASE_IMAGE=nvidia/cuda:13.0.1-devel-ubuntu22.04 --tag vllm-ci:build-image --target build --progress plain -f docker/Dockerfile ."
74
- "mkdir artifacts"
@@ -81,7 +81,7 @@ steps:
81
depends_on: ~
82
id: build-wheel-x86-cpu
83
agents:
84
- queue: cpu_queue_postmerge
84
+ queue: cpu_queue_release
85
commands:
86
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg GIT_REPO_CHECK=1 --build-arg VLLM_CPU_X86=true --tag vllm-ci:build-image --target vllm-build --progress plain -f docker/Dockerfile.cpu ."
87
- "mkdir artifacts"
@@ -97,7 +97,7 @@ steps:
97
depends_on: ~
98
id: build-release-image-x86
99
agents:
100
- queue: cpu_queue_postmerge
100
+ queue: cpu_queue_release
101
commands:
102
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
103
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.9.1 --build-arg FLASHINFER_AOT_COMPILE=true --build-arg INSTALL_KV_CONNECTORS=true --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m) --target vllm-openai --progress plain -f docker/Dockerfile ."
@@ -110,7 +110,7 @@ steps:
110
depends_on: ~
111
id: build-release-image-arm64
112
agents:
113
- queue: arm64_cpu_queue_postmerge
113
+ queue: arm64_cpu_queue_release
114
commands:
115
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
116
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.9.1 --build-arg FLASHINFER_AOT_COMPILE=true --build-arg torch_cuda_arch_list='8.7 8.9 9.0 10.0+PTX 12.0' --build-arg INSTALL_KV_CONNECTORS=true --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m) --target vllm-openai --progress plain -f docker/Dockerfile ."
@@ -120,7 +120,7 @@ steps:
120
depends_on: ~
121
id: build-release-image-x86-cuda-13-0
122
agents:
123
- queue: cpu_queue_postmerge
123
+ queue: cpu_queue_release
124
commands:
125
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
126
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=13.0.1 --build-arg INSTALL_KV_CONNECTORS=true --build-arg BUILD_BASE_IMAGE=nvidia/cuda:13.0.1-devel-ubuntu22.04 --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m)-cu130 --target vllm-openai --progress plain -f docker/Dockerfile ."
@@ -133,7 +133,7 @@ steps:
133
depends_on: ~
134
id: build-release-image-arm64-cuda-13-0
135
agents:
136
- queue: arm64_cpu_queue_postmerge
136
+ queue: arm64_cpu_queue_release
137
commands:
138
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
139
# compute capability 12.0 for RTX-50 series / RTX PRO 6000 Blackwell, 12.1 for DGX Spark
@@ -149,7 +149,7 @@ steps:
149
- block-cpu-release-image-build
150
- input-release-version
151
agents:
152
- queue: cpu_queue_postmerge
152
+ queue: cpu_queue_release
153
commands:
154
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
155
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg GIT_REPO_CHECK=1 --build-arg VLLM_CPU_X86=true --tag public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:$(buildkite-agent meta-data get release-version) --tag public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:latest --progress plain --target vllm-openai -f docker/Dockerfile.cpu ."
@@ -167,7 +167,7 @@ steps:
167
- block-arm64-cpu-release-image-build
168
- input-release-version
169
agents:
170
- queue: arm64_cpu_queue_postmerge
170
+ queue: arm64_cpu_queue_release
171
commands:
172
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
173
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg GIT_REPO_CHECK=1 --tag public.ecr.aws/q9t5s3a7/vllm-arm64-cpu-release-repo:$(buildkite-agent meta-data get release-version) --tag public.ecr.aws/q9t5s3a7/vllm-arm64-cpu-release-repo:latest --progress plain --target vllm-openai -f docker/Dockerfile.cpu ."
@@ -185,7 +185,7 @@ steps:
185
- build-release-image-arm64
186
id: create-multi-arch-manifest
187
agents:
188
- queue: small_cpu_queue_postmerge
188
+ queue: small_cpu_queue_release
189
commands:
190
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
191
- "docker manifest create public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-x86_64 public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-aarch64 --amend"
@@ -196,7 +196,7 @@ steps:
196
- create-multi-arch-manifest
197
id: annotate-release-workflow
198
agents:
199
- queue: small_cpu_queue_postmerge
199
+ queue: small_cpu_queue_release
200
commands:
201
- "bash .buildkite/scripts/annotate-release.sh"
202
@@ -206,7 +206,7 @@ steps:
206
- build-release-image-arm64-cuda-13-0
207
id: create-multi-arch-manifest-cuda-13-0
208
agents:
209
- queue: small_cpu_queue_postmerge
209
+ queue: small_cpu_queue_release
210
commands:
211
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
212
- "docker manifest create public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-cu130 public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-x86_64-cu130 public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-aarch64-cu130 --amend"
@@ -217,7 +217,7 @@ steps:
217
- create-multi-arch-manifest
218
if: build.env("NIGHTLY") == "1"
219
agents:
220
- queue: small_cpu_queue_postmerge
220
+ queue: small_cpu_queue_release
221
commands:
222
- "bash .buildkite/scripts/push-nightly-builds.sh"
223
# Clean up old nightly builds (keep only last 14)
@@ -235,7 +235,7 @@ steps:
235
- create-multi-arch-manifest-cuda-13-0
236
if: build.env("NIGHTLY") == "1"
237
agents:
238
- queue: small_cpu_queue_postmerge
238
+ queue: small_cpu_queue_release
239
commands:
240
- "bash .buildkite/scripts/push-nightly-builds.sh cu130"
241
# Clean up old nightly builds (keep only last 14)
@@ -262,7 +262,7 @@ steps:
262
- block-upload-release-wheels
263
id: upload-release-wheels
264
agents:
265
- queue: small_cpu_queue_postmerge
265
+ queue: small_cpu_queue_release
266
commands:
267
- "bash .buildkite/scripts/upload-release-wheels-pypi.sh"
268
@@ -323,7 +323,7 @@ steps:
323
- step: input-rocm-config
324
allow_failure: true # Allow failure so non-UI builds can proceed (input step is skipped)
325
agents:
326
- queue: cpu_queue_postmerge
326
+ queue: cpu_queue_release
327
commands:
328
# Set configuration and check cache
329
- |
@@ -465,7 +465,7 @@ steps:
465
- step: build-rocm-base-wheels
466
allow_failure: false
467
agents:
468
- queue: cpu_queue_postmerge
468
+ queue: cpu_queue_release
469
timeout_in_minutes: 180
470
commands:
471
# Download artifacts and prepare Docker image
@@ -575,7 +575,7 @@ steps:
575
- step: build-rocm-vllm-wheel
576
allow_failure: false
577
agents:
578
- queue: cpu_queue_postmerge
578
+ queue: cpu_queue_release
579
timeout_in_minutes: 60
580
commands:
581
# Download all wheel artifacts and run upload
@@ -624,7 +624,7 @@ steps:
624
- step: input-release-version
625
allow_failure: true
626
agents:
627
- queue: cpu_queue_postmerge
627
+ queue: cpu_queue_release
628
commands:
629
- "bash .buildkite/scripts/annotate-rocm-release.sh"
630
env:
@@ -641,7 +641,7 @@ steps:
641
depends_on: block-generate-root-index-rocm-wheels
642
id: generate-root-index-rocm-wheels
643
agents:
644
- queue: cpu_queue_postmerge
644
+ queue: cpu_queue_release
645
commands:
646
- "bash tools/vllm-rocm/generate-rocm-wheels-root-index.sh"
647
env:
@@ -655,7 +655,7 @@ steps:
655
- step: build-rocm-base-wheels
656
allow_failure: false
657
agents:
658
- queue: cpu_queue_postmerge
658
+ queue: cpu_queue_release
659
timeout_in_minutes: 60
660
commands:
661
- |