| 1 | group: Entrypoints |
| 2 | depends_on: |
| 3 | - image-build |
| 4 | steps: |
| 5 | - label: ":nvidia: (H200) Initialized Snapshot E2E" |
| 6 | key: initialized-snapshot-e2e |
| 7 | device: h200 |
| 8 | num_devices: 1 |
| 9 | no_plugin: true |
| 10 | optional: true |
| 11 | timeout_in_minutes: 60 |
| 12 | working_dir: "." |
| 13 | env: |
| 14 | SNAPSHOT_E2E_RUN_LIMIT_S: "1200" |
| 15 | NVIDIA_VISIBLE_DEVICES: "0" |
| 16 | source_file_dependencies: |
| 17 | - .buildkite/scripts/initialized-snapshot-e2e.sh |
| 18 | - .buildkite/test_areas/entrypoints.yaml |
| 19 | - docker/Dockerfile |
| 20 | - tools/install_snapshot_runtime.sh |
| 21 | - vllm/entrypoints/cli/ |
| 22 | - vllm/snapshot/ |
| 23 | commands: |
| 24 | - bash .buildkite/scripts/initialized-snapshot-e2e.sh "$IMAGE_TAG" "$BUILDKITE_COMMIT" |
| 25 | |
| 26 | - label: ":nvidia: (H200 MIG 35GB) Entrypoints Unit" |
| 27 | device: h200_35gb |
| 28 | key: entrypoints-unit-tests |
| 29 | timeout_in_minutes: 40 |
| 30 | working_dir: "/vllm-workspace/tests" |
| 31 | source_file_dependencies: |
| 32 | - vllm/entrypoints |
| 33 | - tests/entrypoints/unit_tests |
| 34 | - tests/entrypoints/weight_transfer |
| 35 | - tests/entrypoints/launchers |
| 36 | commands: |
| 37 | - pytest -v -s entrypoints/unit_tests |
| 38 | - pytest -v -s entrypoints/weight_transfer |
| 39 | - pytest -v -s entrypoints/launchers |
| 40 | mirror: |
| 41 | amd: |
| 42 | label: ":amd: (MI355 DPX) Entrypoints Unit" |
| 43 | dind: false |
| 44 | device: mi355_dpx |
| 45 | timeout_in_minutes: 50 |
| 46 | depends_on: |
| 47 | - image-build-amd |
| 48 | source_file_dependencies: |
| 49 | - vllm/platforms/rocm.py |
| 50 | |
| 51 | - label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (LLM)" |
| 52 | device: h200_35gb |
| 53 | key: entrypoints-integration-llm |
| 54 | timeout_in_minutes: 60 |
| 55 | working_dir: "/vllm-workspace/tests" |
| 56 | source_file_dependencies: |
| 57 | - vllm/ |
| 58 | - "!vllm/distributed/kv_transfer/" |
| 59 | - tests/entrypoints/llm |
| 60 | commands: |
| 61 | - export VLLM_WORKER_MULTIPROC_METHOD=spawn |
| 62 | - pytest -v -s entrypoints/llm --ignore=entrypoints/llm/test_generate.py --ignore=entrypoints/llm/test_collective_rpc.py --ignore=entrypoints/llm/offline_mode |
| 63 | - pytest -v -s entrypoints/llm/test_generate.py # it needs a clean process |
| 64 | - pytest -v -s entrypoints/llm/offline_mode # Needs to avoid interference with other tests |
| 65 | mirror: |
| 66 | amd: |
| 67 | label: ":amd: (MI355 DPX) Entrypoints Integration (LLM)" |
| 68 | dind: false |
| 69 | device: mi355_dpx |
| 70 | timeout_in_minutes: 90 |
| 71 | depends_on: |
| 72 | - image-build-amd |
| 73 | |
| 74 | - label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (API Server) %N" |
| 75 | key: entrypoints-integration-api-server |
| 76 | device: h200_35gb |
| 77 | timeout_in_minutes: 75 |
| 78 | parallelism: 4 |
| 79 | working_dir: "/vllm-workspace/tests" |
| 80 | source_file_dependencies: |
| 81 | - vllm/ |
| 82 | - "!vllm/distributed/kv_transfer/" |
| 83 | - tests/entrypoints/serve |
| 84 | - tests/entrypoints/scale_out |
| 85 | commands: |
| 86 | - export VLLM_WORKER_MULTIPROC_METHOD=spawn |
| 87 | - pytest -v -s entrypoints/serve --ignore=entrypoints/serve/dev/rpc --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --shard-id=$$BUILDKITE_PARALLEL_JOB |
| 88 | - if [ "$$BUILDKITE_PARALLEL_JOB" = "1" ]; then PYTHONPATH=/vllm-workspace pytest -v -s entrypoints/serve/dev/rpc; fi |
| 89 | - pytest -v -s entrypoints/scale_out --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --shard-id=$$BUILDKITE_PARALLEL_JOB |
| 90 | mirror: |
| 91 | amd: |
| 92 | label: ":amd: (MI355 DPX) Entrypoints Integration (API Server) %N" |
| 93 | dind: false |
| 94 | device: mi355_dpx |
| 95 | timeout_in_minutes: 65 |
| 96 | depends_on: |
| 97 | - image-build-amd |
| 98 | |
| 99 | - label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (OpenAI API completion)" |
| 100 | device: h200_35gb |
| 101 | key: entrypoints-integration-api-server-openai-completion |
| 102 | timeout_in_minutes: 68 |
| 103 | working_dir: "/vllm-workspace/tests" |
| 104 | source_file_dependencies: |
| 105 | - vllm/ |
| 106 | - "!vllm/distributed/kv_transfer/" |
| 107 | - tests/entrypoints/openai |
| 108 | commands: |
| 109 | - export VLLM_WORKER_MULTIPROC_METHOD=spawn |
| 110 | - pytest -v -s entrypoints/openai/completion --ignore=entrypoints/openai/completion/test_tensorizer_entrypoint.py |
| 111 | - pytest -v -s entrypoints/openai --ignore=entrypoints/openai/completion --ignore=entrypoints/openai/chat_completion --ignore=entrypoints/openai/responses --ignore=entrypoints/openai/correctness |
| 112 | mirror: |
| 113 | amd: |
| 114 | label: ":amd: (MI355 DPX) Entrypoints Integration (OpenAI API completion)" |
| 115 | dind: false |
| 116 | device: mi355_dpx |
| 117 | timeout_in_minutes: 65 |
| 118 | depends_on: |
| 119 | - image-build-amd |
| 120 | |
| 121 | - label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (OpenAI API chat_completion)" |
| 122 | device: h200_35gb |
| 123 | key: entrypoints-integration-api-server-openai-chat_completion |
| 124 | timeout_in_minutes: 83 |
| 125 | working_dir: "/vllm-workspace/tests" |
| 126 | source_file_dependencies: |
| 127 | - vllm/ |
| 128 | - "!vllm/distributed/kv_transfer/" |
| 129 | - tests/entrypoints/openai |
| 130 | commands: |
| 131 | - export VLLM_WORKER_MULTIPROC_METHOD=spawn |
| 132 | - pytest -v -s entrypoints/openai/chat_completion |
| 133 | mirror: |
| 134 | amd: |
| 135 | label: ":amd: (MI355 DPX) Entrypoints Integration (OpenAI API chat_completion)" |
| 136 | dind: false |
| 137 | device: mi355_dpx |
| 138 | timeout_in_minutes: 70 |
| 139 | depends_on: |
| 140 | - image-build-amd |
| 141 | |
| 142 | - label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (API Server Generate)" |
| 143 | device: h200_35gb |
| 144 | key: entrypoints-integration-api-server-generate |
| 145 | timeout_in_minutes: 50 |
| 146 | working_dir: "/vllm-workspace/tests" |
| 147 | source_file_dependencies: |
| 148 | - vllm/ |
| 149 | - "!vllm/distributed/kv_transfer/" |
| 150 | - tests/tool_use |
| 151 | - tests/entrypoints/tool_parsers |
| 152 | - tests/entrypoints/generate |
| 153 | - tests/entrypoints/anthropic |
| 154 | - tests/entrypoints/cohere |
| 155 | commands: |
| 156 | - pytest -v -s tool_use |
| 157 | - pytest -v -s entrypoints/tool_parsers |
| 158 | - pytest -v -s entrypoints/generate |
| 159 | - pytest -v -s entrypoints/anthropic |
| 160 | - pytest -v -s entrypoints/cohere |
| 161 | mirror: |
| 162 | amd: |
| 163 | label: ":amd: (MI355 DPX) Entrypoints Integration (API Server Generate)" |
| 164 | dind: false |
| 165 | device: mi355_dpx |
| 166 | timeout_in_minutes: 65 |
| 167 | depends_on: |
| 168 | - image-build-amd |
| 169 | |
| 170 | - label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (Responses API)" |
| 171 | device: h200_35gb |
| 172 | key: entrypoints-integration-responses-api |
| 173 | timeout_in_minutes: 50 |
| 174 | working_dir: "/vllm-workspace/tests" |
| 175 | source_file_dependencies: |
| 176 | - vllm/ |
| 177 | - "!vllm/distributed/kv_transfer/" |
| 178 | - tests/entrypoints/openai/responses |
| 179 | commands: |
| 180 | - pytest -v -s entrypoints/openai/responses |
| 181 | mirror: |
| 182 | amd: |
| 183 | label: ":amd: (MI355 DPX) Entrypoints Integration (Responses API)" |
| 184 | dind: false |
| 185 | device: mi355_dpx |
| 186 | timeout_in_minutes: 50 |
| 187 | depends_on: |
| 188 | - image-build-amd |
| 189 | |
| 190 | - label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (Speech to Text)" |
| 191 | device: h200_35gb |
| 192 | key: entrypoints-integration-speech_to_text |
| 193 | timeout_in_minutes: 45 |
| 194 | working_dir: "/vllm-workspace/tests" |
| 195 | source_file_dependencies: |
| 196 | - vllm/ |
| 197 | - "!vllm/distributed/kv_transfer/" |
| 198 | - tests/entrypoints/speech_to_text |
| 199 | commands: |
| 200 | - export VLLM_WORKER_MULTIPROC_METHOD=spawn |
| 201 | - pytest -v -s entrypoints/speech_to_text |
| 202 | mirror: |
| 203 | amd: |
| 204 | label: ":amd: (MI355 DPX) Entrypoints Integration (Speech to Text)" |
| 205 | dind: false |
| 206 | device: mi355_dpx |
| 207 | timeout_in_minutes: 60 |
| 208 | working_dir: "/vllm-workspace/tests" |
| 209 | depends_on: |
| 210 | - image-build-amd |
| 211 | |
| 212 | - label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (Multimodal)" |
| 213 | device: h200_35gb |
| 214 | key: entrypoints-integration-multimodal |
| 215 | timeout_in_minutes: 45 |
| 216 | working_dir: "/vllm-workspace/tests" |
| 217 | source_file_dependencies: |
| 218 | - vllm/ |
| 219 | - "!vllm/distributed/kv_transfer/" |
| 220 | - tests/entrypoints/multimodal |
| 221 | commands: |
| 222 | - export VLLM_WORKER_MULTIPROC_METHOD=spawn |
| 223 | - pytest -v -s entrypoints/multimodal |
| 224 | mirror: |
| 225 | amd: |
| 226 | label: ":amd: (MI355 DPX) Entrypoints Integration (Multimodal)" |
| 227 | dind: false |
| 228 | device: mi355_dpx |
| 229 | timeout_in_minutes: 55 |
| 230 | depends_on: |
| 231 | - image-build-amd |
| 232 | |
| 233 | - label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (Pooling)" |
| 234 | device: h200_35gb |
| 235 | key: entrypoints-integration-pooling |
| 236 | timeout_in_minutes: 75 |
| 237 | working_dir: "/vllm-workspace/tests" |
| 238 | source_file_dependencies: |
| 239 | - vllm/ |
| 240 | - "!vllm/distributed/kv_transfer/" |
| 241 | - tests/entrypoints/pooling |
| 242 | commands: |
| 243 | - export VLLM_WORKER_MULTIPROC_METHOD=spawn |
| 244 | - pytest -v -s entrypoints/pooling |
| 245 | mirror: |
| 246 | amd: |
| 247 | label: ":amd: (MI355 DPX) Entrypoints Integration (Pooling)" |
| 248 | dind: false |
| 249 | device: mi355_dpx |
| 250 | timeout_in_minutes: 65 |
| 251 | working_dir: "/vllm-workspace/tests" |
| 252 | depends_on: |
| 253 | - image-build-amd |
| 254 | |
| 255 | - label: ":nvidia: (H200 MIG 18GB) OpenAI API Correctness" |
| 256 | key: openai-api-correctness |
| 257 | timeout_in_minutes: 20 |
| 258 | device: h200_18gb |
| 259 | source_file_dependencies: |
| 260 | - csrc/ |
| 261 | - vllm/entrypoints/openai/ |
| 262 | commands: # LMEval |
| 263 | - pytest -s entrypoints/openai/correctness/ |
| 264 | mirror: |
| 265 | amd: |
| 266 | label: ":amd: (MI355 DPX) OpenAI API Correctness" |
| 267 | dind: false |
| 268 | device: mi355_dpx |
| 269 | timeout_in_minutes: 30 |
| 270 | depends_on: |
| 271 | - image-build-amd |
| 272 | source_file_dependencies: |
| 273 | - csrc/ |
| 274 | - vllm/entrypoints/openai/ |
| 275 | - vllm/model_executor/layers/ |
| 276 | - vllm/v1/attention/backends/ |
| 277 | - vllm/v1/attention/selector.py |
| 278 | - vllm/_aiter_ops.py |
| 279 | - vllm/platforms/rocm.py |
| 280 | - vllm/model_executor/model_loader/ |
| 281 | commands: |
| 282 | - bash ../tools/install_torchcodec_rocm.sh || exit 1 |
| 283 | - pytest -s entrypoints/openai/correctness/ |