| 1 | group: Entrypoints |
| 2 | depends_on: |
| 3 | - image-build-xpu |
| 4 | steps: |
| 5 | - label: ":intel: (BMG) Entrypoints Unit" |
| 6 | key: entrypoints-unit-tests-intel |
| 7 | timeout_in_minutes: 45 |
| 8 | device: intel_gpu |
| 9 | agent_tags: |
| 10 | label: production |
| 11 | gpu: 1+ |
| 12 | mem: 16+ |
| 13 | no_plugin: true |
| 14 | working_dir: "." |
| 15 | env: |
| 16 | REGISTRY: "public.ecr.aws/q9t5s3a7" |
| 17 | REPO: "vllm-ci-test-repo" |
| 18 | VLLM_TEST_DEVICE: "xpu" |
| 19 | source_file_dependencies: |
| 20 | - vllm/entrypoints |
| 21 | - tests/entrypoints/unit_tests |
| 22 | - tests/entrypoints/weight_transfer |
| 23 | - tests/entrypoints/launchers |
| 24 | commands: |
| 25 | - >- |
| 26 | bash .buildkite/scripts/hardware_ci/run-intel-test.sh |
| 27 | 'cd tests && |
| 28 | pytest -v -s entrypoints/unit_tests && |
| 29 | pytest -v -s entrypoints/weight_transfer' |
| 30 | |
| 31 | - label: ":intel: (BMG) Entrypoints Integration (LLM)" |
| 32 | key: entrypoints-integration-llm-intel |
| 33 | timeout_in_minutes: 70 |
| 34 | device: intel_gpu |
| 35 | agent_tags: |
| 36 | label: production |
| 37 | gpu: 1+ |
| 38 | mem: 16+ |
| 39 | no_plugin: true |
| 40 | working_dir: "." |
| 41 | env: |
| 42 | REGISTRY: "public.ecr.aws/q9t5s3a7" |
| 43 | REPO: "vllm-ci-test-repo" |
| 44 | VLLM_TEST_DEVICE: "xpu" |
| 45 | source_file_dependencies: |
| 46 | - vllm/ |
| 47 | - tests/entrypoints/llm |
| 48 | commands: |
| 49 | - >- |
| 50 | bash .buildkite/scripts/hardware_ci/run-intel-test.sh |
| 51 | 'export VLLM_WORKER_MULTIPROC_METHOD=spawn && |
| 52 | cd tests && |
| 53 | pytest -v -s entrypoints/llm/test_generate.py && |
| 54 | pytest -v -s entrypoints/llm/offline_mode' |
| 55 | |
| 56 | - label: ":intel: (BMG) Entrypoints Integration (API Server)" |
| 57 | key: entrypoints-integration-api-server-intel |
| 58 | timeout_in_minutes: 70 |
| 59 | device: intel_gpu |
| 60 | agent_tags: |
| 61 | label: production |
| 62 | gpu: 1+ |
| 63 | mem: 16+ |
| 64 | no_plugin: true |
| 65 | working_dir: "." |
| 66 | env: |
| 67 | REGISTRY: "public.ecr.aws/q9t5s3a7" |
| 68 | REPO: "vllm-ci-test-repo" |
| 69 | VLLM_TEST_DEVICE: "xpu" |
| 70 | source_file_dependencies: |
| 71 | - vllm/ |
| 72 | - "!vllm/distributed/kv_transfer/" |
| 73 | - tests/entrypoints/serve |
| 74 | commands: |
| 75 | - >- |
| 76 | bash .buildkite/scripts/hardware_ci/run-intel-test.sh |
| 77 | 'export VLLM_WORKER_MULTIPROC_METHOD=spawn && |
| 78 | cd tests && |
| 79 | pytest -v -s entrypoints/serve --ignore=entrypoints/serve/dev/rpc |
| 80 | --ignore=entrypoints/serve/dev/rlhf/test_weight_operation_metrics.py |
| 81 | --deselect="tests/entrypoints/serve/dev/rlhf/state_transitions/test_pause_resume.py::TestPauseResume::test_clear_cache_preserves_output_and_controls_prefix_cache[MRV1]" && |
| 82 | PYTHONPATH=/workspace/vllm pytest -v -s entrypoints/serve/dev/rpc' |
| 83 | |
| 84 | - label: ":intel: (BMG) Entrypoints Integration (API Server Scale Out)" |
| 85 | key: entrypoints-integration-api-server-scale-out-intel |
| 86 | timeout_in_minutes: 40 |
| 87 | device: intel_gpu |
| 88 | agent_tags: |
| 89 | label: production |
| 90 | gpu: 1+ |
| 91 | mem: 16+ |
| 92 | no_plugin: true |
| 93 | working_dir: "." |
| 94 | env: |
| 95 | REGISTRY: "public.ecr.aws/q9t5s3a7" |
| 96 | REPO: "vllm-ci-test-repo" |
| 97 | VLLM_TEST_DEVICE: "xpu" |
| 98 | source_file_dependencies: |
| 99 | - vllm/ |
| 100 | - "!vllm/distributed/kv_transfer/" |
| 101 | - tests/entrypoints/scale_out |
| 102 | commands: |
| 103 | - >- |
| 104 | bash .buildkite/scripts/hardware_ci/run-intel-test.sh |
| 105 | 'export VLLM_WORKER_MULTIPROC_METHOD=spawn && |
| 106 | cd tests && |
| 107 | pytest -v -s entrypoints/scale_out' |
| 108 | |
| 109 | - label: ":intel: (BMG) Entrypoints Integration (API Server Generate)" |
| 110 | key: entrypoints-integration-api-server-generate-intel |
| 111 | timeout_in_minutes: 65 |
| 112 | device: intel_gpu |
| 113 | agent_tags: |
| 114 | label: production |
| 115 | gpu: 1+ |
| 116 | mem: 24+ |
| 117 | no_plugin: true |
| 118 | working_dir: "." |
| 119 | env: |
| 120 | REGISTRY: "public.ecr.aws/q9t5s3a7" |
| 121 | REPO: "vllm-ci-test-repo" |
| 122 | VLLM_TEST_DEVICE: "xpu" |
| 123 | optional: true |
| 124 | source_file_dependencies: |
| 125 | - vllm/ |
| 126 | - "!vllm/distributed/kv_transfer/" |
| 127 | - tests/tool_use |
| 128 | - tests/entrypoints/tool_parsers |
| 129 | - tests/entrypoints/generate |
| 130 | - tests/entrypoints/anthropic |
| 131 | - tests/entrypoints/cohere |
| 132 | commands: |
| 133 | - >- |
| 134 | bash .buildkite/scripts/hardware_ci/run-intel-test.sh |
| 135 | 'cd tests && |
| 136 | pytest -v -s tool_use |
| 137 | --deselect="tests/tool_use/mistral/test_mistral_tool_calls.py::test_tool_call_none_with_tools[ministral-3b]" |
| 138 | --deselect="tests/tool_use/test_chat_completions.py::test_chat_completion_with_tools[granite-3.0-8b]" |
| 139 | --deselect="tests/tool_use/test_chat_completions.py::test_chat_completion_with_tools[llama3.2]" && |
| 140 | pytest -v -s entrypoints/tool_parsers && |
| 141 | pytest -v -s entrypoints/generate && |
| 142 | pytest -v -s entrypoints/anthropic && |
| 143 | pytest -v -s entrypoints/cohere' |
| 144 | |
| 145 | - label: ":intel: (BMG) Entrypoints Integration (Responses API)" |
| 146 | key: entrypoints-integration-responses-api-intel |
| 147 | timeout_in_minutes: 60 |
| 148 | device: intel_gpu |
| 149 | agent_tags: |
| 150 | label: production |
| 151 | gpu: 1+ |
| 152 | mem: 24+ |
| 153 | no_plugin: true |
| 154 | working_dir: "." |
| 155 | env: |
| 156 | REGISTRY: "public.ecr.aws/q9t5s3a7" |
| 157 | REPO: "vllm-ci-test-repo" |
| 158 | VLLM_TEST_DEVICE: "xpu" |
| 159 | optional: true |
| 160 | source_file_dependencies: |
| 161 | - vllm/ |
| 162 | - "!vllm/distributed/kv_transfer/" |
| 163 | - tests/entrypoints/openai/responses |
| 164 | commands: |
| 165 | - >- |
| 166 | bash .buildkite/scripts/hardware_ci/run-intel-test.sh |
| 167 | 'cd tests && |
| 168 | pytest -v -s entrypoints/openai/responses' |
| 169 | |
| 170 | - label: ":intel: (BMG) OpenAI API Correctness" |
| 171 | key: openai-api-correctness-intel |
| 172 | timeout_in_minutes: 30 |
| 173 | device: intel_gpu |
| 174 | agent_tags: |
| 175 | label: production |
| 176 | gpu: 1+ |
| 177 | mem: 16+ |
| 178 | no_plugin: true |
| 179 | working_dir: "." |
| 180 | env: |
| 181 | REGISTRY: "public.ecr.aws/q9t5s3a7" |
| 182 | REPO: "vllm-ci-test-repo" |
| 183 | VLLM_TEST_DEVICE: "xpu" |
| 184 | source_file_dependencies: |
| 185 | - csrc/ |
| 186 | - vllm/entrypoints/openai/ |
| 187 | commands: |
| 188 | - >- |
| 189 | bash .buildkite/scripts/hardware_ci/run-intel-test.sh |
| 190 | 'cd tests && |
| 191 | pytest -v -s entrypoints/openai/correctness/' |