[CI] Add key field to all test_areas pipeline steps (#41201)

Signed-off-by: khluu <khluu000@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
This commit is contained in:
Kevin H. Luu
2026-04-29 16:59:16 -07:00
committed by GitHub
co-authored by Claude
parent 3795d7acf4
commit 0ab67c0222
28 changed files with 162 additions and 3 deletions
+2
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: V1 attention (H100)
key: v1-attention-h100
timeout_in_minutes: 30
device: h100
source_file_dependencies:
@@ -14,6 +15,7 @@ steps:
- pytest -v -s v1/attention
- label: V1 attention (B200)
key: v1-attention-b200
timeout_in_minutes: 30
device: b200
source_file_dependencies:
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Basic Correctness
key: basic-correctness
timeout_in_minutes: 30
device: h200_18gb
source_file_dependencies:
+2
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Benchmarks CLI Test
key: benchmarks-cli-test
timeout_in_minutes: 20
device: h200_18gb
source_file_dependencies:
@@ -12,6 +13,7 @@ steps:
- pytest -v -s benchmarks/
- label: Attention Benchmarks Smoke Test (B200)
key: attention-benchmarks-smoke-test-b200
device: b200
num_gpus: 2
optional: true
+13
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Sequence Parallel Correctness Tests (2 GPUs)
key: sequence-parallel-correctness-tests-2-gpus
timeout_in_minutes: 50
working_dir: "/vllm-workspace/"
num_devices: 2
@@ -17,6 +18,7 @@ steps:
- pytest -v -s tests/compile/correctness_e2e/test_sequence_parallel.py
- label: Sequence Parallel Correctness Tests (2xH100)
key: sequence-parallel-correctness-tests-2xh100
timeout_in_minutes: 50
working_dir: "/vllm-workspace/"
device: h100
@@ -27,6 +29,7 @@ steps:
- pytest -v -s tests/compile/correctness_e2e/test_sequence_parallel.py
- label: AsyncTP Correctness Tests (2xH100)
key: asynctp-correctness-tests-2xh100
timeout_in_minutes: 50
working_dir: "/vllm-workspace/"
device: h100
@@ -37,6 +40,7 @@ steps:
- pytest -v -s tests/compile/correctness_e2e/test_async_tp.py
- label: AsyncTP Correctness Tests (B200)
key: asynctp-correctness-tests-b200
timeout_in_minutes: 50
working_dir: "/vllm-workspace/"
device: b200
@@ -47,6 +51,7 @@ steps:
- pytest -v -s tests/compile/correctness_e2e/test_async_tp.py
- label: Distributed Compile Unit Tests (2xH100)
key: distributed-compile-unit-tests-2xh100
timeout_in_minutes: 20
working_dir: "/vllm-workspace/"
device: h100
@@ -60,6 +65,7 @@ steps:
- pytest -s -v tests/compile/passes/distributed
- label: Fusion and Compile Unit Tests (2xB200)
key: fusion-and-compile-unit-tests-2xb200
timeout_in_minutes: 20
working_dir: "/vllm-workspace/"
device: b200
@@ -89,6 +95,7 @@ steps:
- pytest -v -s tests/compile/fullgraph/test_full_graph.py::test_fp8_kv_scale_compile
- label: Fusion E2E Quick (H100)
key: fusion-e2e-quick-h100
timeout_in_minutes: 15
working_dir: "/vllm-workspace/"
device: h100
@@ -107,6 +114,7 @@ steps:
- pytest -v -s tests/compile/fusions_e2e/test_tp1_quant.py -k "inductor_partition and not +rms_norm and +quant_fp8 and (qwen3 or deepseek)"
- label: Fusion E2E Config Sweep (H100)
key: fusion-e2e-config-sweep-h100
timeout_in_minutes: 30
working_dir: "/vllm-workspace/"
device: h100
@@ -126,6 +134,7 @@ steps:
- pytest -v -s tests/compile/fusions_e2e/test_tp1_quant.py -k "llama-3"
- label: Fusion E2E Config Sweep (B200)
key: fusion-e2e-config-sweep-b200
timeout_in_minutes: 30
working_dir: "/vllm-workspace/"
device: b200
@@ -139,6 +148,7 @@ steps:
- pytest -v -s tests/compile/fusions_e2e/test_tp1_quant.py -k "inductor_partition and (FLASHINFER and not +rms_norm and (not +quant_fp8 or +quant_fp8 and (qwen3 or deepseek)) or llama-3)"
- label: Fusion E2E TP2 Quick (H100)
key: fusion-e2e-tp2-quick-h100
timeout_in_minutes: 20
working_dir: "/vllm-workspace/"
device: h100
@@ -156,6 +166,7 @@ steps:
- pytest -v -s tests/compile/fusions_e2e/test_tp2_async_tp.py -k "inductor_partition and not +rms_norm and (not +quant_fp8 or +quant_fp8 and (qwen3 or deepseek))"
- label: Fusion E2E TP2 AR-RMS Config Sweep (H100)
key: fusion-e2e-tp2-ar-rms-config-sweep-h100
timeout_in_minutes: 40
working_dir: "/vllm-workspace/"
device: h100
@@ -175,6 +186,7 @@ steps:
- pytest -v -s tests/compile/fusions_e2e/test_tp2_ar_rms.py -k "llama-3"
- label: Fusion E2E TP2 AsyncTP Config Sweep (H100)
key: fusion-e2e-tp2-asynctp-config-sweep-h100
timeout_in_minutes: 40
working_dir: "/vllm-workspace/"
device: h100
@@ -194,6 +206,7 @@ steps:
- pytest -v -s tests/compile/fusions_e2e/test_tp2_async_tp.py -k "llama-3"
- label: Fusion E2E TP2 (B200)
key: fusion-e2e-tp2-b200
timeout_in_minutes: 20
working_dir: "/vllm-workspace/"
device: b200
+2
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Platform Tests (CUDA)
key: platform-tests-cuda
timeout_in_minutes: 15
device: h200_18gb
source_file_dependencies:
@@ -13,6 +14,7 @@ steps:
- pytest -v -s cuda/test_platform_no_cuda_init.py
- label: Cudagraph
key: cudagraph
timeout_in_minutes: 20
source_file_dependencies:
- tests/v1/cudagraph
+8
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Distributed NixlConnector PD accuracy (4 GPUs)
key: distributed-nixlconnector-pd-accuracy-4-gpus
timeout_in_minutes: 30
working_dir: "/vllm-workspace/tests"
num_devices: 4
@@ -13,6 +14,7 @@ steps:
- uv pip install --system -r /vllm-workspace/requirements/kv_connectors.txt
- bash v1/kv_connector/nixl_integration/config_sweep_accuracy_test.sh
- label: Distributed FlashInfer NixlConnector PD accuracy (4 GPUs)
key: distributed-flashinfer-nixlconnector-pd-accuracy-4-gpus
timeout_in_minutes: 30
working_dir: "/vllm-workspace/tests"
num_devices: 4
@@ -24,6 +26,7 @@ steps:
- FLASHINFER=1 bash v1/kv_connector/nixl_integration/config_sweep_accuracy_test.sh
- label: DP EP Distributed NixlConnector PD accuracy tests (4 GPUs)
key: dp-ep-distributed-nixlconnector-pd-accuracy-tests-4-gpus
timeout_in_minutes: 30
working_dir: "/vllm-workspace/tests"
num_devices: 4
@@ -35,6 +38,7 @@ steps:
- DP_EP=1 bash v1/kv_connector/nixl_integration/config_sweep_accuracy_test.sh
- label: CrossLayer KV layout Distributed NixlConnector PD accuracy tests (4 GPUs)
key: crosslayer-kv-layout-distributed-nixlconnector-pd-accuracy-tests-4-gpus
timeout_in_minutes: 30
working_dir: "/vllm-workspace/tests"
num_devices: 4
@@ -46,6 +50,7 @@ steps:
- CROSS_LAYERS_BLOCKS=True bash v1/kv_connector/nixl_integration/config_sweep_accuracy_test.sh
- label: Hybrid SSM NixlConnector PD accuracy tests (4 GPUs)
key: hybrid-ssm-nixlconnector-pd-accuracy-tests-4-gpus
timeout_in_minutes: 20
working_dir: "/vllm-workspace/tests"
num_devices: 4
@@ -57,6 +62,7 @@ steps:
- HYBRID_SSM=1 bash v1/kv_connector/nixl_integration/config_sweep_accuracy_test.sh
- label: MultiConnector (Nixl+Offloading) PD accuracy (2 GPUs)
key: multiconnector-nixl-offloading-pd-accuracy-2-gpus
timeout_in_minutes: 30
working_dir: "/vllm-workspace/tests"
num_devices: 2
@@ -71,6 +77,7 @@ steps:
- bash v1/kv_connector/nixl_integration/run_multi_connector_accuracy_test.sh
- label: NixlConnector PD + Spec Decode acceptance (2 GPUs)
key: nixlconnector-pd-spec-decode-acceptance-2-gpus
timeout_in_minutes: 30
device: a100
working_dir: "/vllm-workspace/tests"
@@ -84,6 +91,7 @@ steps:
- bash v1/kv_connector/nixl_integration/spec_decode_acceptance_test.sh
- label: MultiConnector (Nixl+Offloading) PD edge cases (2 GPUs)
key: multiconnector-nixl-offloading-pd-edge-cases-2-gpus
timeout_in_minutes: 30
working_dir: "/vllm-workspace/tests"
num_devices: 2
+14
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Distributed Comm Ops
key: distributed-comm-ops
timeout_in_minutes: 20
working_dir: "/vllm-workspace/tests"
num_devices: 2
@@ -16,6 +17,7 @@ steps:
- pytest -v -s distributed/test_shm_storage.py
- label: Distributed DP Tests (2 GPUs)
key: distributed-dp-tests-2-gpus
timeout_in_minutes: 20
working_dir: "/vllm-workspace/tests"
num_devices: 2
@@ -37,6 +39,7 @@ steps:
- DP_SIZE=2 pytest -v -s entrypoints/openai/test_multi_api_servers.py
- label: Distributed Compile + RPC Tests (2 GPUs)
key: distributed-compile-rpc-tests-2-gpus
timeout_in_minutes: 20
working_dir: "/vllm-workspace/tests"
num_devices: 2
@@ -59,6 +62,7 @@ steps:
- pytest -v -s ./compile/test_wrapper.py
- label: Distributed Torchrun + Shutdown Tests (2 GPUs)
key: distributed-torchrun-shutdown-tests-2-gpus
timeout_in_minutes: 20
working_dir: "/vllm-workspace/tests"
num_devices: 2
@@ -81,6 +85,7 @@ steps:
- pytest -v -s v1/worker/test_worker_memory_snapshot.py
- label: Distributed Torchrun + Examples (4 GPUs)
key: distributed-torchrun-examples-4-gpus
timeout_in_minutes: 30
working_dir: "/vllm-workspace"
num_devices: 4
@@ -112,6 +117,7 @@ steps:
- VLLM_ALLOW_INSECURE_SERIALIZATION=1 python3 examples/rl/rlhf_ipc.py
- label: Distributed DP Tests (4 GPUs)
key: distributed-dp-tests-4-gpus
timeout_in_minutes: 30
working_dir: "/vllm-workspace/tests"
num_devices: 4
@@ -132,6 +138,7 @@ steps:
- pytest -v -s distributed/test_utils.py
- label: Distributed Compile + Comm (4 GPUs)
key: distributed-compile-comm-4-gpus
timeout_in_minutes: 30
working_dir: "/vllm-workspace/tests"
num_devices: 4
@@ -153,6 +160,7 @@ steps:
- pytest -v -s distributed/test_multiproc_executor.py::test_multiproc_executor_multi_node
- label: Distributed Tests (8 GPUs)(H100)
key: distributed-tests-8-gpus-h100
timeout_in_minutes: 10
device: h100
num_devices: 8
@@ -173,6 +181,7 @@ steps:
- torchrun --nproc-per-node=8 ../examples/features/torchrun/torchrun_dp_example_offline.py --tp-size=2 --pp-size=1 --dp-size=4 --enable-ep
- label: Distributed Tests (4 GPUs)(A100)
key: distributed-tests-4-gpus-a100
device: a100
optional: true
num_devices: 4
@@ -187,6 +196,7 @@ steps:
- pytest -v -s -x lora/test_mixtral.py
- label: Distributed Tests (2 GPUs)(H100)
key: distributed-tests-2-gpus-h100
timeout_in_minutes: 15
device: h100
optional: true
@@ -201,6 +211,7 @@ steps:
- pytest -v -s tests/distributed/test_packed_tensor.py
- label: Distributed Tests (2 GPUs)(B200)
key: distributed-tests-2-gpus-b200
device: b200
optional: true
working_dir: "/vllm-workspace/"
@@ -214,6 +225,7 @@ steps:
- label: 2 Node Test (4 GPUs)
key: 2-node-test-4-gpus
timeout_in_minutes: 30
working_dir: "/vllm-workspace/tests"
num_devices: 2
@@ -231,6 +243,7 @@ steps:
- ./.buildkite/scripts/run-multi-node-test.sh /vllm-workspace/tests 2 2 $IMAGE_TAG "VLLM_TEST_SAME_HOST=0 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_same_node.py | grep 'Same node test passed' && NUM_NODES=2 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_node_count.py | grep 'Node count test passed' && python3 ../examples/features/data_parallel/data_parallel_offline.py -dp=2 -tp=1 --dp-num-nodes=2 --dp-node-rank=0 --dp-master-addr=192.168.10.10 --dp-master-port=12345 --enforce-eager --trust-remote-code && VLLM_MULTI_NODE=1 pytest -v -s distributed/test_multi_node_assignment.py && VLLM_MULTI_NODE=1 pytest -v -s distributed/test_pipeline_parallel.py" "VLLM_TEST_SAME_HOST=0 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_same_node.py | grep 'Same node test passed' && NUM_NODES=2 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_node_count.py | grep 'Node count test passed' && python3 ../examples/features/data_parallel/data_parallel_offline.py -dp=2 -tp=1 --dp-num-nodes=2 --dp-node-rank=1 --dp-master-addr=192.168.10.10 --dp-master-port=12345 --enforce-eager --trust-remote-code"
- label: Pipeline + Context Parallelism (4 GPUs)
key: pipeline-context-parallelism-4-gpus
timeout_in_minutes: 60
working_dir: "/vllm-workspace/tests"
num_devices: 4
@@ -245,6 +258,7 @@ steps:
- pytest -v -s distributed/test_pipeline_parallel.py
- label: RayExecutorV2 (4 GPUs)
key: rayexecutorv2-4-gpus
timeout_in_minutes: 60
working_dir: "/vllm-workspace/tests"
num_devices: 4
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: DeepSeek V2-Lite Accuracy
key: deepseek-v2-lite-accuracy
timeout_in_minutes: 60
device: h100
optional: true
@@ -12,6 +13,7 @@ steps:
- bash .buildkite/scripts/scheduled_integration_test/deepseek_v2_lite_ep_eplb.sh 0.25 200 8010
- label: Qwen3-30B-A3B-FP8-block Accuracy
key: qwen3-30b-a3b-fp8-block-accuracy
timeout_in_minutes: 60
device: h100
optional: true
@@ -21,6 +23,7 @@ steps:
- bash .buildkite/scripts/scheduled_integration_test/qwen30b_a3b_fp8_block_ep_eplb.sh 0.8 200 8020
- label: Qwen3-30B-A3B-FP8-block Accuracy (B200)
key: qwen3-30b-a3b-fp8-block-accuracy-b200
timeout_in_minutes: 60
device: b200
optional: true
@@ -30,6 +33,7 @@ steps:
- bash .buildkite/scripts/scheduled_integration_test/qwen30b_a3b_fp8_block_ep_eplb.sh 0.8 200 8020 2 1
- label: Qwen3-30B-A3B-FP8 DP4 Async EPLB Accuracy
key: qwen3-30b-a3b-fp8-dp4-async-eplb-accuracy
timeout_in_minutes: 60
device: h100
optional: true
@@ -39,6 +43,7 @@ steps:
- bash .buildkite/scripts/scheduled_integration_test/qwen30b_a3b_fp8_dp4_async_eplb.sh 0.8 200 8050
- label: DeepSeek V2-Lite Prefetch Offload Accuracy (H100)
key: deepseek-v2-lite-prefetch-offload-accuracy-h100
timeout_in_minutes: 60
device: h100
optional: true
+7
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Engine
key: engine
timeout_in_minutes: 15
device: h200_18gb
source_file_dependencies:
@@ -16,6 +17,7 @@ steps:
- pytest -v -s engine test_sequence.py test_config.py test_logger.py test_vllm_port.py
- label: Engine (1 GPU)
key: engine-1-gpu
timeout_in_minutes: 30
source_file_dependencies:
- vllm/v1/engine/
@@ -25,6 +27,7 @@ steps:
- pytest -v -s v1/engine --ignore v1/engine/test_preprocess_error_handling.py
- label: e2e Scheduling (1 GPU)
key: e2e-scheduling-1-gpu
timeout_in_minutes: 30
device: h200_18gb
source_file_dependencies:
@@ -34,6 +37,7 @@ steps:
- pytest -v -s v1/e2e/general/test_async_scheduling.py
- label: e2e Core (1 GPU)
key: e2e-core-1-gpu
timeout_in_minutes: 30
source_file_dependencies:
- vllm/v1/
@@ -42,6 +46,7 @@ steps:
- pytest -v -s v1/e2e/general --ignore v1/e2e/general/test_async_scheduling.py
- label: V1 e2e (2 GPUs)
key: v1-e2e-2-gpus
timeout_in_minutes: 60 # TODO: Fix timeout after we have more confidence in the test stability
optional: true
num_devices: 2
@@ -58,6 +63,7 @@ steps:
- image-build-amd
- label: V1 e2e (4 GPUs)
key: v1-e2e-4-gpus
timeout_in_minutes: 60 # TODO: Fix timeout after we have more confidence in the test stability
optional: true
num_devices: 4
@@ -74,6 +80,7 @@ steps:
- image-build-amd
- label: V1 e2e (4xH100)
key: v1-e2e-4xh100
timeout_in_minutes: 60
device: h100
num_devices: 4
+10 -1
View File
@@ -2,7 +2,8 @@ group: Entrypoints
depends_on:
- image-build
steps:
- label: Entrypoints Unit Tests
- label: Entrypoints Unit Tests
key: entrypoints-unit-tests
timeout_in_minutes: 10
working_dir: "/vllm-workspace/tests"
source_file_dependencies:
@@ -13,6 +14,7 @@ steps:
- pytest -v -s entrypoints/ --ignore=entrypoints/llm --ignore=entrypoints/rpc --ignore=entrypoints/sleep --ignore=entrypoints/serve/instrumentator --ignore=entrypoints/openai --ignore=entrypoints/offline_mode --ignore=entrypoints/test_chat_utils.py --ignore=entrypoints/pooling
- label: Entrypoints Integration (LLM)
key: entrypoints-integration-llm
timeout_in_minutes: 40
working_dir: "/vllm-workspace/tests"
source_file_dependencies:
@@ -26,6 +28,7 @@ steps:
- pytest -v -s entrypoints/offline_mode # Needs to avoid interference with other tests
- label: Entrypoints Integration (API Server openai - Part 1)
key: entrypoints-integration-api-server-openai-part-1
timeout_in_minutes: 50
working_dir: "/vllm-workspace/tests"
source_file_dependencies:
@@ -43,6 +46,7 @@ steps:
- label: Entrypoints Integration (API Server openai - Part 2)
key: entrypoints-integration-api-server-openai-part-2
timeout_in_minutes: 50
working_dir: "/vllm-workspace/tests"
source_file_dependencies:
@@ -60,6 +64,7 @@ steps:
- image-build-amd
- label: Entrypoints Integration (API Server openai - Part 3)
key: entrypoints-integration-api-server-openai-part-3
timeout_in_minutes: 50
device: h200_18gb
working_dir: "/vllm-workspace/tests"
@@ -72,6 +77,7 @@ steps:
- pytest -v -s entrypoints/openai --ignore=entrypoints/openai/chat_completion --ignore=entrypoints/openai/completion --ignore=entrypoints/openai/speech_to_text/ --ignore=entrypoints/openai/correctness/ --ignore=entrypoints/openai/tool_parsers/ --ignore=entrypoints/openai/responses --ignore=entrypoints/openai/test_multi_api_servers.py
- label: Entrypoints Integration (API Server 2)
key: entrypoints-integration-api-server-2
timeout_in_minutes: 130
working_dir: "/vllm-workspace/tests"
source_file_dependencies:
@@ -86,6 +92,7 @@ steps:
- pytest -v -s tool_use
- label: Entrypoints Integration (Pooling)
key: entrypoints-integration-pooling
timeout_in_minutes: 50
working_dir: "/vllm-workspace/tests"
source_file_dependencies:
@@ -96,6 +103,7 @@ steps:
- pytest -v -s entrypoints/pooling
- label: Entrypoints Integration (Responses API)
key: entrypoints-integration-responses-api
timeout_in_minutes: 50
working_dir: "/vllm-workspace/tests"
source_file_dependencies:
@@ -105,6 +113,7 @@ steps:
- pytest -v -s entrypoints/openai/responses
- label: OpenAI API Correctness
key: openai-api-correctness
timeout_in_minutes: 30
device: h200_18gb
source_file_dependencies:
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: EPLB Algorithm
key: eplb-algorithm
timeout_in_minutes: 15
device: h200_18gb
working_dir: "/vllm-workspace/tests"
@@ -15,6 +16,7 @@ steps:
- pytest -v -s distributed/test_eplb_utils.py
- label: EPLB Execution # 17min
key: eplb-execution
timeout_in_minutes: 27
working_dir: "/vllm-workspace/tests"
num_devices: 4
@@ -26,6 +28,7 @@ steps:
- pytest -v -s distributed/test_eplb_spec_decode.py
- label: Elastic EP Scaling Test
key: elastic-ep-scaling-test
timeout_in_minutes: 20
device: h100
working_dir: "/vllm-workspace/tests"
+15
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: vLLM IR Tests
key: vllm-ir-tests
timeout_in_minutes: 10
device: h200_18gb
working_dir: "/vllm-workspace/"
@@ -14,6 +15,7 @@ steps:
- pytest -v -s tests/kernels/ir
- label: Kernels Core Operation Test
key: kernels-core-operation-test
timeout_in_minutes: 75
source_file_dependencies:
- csrc/
@@ -23,6 +25,7 @@ steps:
- pytest -v -s kernels/core --ignore=kernels/core/test_minimax_reduce_rms.py kernels/test_concat_mla_q.py
- label: Kernels MiniMax Reduce RMS Test (2 GPUs)
key: kernels-minimax-reduce-rms-test-2-gpus
timeout_in_minutes: 15
num_devices: 2
device: h100
@@ -36,6 +39,7 @@ steps:
- pytest -v -s kernels/core/test_minimax_reduce_rms.py
- label: Kernels Attention Test %N
key: kernels-attention-test
timeout_in_minutes: 35
source_file_dependencies:
- csrc/attention/
@@ -49,6 +53,7 @@ steps:
parallelism: 2
- label: Kernels Quantization Test %N
key: kernels-quantization-test
timeout_in_minutes: 90
source_file_dependencies:
- csrc/quantization/
@@ -59,6 +64,7 @@ steps:
parallelism: 2
- label: Kernels MoE Test %N
key: kernels-moe-test
timeout_in_minutes: 25
source_file_dependencies:
- csrc/quantization/cutlass_w8a8/moe/
@@ -74,6 +80,7 @@ steps:
parallelism: 5
- label: Kernels Mamba Test
key: kernels-mamba-test
timeout_in_minutes: 45
source_file_dependencies:
- csrc/mamba/
@@ -83,6 +90,7 @@ steps:
- pytest -v -s kernels/mamba
- label: Kernels DeepGEMM Test (H100)
key: kernels-deepgemm-test-h100
timeout_in_minutes: 45
device: h100
num_devices: 1
@@ -104,6 +112,7 @@ steps:
- pytest -v -s quantization/test_cutlass_w4a16.py
- label: Kernels (B200)
key: kernels-b200
timeout_in_minutes: 30
working_dir: "/vllm-workspace/"
device: b200
@@ -152,6 +161,7 @@ steps:
- pytest -v -s tests/models/quantization/test_nvfp4.py
- label: Kernels Helion Test
key: kernels-helion-test
timeout_in_minutes: 30
device: h100
source_file_dependencies:
@@ -163,6 +173,7 @@ steps:
- label: Kernels FP8 MoE Test (1 H100)
key: kernels-fp8-moe-test-1-h100
timeout_in_minutes: 90
device: h100
num_devices: 1
@@ -179,6 +190,7 @@ steps:
- pytest -v -s kernels/moe/test_triton_moe_ptpc_fp8.py
- label: Kernels FP8 MoE Test (2 H100s)
key: kernels-fp8-moe-test-2-h100s
timeout_in_minutes: 90
device: h100
num_devices: 2
@@ -188,6 +200,7 @@ steps:
- pytest -v -s kernels/moe/test_deepep_moe.py
- label: Kernels Fp4 MoE Test (B200)
key: kernels-fp4-moe-test-b200
timeout_in_minutes: 60
device: b200
num_devices: 1
@@ -200,6 +213,7 @@ steps:
- label: Kernels FusedMoE Layer Test (2 H100s)
key: kernels-fusedmoe-layer-test-2-h100s
timeout_in_minutes: 90
device: h100
num_devices: 2
@@ -216,6 +230,7 @@ steps:
- label: Kernels FusedMoE Layer Test (2 B200s)
key: kernels-fusedmoe-layer-test-2-b200s
timeout_in_minutes: 90
device: b200
num_devices: 2
+11
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: LM Eval Small Models
key: lm-eval-small-models
timeout_in_minutes: 75
source_file_dependencies:
- csrc/
@@ -24,6 +25,7 @@ steps:
# - pytest -s -v test_lm_eval_correctness.py --config-list-file=configs/models-large.txt --tp-size=4
- label: LM Eval Large Models (4 GPUs)(H100)
key: lm-eval-large-models-4-gpus-h100
device: h100
optional: true
num_devices: 4
@@ -36,6 +38,7 @@ steps:
- pytest -s -v test_lm_eval_correctness.py --config-list-file=configs/models-large-hopper.txt --tp-size=4
- label: LM Eval Small Models (B200)
key: lm-eval-small-models-b200
timeout_in_minutes: 120
device: b200
optional: true
@@ -46,6 +49,7 @@ steps:
- pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=configs/models-blackwell.txt
- label: LM Eval Qwen3.5 Models (B200)
key: lm-eval-qwen3-5-models-b200
timeout_in_minutes: 120
device: b200
optional: true
@@ -62,6 +66,7 @@ steps:
- pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=configs/models-qwen35-blackwell.txt
- label: LM Eval Large Models (H200)
key: lm-eval-large-models-h200
timeout_in_minutes: 60
device: h200
optional: true
@@ -70,6 +75,7 @@ steps:
- pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=configs/models-h200.txt
- label: MoE Refactor Integration Test (H100 - TEMPORARY)
key: moe-refactor-integration-test-h100-temporary
device: h100
optional: true
num_devices: 2
@@ -77,6 +83,7 @@ steps:
- pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/moe-refactor/config-h100.txt
- label: MoE Refactor Integration Test (B200 - TEMPORARY)
key: moe-refactor-integration-test-b200-temporary
device: b200
optional: true
num_devices: 2
@@ -84,6 +91,7 @@ steps:
- pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/moe-refactor/config-b200.txt
- label: MoE Refactor Integration Test (B200 DP - TEMPORARY)
key: moe-refactor-integration-test-b200-dp-temporary
device: b200
optional: true
num_devices: 2
@@ -92,6 +100,7 @@ steps:
- label: LM Eval TurboQuant KV Cache
key: lm-eval-turboquant-kv-cache
timeout_in_minutes: 75
source_file_dependencies:
- vllm/model_executor/layers/quantization/turboquant/
@@ -102,6 +111,7 @@ steps:
- pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/models-turboquant.txt
- label: GPQA Eval (GPT-OSS) (H100)
key: gpqa-eval-gpt-oss-h100
timeout_in_minutes: 120
device: h100
optional: true
@@ -115,6 +125,7 @@ steps:
- pytest -s -v evals/gpt_oss/test_gpqa_correctness.py --config-list-file=configs/models-h100.txt
- label: GPQA Eval (GPT-OSS) (B200)
key: gpqa-eval-gpt-oss-b200
timeout_in_minutes: 120
device: b200
optional: true
+2
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: LoRA %N
key: lora
timeout_in_minutes: 30
source_file_dependencies:
- vllm/lora
@@ -13,6 +14,7 @@ steps:
- label: LoRA TP (Distributed)
key: lora-tp-distributed
timeout_in_minutes: 30
num_devices: 4
source_file_dependencies:
+14 -1
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: V1 Spec Decode
key: v1-spec-decode
timeout_in_minutes: 30
source_file_dependencies:
- vllm/
@@ -18,6 +19,7 @@ steps:
- image-build-amd
- label: V1 Sample + Logits
key: v1-sample-logits
timeout_in_minutes: 30
device: h200_18gb
source_file_dependencies:
@@ -41,6 +43,7 @@ steps:
- image-build-amd
- label: V1 Core + KV + Metrics
key: v1-core-kv-metrics
timeout_in_minutes: 30
source_file_dependencies:
- vllm/
@@ -71,6 +74,7 @@ steps:
- image-build-amd
- label: V1 Others (CPU)
key: v1-others-cpu
depends_on:
- image-build-cpu
source_file_dependencies:
@@ -86,6 +90,7 @@ steps:
- pytest -v -s -m 'cpu_test' v1/metrics
- label: Regression
key: regression
timeout_in_minutes: 20
device: h200_18gb
source_file_dependencies:
@@ -97,6 +102,7 @@ steps:
working_dir: "/vllm-workspace/tests" # optional
- label: Examples
key: examples
timeout_in_minutes: 45
working_dir: "/vllm-workspace/examples"
source_file_dependencies:
@@ -128,6 +134,7 @@ steps:
- python3 features/speculative_decoding/spec_decode_offline.py --test --method eagle3 --num_spec_tokens 3 --dataset-name hf --dataset-path philschmid/mt-bench --num-prompts 80 --temp 0 --top-p 1.0 --top-k -1 --tp 1 --enable-chunked-prefill --max-model-len 1536
- label: Metrics, Tracing (2 GPUs)
key: metrics-tracing-2-gpus
timeout_in_minutes: 20
num_devices: 2
source_file_dependencies:
@@ -142,6 +149,7 @@ steps:
- pytest -v -s v1/tracing
- label: Python-only Installation
key: python-only-installation
depends_on: ~
timeout_in_minutes: 20
source_file_dependencies:
@@ -151,6 +159,7 @@ steps:
- bash standalone_tests/python_only_compile.sh
- label: Async Engine, Inputs, Utils, Worker
key: async-engine-inputs-utils-worker
timeout_in_minutes: 50
source_file_dependencies:
- vllm/
@@ -163,7 +172,8 @@ steps:
- pytest -v -s utils_
- label: Async Engine, Inputs, Utils, Worker, Config (CPU)
depends_on:
key: async-engine-inputs-utils-worker-config-cpu
depends_on:
- image-build-cpu
timeout_in_minutes: 30
source_file_dependencies:
@@ -196,6 +206,7 @@ steps:
- pytest -v -s config
- label: Batch Invariance (H100)
key: batch-invariance-h100
timeout_in_minutes: 30
device: h100
source_file_dependencies:
@@ -211,6 +222,7 @@ steps:
- VLLM_TEST_MODEL=Qwen/Qwen3-30B-A3B-Thinking-2507-FP8 pytest -v -s v1/determinism/test_batch_invariance.py::test_v1_generation_is_deterministic_across_batch_sizes_with_needle[FLASH_ATTN]
- label: Batch Invariance (B200)
key: batch-invariance-b200
timeout_in_minutes: 30
device: b200
source_file_dependencies:
@@ -227,6 +239,7 @@ steps:
- pytest -v -s v1/determinism/test_nvfp4_batch_invariant.py
- label: Acceptance Length Test (Large Models) # optional
key: acceptance-length-test-large-models
timeout_in_minutes: 25
gpu: h100
optional: true
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Model Executor
key: model-executor
timeout_in_minutes: 35
source_file_dependencies:
- vllm/engine/arg_utils.py
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Model Runner V2 Core Tests
key: model-runner-v2-core-tests
timeout_in_minutes: 45
source_file_dependencies:
- vllm/v1/worker/gpu/
@@ -25,6 +26,7 @@ steps:
- pytest -v -s entrypoints/llm/test_struct_output_generate.py -k "xgrammar and not speculative_config6 and not speculative_config7 and not speculative_config8 and not speculative_config0"
- label: Model Runner V2 Examples
key: model-runner-v2-examples
timeout_in_minutes: 45
working_dir: "/vllm-workspace/examples"
source_file_dependencies:
@@ -60,6 +62,7 @@ steps:
- python3 features/speculative_decoding/spec_decode_offline.py --test --method eagle3 --num_spec_tokens 3 --dataset-name hf --dataset-path philschmid/mt-bench --num-prompts 80 --temp 0 --top-p 1.0 --top-k -1 --tp 1 --enable-chunked-prefill --max-model-len 1536
- label: Model Runner V2 Distributed (2 GPUs)
key: model-runner-v2-distributed-2-gpus
timeout_in_minutes: 45
working_dir: "/vllm-workspace/tests"
num_devices: 2
@@ -80,6 +83,7 @@ steps:
- TP_SIZE=1 DP_SIZE=2 pytest -v -s v1/distributed/test_eagle_dp.py
- label: Model Runner V2 Pipeline Parallelism (4 GPUs)
key: model-runner-v2-pipeline-parallelism-4-gpus
timeout_in_minutes: 60
working_dir: "/vllm-workspace/tests"
num_devices: 4
@@ -95,6 +99,7 @@ steps:
- pytest -v -s distributed/test_pp_cudagraph.py -k "not ray"
- label: Model Runner V2 Spec Decode
key: model-runner-v2-spec-decode
timeout_in_minutes: 30
working_dir: "/vllm-workspace/tests"
source_file_dependencies:
+6
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Basic Models Tests (Initialization)
key: basic-models-tests-initialization
timeout_in_minutes: 45
torch_nightly: true
source_file_dependencies:
@@ -16,6 +17,7 @@ steps:
torch_nightly: {}
- label: Basic Models Tests (Extra Initialization) %N
key: basic-models-tests-extra-initialization
timeout_in_minutes: 45
source_file_dependencies:
- vllm/model_executor/models/
@@ -31,6 +33,7 @@ steps:
torch_nightly: {}
- label: Basic Models Tests (Other)
key: basic-models-tests-other
timeout_in_minutes: 45
source_file_dependencies:
- vllm/
@@ -47,6 +50,7 @@ steps:
- label: Basic Models Test (Other CPU) # 5min
key: basic-models-test-other-cpu
depends_on:
- image-build-cpu
timeout_in_minutes: 10
@@ -59,6 +63,7 @@ steps:
- pytest -v -s models/test_utils.py models/test_vision.py
- label: Transformers Nightly Models
key: transformers-nightly-models
working_dir: "/vllm-workspace/"
optional: true
soft_fail: true
@@ -74,6 +79,7 @@ steps:
- VLLM_WORKER_MULTIPROC_METHOD=spawn python3 examples/generate/multimodal/audio_language_offline.py --model-type whisper
- label: Transformers Backward Compatibility Models Test
key: transformers-backward-compatibility-models-test
working_dir: "/vllm-workspace/"
optional: true
soft_fail: true
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Distributed Model Tests (2 GPUs)
key: distributed-model-tests-2-gpus
timeout_in_minutes: 50
working_dir: "/vllm-workspace/tests"
num_devices: 2
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Language Models Tests (Standard)
key: language-models-tests-standard
timeout_in_minutes: 25
source_file_dependencies:
- vllm/
@@ -15,6 +16,7 @@ steps:
torch_nightly: {}
- label: Language Models Tests (Extra Standard) %N
key: language-models-tests-extra-standard
timeout_in_minutes: 45
source_file_dependencies:
- vllm/model_executor/models/
@@ -31,6 +33,7 @@ steps:
torch_nightly: {}
- label: Language Models Tests (Hybrid) %N
key: language-models-tests-hybrid
timeout_in_minutes: 75
source_file_dependencies:
- vllm/
@@ -47,6 +50,7 @@ steps:
torch_nightly: {}
- label: Language Models Test (Extended Generation) # 80min
key: language-models-test-extended-generation
timeout_in_minutes: 110
optional: true
source_file_dependencies:
@@ -69,6 +73,7 @@ steps:
- pytest -v -s models/language/generation -m '(not core_model) and (not hybrid_model)'
- label: Language Models Test (PPL)
key: language-models-test-ppl
timeout_in_minutes: 110
device: h200_18gb
optional: true
@@ -79,6 +84,7 @@ steps:
- pytest -v -s models/language/generation_ppl_test
- label: Language Models Test (Extended Pooling) # 36min
key: language-models-test-extended-pooling
timeout_in_minutes: 50
optional: true
source_file_dependencies:
@@ -93,6 +99,7 @@ steps:
- image-build-amd
- label: Language Models Test (MTEB)
key: language-models-test-mteb
timeout_in_minutes: 110
device: h200_18gb
optional: true
+12 -1
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: "Multi-Modal Models (Standard) 1: qwen2"
key: multi-modal-models-standard-1-qwen2
timeout_in_minutes: 45
device: h200_18gb
source_file_dependencies:
@@ -19,6 +20,7 @@ steps:
- image-build-amd
- label: "Multi-Modal Models (Standard) 2: qwen3 + gemma"
key: multi-modal-models-standard-2-qwen3-gemma
timeout_in_minutes: 45
device: h200_18gb
source_file_dependencies:
@@ -36,6 +38,7 @@ steps:
- image-build-amd
- label: "Multi-Modal Models (Standard) 3: llava + qwen2_vl"
key: multi-modal-models-standard-3-llava-qwen2-vl
timeout_in_minutes: 45
source_file_dependencies:
- vllm/
@@ -51,6 +54,7 @@ steps:
- image-build-amd
- label: "Multi-Modal Models (Standard) 4: other + whisper"
key: multi-modal-models-standard-4-other-whisper
timeout_in_minutes: 45
source_file_dependencies:
- vllm/
@@ -67,7 +71,8 @@ steps:
- image-build-amd
- label: Multi-Modal Processor (CPU)
depends_on:
key: multi-modal-processor-cpu
depends_on:
- image-build-cpu
timeout_in_minutes: 60
source_file_dependencies:
@@ -80,6 +85,7 @@ steps:
- pytest -v -s models/multimodal/processing --ignore models/multimodal/processing/test_tensor_schema.py
- label: Multi-Modal Processor # 44min
key: multi-modal-processor
timeout_in_minutes: 60
device: h200_18gb
source_file_dependencies:
@@ -91,6 +97,7 @@ steps:
- pytest -v -s models/multimodal/processing/test_tensor_schema.py
- label: Multi-Modal Accuracy Eval (Small Models) # 50min
key: multi-modal-accuracy-eval-small-models
timeout_in_minutes: 70
working_dir: "/vllm-workspace/.buildkite/lm-eval-harness"
source_file_dependencies:
@@ -101,6 +108,7 @@ steps:
- pytest -s -v test_lm_eval_correctness.py --config-list-file=configs/models-mm-small.txt --tp-size=1
- label: Multi-Modal Models (Extended Generation 1)
key: multi-modal-models-extended-generation-1
optional: true
source_file_dependencies:
- vllm/
@@ -117,6 +125,7 @@ steps:
- image-build-amd
- label: Multi-Modal Models (Extended Generation 2)
key: multi-modal-models-extended-generation-2
optional: true
source_file_dependencies:
- vllm/
@@ -126,6 +135,7 @@ steps:
- pytest -v -s models/multimodal/generation/test_common.py -m 'split(group=0) and not core_model'
- label: Multi-Modal Models (Extended Generation 3)
key: multi-modal-models-extended-generation-3
optional: true
source_file_dependencies:
- vllm/
@@ -135,6 +145,7 @@ steps:
- pytest -v -s models/multimodal/generation/test_common.py -m 'split(group=1) and not core_model'
- label: Multi-Modal Models (Extended Pooling)
key: multi-modal-models-extended-pooling
optional: true
device: h200_18gb
source_file_dependencies:
+1
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Plugin Tests (2 GPUs)
key: plugin-tests-2-gpus
timeout_in_minutes: 60
working_dir: "/vllm-workspace/tests"
num_devices: 2
+6
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: PyTorch Compilation Unit Tests
key: pytorch-compilation-unit-tests
timeout_in_minutes: 10
source_file_dependencies:
- vllm/
@@ -18,6 +19,7 @@ steps:
- "find compile/ -maxdepth 1 -name 'test_*.py' -print0 | xargs -0 -n1 -I{} pytest -s -v '{}'"
- label: PyTorch Compilation Unit Tests (H100)
key: pytorch-compilation-unit-tests-h100
timeout_in_minutes: 30
device: h100
num_devices: 1
@@ -28,6 +30,7 @@ steps:
- "find compile/h100/ -name 'test_*.py' -print0 | xargs -0 -n1 -I{} pytest -s -v '{}'"
- label: PyTorch Compilation Passes Unit Tests
key: pytorch-compilation-passes-unit-tests
timeout_in_minutes: 20
source_file_dependencies:
- vllm/
@@ -36,6 +39,7 @@ steps:
- pytest -s -v compile/passes --ignore compile/passes/distributed
- label: PyTorch Fullgraph Smoke Test
key: pytorch-fullgraph-smoke-test
timeout_in_minutes: 35
source_file_dependencies:
- vllm/
@@ -48,6 +52,7 @@ steps:
- "find compile/fullgraph/ -name 'test_*.py' -not -name 'test_full_graph.py' -print0 | xargs -0 -n1 -I{} pytest -s -v '{}'"
- label: PyTorch Fullgraph
key: pytorch-fullgraph
timeout_in_minutes: 30
device: h200_18gb
source_file_dependencies:
@@ -58,6 +63,7 @@ steps:
- pytest -v -s compile/fullgraph/test_full_graph.py -k 'not test_fp8_kv_scale_compile'
- label: Pytorch Nightly Dependency Override Check # 2min
key: pytorch-nightly-dependency-override-check
# if this test fails, it means the nightly torch version is not compatible with some
# of the dependencies. Please check the error message and add the package to whitelist
# in /vllm/tools/pre_commit/generate_nightly_torch_test.py
+3
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Quantization
key: quantization
timeout_in_minutes: 90
source_file_dependencies:
- csrc/
@@ -21,6 +22,7 @@ steps:
- VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization/ --ignore quantization/test_blackwell_moe.py
- label: Quantized MoE Test (B200)
key: quantized-moe-test-b200
timeout_in_minutes: 60
working_dir: "/vllm-workspace/"
device: b200
@@ -38,6 +40,7 @@ steps:
- pytest -s -v tests/quantization/test_blackwell_moe.py
- label: Quantized Models Test
key: quantized-models-test
timeout_in_minutes: 60
source_file_dependencies:
- vllm/model_executor/layers/quantization
+1
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Ray Dependency Compatibility Check
key: ray-dependency-compatibility-check
# Informational only — does not block the pipeline.
# If this fails, it means the PR introduces a dependency that
# conflicts with Ray's dependency constraints.
+1
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Samplers Test
key: samplers-test
timeout_in_minutes: 75
source_file_dependencies:
- vllm/model_executor/layers
+8
View File
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Spec Decode Eagle
key: spec-decode-eagle
timeout_in_minutes: 30
device: h200_18gb
source_file_dependencies:
@@ -13,6 +14,7 @@ steps:
- pytest -v -s v1/e2e/spec_decode -k "eagle_correctness"
- label: Spec Decode Eagle Nightly B200
key: spec-decode-eagle-nightly-b200
timeout_in_minutes: 30
device: b200
optional: true
@@ -24,6 +26,7 @@ steps:
- pytest -v -s v1/e2e/spec_decode -k "eagle_correctness"
- label: Spec Decode Speculators + MTP
key: spec-decode-speculators-mtp
timeout_in_minutes: 30
device: h200_18gb
source_file_dependencies:
@@ -35,6 +38,7 @@ steps:
- pytest -v -s v1/e2e/spec_decode -k "speculators or mtp_correctness"
- label: Spec Decode Speculators + MTP Nightly B200
key: spec-decode-speculators-mtp-nightly-b200
timeout_in_minutes: 30
device: b200
optional: true
@@ -47,6 +51,7 @@ steps:
- pytest -v -s v1/e2e/spec_decode -k "speculators or mtp_correctness"
- label: Spec Decode Ngram + Suffix
key: spec-decode-ngram-suffix
timeout_in_minutes: 30
device: h200_18gb
source_file_dependencies:
@@ -57,6 +62,7 @@ steps:
- pytest -v -s v1/e2e/spec_decode -k "ngram or suffix"
- label: Spec Decode Draft Model
key: spec-decode-draft-model
timeout_in_minutes: 30
device: h200_18gb
source_file_dependencies:
@@ -67,6 +73,7 @@ steps:
- pytest -v -s v1/e2e/spec_decode -k "draft_model or no_sync or batch_inference"
- label: Spec Decode Draft Model Nightly B200
key: spec-decode-draft-model-nightly-b200
timeout_in_minutes: 30
device: b200
optional: true
@@ -78,6 +85,7 @@ steps:
- pytest -v -s v1/e2e/spec_decode -k "draft_model or no_sync or batch_inference"
- label: DFlash Speculators Correctness
key: dflash-speculators-correctness
timeout_in_minutes: 30
device: h100
optional: true
@@ -3,6 +3,7 @@ depends_on:
- image-build
steps:
- label: Weight Loading Multiple GPU # 33min
key: weight-loading-multiple-gpu
timeout_in_minutes: 45
working_dir: "/vllm-workspace/tests"
num_devices: 2