diff --git a/.buildkite/test_areas/distributed.yaml b/.buildkite/test_areas/distributed.yaml index 8825a67b1b..44e96ce632 100644 --- a/.buildkite/test_areas/distributed.yaml +++ b/.buildkite/test_areas/distributed.yaml @@ -36,7 +36,7 @@ steps: - label: ":nvidia: (L4) Distributed DP Basic" key: distributed-dp-tests-2-gpus - timeout_in_minutes: 35 + timeout_in_minutes: 45 working_dir: "/vllm-workspace/tests" device: l4 num_devices: 2 @@ -50,6 +50,9 @@ steps: - vllm/v1/worker/ - tests/v1/distributed - tests/entrypoints/launchers/api_server/test_multi_api_servers.py + - tests/v1/e2e/general/test_sharded_sampling.py + - tests/v1/e2e/spec_decode/test_sharded_sampling.py + - tests/v1/e2e/spec_decode/utils.py commands: # https://github.com/NVIDIA/nccl/issues/1838 - export NCCL_CUMEM_HOST_ENABLE=0 @@ -57,6 +60,9 @@ steps: - TP_SIZE=1 DP_SIZE=2 pytest -v -s v1/distributed/test_eagle_dp.py - TP_SIZE=1 DP_SIZE=2 pytest -v -s v1/distributed/test_external_lb_dp.py - DP_SIZE=2 pytest -v -s entrypoints/launchers/api_server/test_multi_api_servers.py + # Batch-sharded sampling requires TP=2 + - pytest -v -s v1/e2e/general/test_sharded_sampling.py + - pytest -v -s v1/e2e/spec_decode/test_sharded_sampling.py mirror: amd: label: ":amd: (MI300) Distributed DP Basic" @@ -324,9 +330,11 @@ steps: - vllm/executor/ - vllm/model_executor/models/ - tests/distributed/ + - tests/v1/distributed/test_pp_dp_v2.py commands: - pytest -v -s distributed/test_pp_cudagraph.py - pytest -v -s distributed/test_pipeline_parallel.py + - pytest -v -s v1/distributed/test_pp_dp_v2.py - label: ":nvidia: (L4) RayExecutorV2" key: rayexecutorv2-4-gpus diff --git a/.buildkite/test_areas/model_runner_v2.yaml b/.buildkite/test_areas/model_runner_v2.yaml deleted file mode 100644 index 60895f754a..0000000000 --- a/.buildkite/test_areas/model_runner_v2.yaml +++ /dev/null @@ -1,169 +0,0 @@ -group: Model Runner V2 -depends_on: - - image-build -steps: -- label: ":nvidia: (H200 MIG 35GB) Model Runner V2 Core" - device: h200_35gb - key: model-runner-v2-core-tests - timeout_in_minutes: 35 - source_file_dependencies: - - vllm/v1/worker/gpu/ - - vllm/v1/worker/gpu_worker.py - - vllm/v1/core/sched/ - - vllm/v1/attention/ - - tests/v1/engine/test_llm_engine.py - - tests/v1/e2e/ - - tests/entrypoints/llm/test_struct_output_generate.py - commands: - - set -x - - export VLLM_USE_V2_MODEL_RUNNER=1 - - pytest -v -s v1/engine/test_llm_engine.py -k "not test_engine_metrics" - - pytest -v -s v1/e2e/general/test_async_scheduling.py -k "not ngram" - - pytest -v -s v1/e2e/general/test_context_length.py - - pytest -v -s v1/e2e/general/test_min_tokens.py - # Temporary hack filter to exclude ngram spec decoding based tests. - - 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: ":nvidia: (H200 MIG 35GB) Model Runner V2 Examples" - device: h200_35gb - key: model-runner-v2-examples - timeout_in_minutes: 35 - working_dir: "/vllm-workspace/examples" - source_file_dependencies: - - vllm/v1/worker/gpu/ - - vllm/v1/core/sched/ - - vllm/v1/worker/gpu_worker.py - - examples/basic/offline_inference/ - - examples/generate/multimodal/ - - examples/features/ - - examples/pooling/embed/vision_embedding_offline.py - - examples/features/tensorize_vllm_model.py - commands: - - set -x - - export VLLM_USE_V2_MODEL_RUNNER=1 - - pip install --no-deps tensorizer # for tensorizer test - - python3 basic/offline_inference/chat.py # for basic - - python3 basic/offline_inference/generate.py --model facebook/opt-125m - #- python3 basic/offline_inference/generate.py --model meta-llama/Llama-2-13b-chat-hf --cpu-offload-gb 10 # TODO - #- python3 basic/offline_inference/embed.py # TODO - # for multi-modal models - - python3 generate/multimodal/audio_language_offline.py --seed 0 - - python3 generate/multimodal/vision_language_offline.py --seed 0 - - python3 generate/multimodal/vision_language_multi_image_offline.py --seed 0 - - python3 generate/multimodal/encoder_decoder_multimodal_offline.py --model-type whisper --seed 0 - # for pooling models - - python3 pooling/embed/vision_embedding_offline.py --seed 0 - # for features demo - - python3 features/automatic_prefix_caching/prefix_caching_offline.py - - python3 deployment/llm_engine_example.py - - python3 features/tensorize_vllm_model.py --model facebook/opt-125m serialize --serialized-directory /tmp/ --suffix v1 && python3 features/tensorize_vllm_model.py --model facebook/opt-125m deserialize --path-to-tensors /tmp/vllm/facebook/opt-125m/v1/model.tensors - - python3 features/speculative_decoding/spec_decode_offline.py --test --method eagle --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 2048 - # https://github.com/vllm-project/vllm/pull/26682 uses slightly more memory in PyTorch 2.9+ causing this test to OOM in 1xL4 GPU - - 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: ":nvidia: (L4) Model Runner V2 Distributed" - key: model-runner-v2-distributed-2-gpus - timeout_in_minutes: 40 - working_dir: "/vllm-workspace/tests" - device: l4 - num_devices: 2 - source_file_dependencies: - - vllm/v1/worker/gpu/ - - vllm/v1/worker/gpu_worker.py - - tests/basic_correctness/test_basic_correctness.py - - tests/v1/distributed/test_async_llm_dp.py - - tests/v1/distributed/test_eagle_dp.py - - tests/v1/e2e/general/test_sharded_sampling.py - - tests/v1/e2e/spec_decode/test_sharded_sampling.py - commands: - - set -x - - export VLLM_USE_V2_MODEL_RUNNER=1 - # The "and not True" here is a hacky way to exclude the prompt_embeds cases which aren't yet supported. - - TARGET_TEST_SUITE=L4 pytest -v -s basic_correctness/test_basic_correctness.py -m 'distributed(num_gpus=2)' -k "not ray and not True" - # https://github.com/NVIDIA/nccl/issues/1838 - - export NCCL_CUMEM_HOST_ENABLE=0 - - TP_SIZE=1 DP_SIZE=2 pytest -v -s v1/distributed/test_async_llm_dp.py -k "not ray" - - TP_SIZE=1 DP_SIZE=2 pytest -v -s v1/distributed/test_eagle_dp.py - - pytest -v -s v1/e2e/general/test_sharded_sampling.py - - pytest -v -s v1/e2e/spec_decode/test_sharded_sampling.py - - mirror: - amd: - label: ":amd: (MI300) Model Runner V2 Distributed" - dind: false - device: mi300_2 - timeout_in_minutes: 45 - depends_on: - - image-build-amd - source_file_dependencies: - - vllm/v1/worker/gpu/ - - vllm/v1/worker/gpu_worker.py - - tests/basic_correctness/test_basic_correctness.py - - tests/v1/distributed/test_async_llm_dp.py - - tests/v1/distributed/test_eagle_dp.py - - vllm/platforms/rocm.py - commands: - - set -x - - export VLLM_USE_V2_MODEL_RUNNER=1 - - TARGET_TEST_SUITE=MI300 pytest -v -s basic_correctness/test_basic_correctness.py -m 'distributed(num_gpus=2)' -k "not ray and not True" - - TP_SIZE=1 DP_SIZE=2 pytest -v -s v1/distributed/test_async_llm_dp.py -k "not ray" - - TP_SIZE=1 DP_SIZE=2 pytest -v -s v1/distributed/test_eagle_dp.py - -- label: ":nvidia: (L4) Model Runner V2 Pipeline Parallelism" - key: model-runner-v2-pipeline-parallelism-4-gpus - timeout_in_minutes: 50 - working_dir: "/vllm-workspace/tests" - device: l4 - num_devices: 4 - source_file_dependencies: - - vllm/v1/worker/gpu/ - - vllm/v1/worker/gpu_worker.py - - tests/distributed/test_pipeline_parallel.py - - tests/distributed/test_pp_cudagraph.py - - tests/v1/distributed/test_pp_dp_v2.py - commands: - - set -x - - export VLLM_USE_V2_MODEL_RUNNER=1 - - pytest -v -s distributed/test_pipeline_parallel.py -k "not ray" - - pytest -v -s distributed/test_pp_cudagraph.py -k "not ray" - - pytest -v -s v1/distributed/test_pp_dp_v2.py - -- label: ":nvidia: (H200 MIG 35GB) Model Runner V2 Spec Decode" - device: h200_35gb - key: model-runner-v2-spec-decode - timeout_in_minutes: 50 - working_dir: "/vllm-workspace/tests" - source_file_dependencies: - - vllm/v1/worker/gpu/ - - vllm/v1/worker/gpu_worker.py - - tests/v1/spec_decode/test_max_len.py - - tests/v1/spec_decode/test_rejection_sampler_utils.py - - tests/v1/spec_decode/test_synthetic_rejection_sampler_utils.py - - tests/v1/e2e/spec_decode/ - commands: - - set -x - - export VLLM_USE_V2_MODEL_RUNNER=1 - - pytest -v -s v1/spec_decode/test_max_len.py -k "eagle or mtp" - - pytest -v -s v1/spec_decode/test_rejection_sampler_utils.py - - pytest -v -s v1/spec_decode/test_synthetic_rejection_sampler_utils.py - - pytest -v -s v1/e2e/spec_decode/eagle/ - - pytest -v -s v1/e2e/spec_decode/speculators/ - - >- - pytest -v -s - 'v1/e2e/spec_decode/mtp/test_mtp.py::test_mtp_correctness[qwen3_5-hybrid]' - mirror: - amd: - label: ":amd: (MI355) Model Runner V2 Spec Decode" - dind: false - device: mi355_1 - timeout_in_minutes: 75 - depends_on: - - image-build-amd - source_file_dependencies: - - vllm/platforms/rocm.py - - vllm/v1/worker/gpu/ - - vllm/v1/worker/gpu_worker.py - - tests/v1/spec_decode/test_max_len.py - - tests/v1/spec_decode/test_rejection_sampler_utils.py - - tests/v1/spec_decode/test_synthetic_rejection_sampler_utils.py - - tests/v1/e2e/spec_decode/ diff --git a/tests/v1/distributed/test_pp_dp_v2.py b/tests/v1/distributed/test_pp_dp_v2.py index 3533154997..b931db1510 100644 --- a/tests/v1/distributed/test_pp_dp_v2.py +++ b/tests/v1/distributed/test_pp_dp_v2.py @@ -37,8 +37,8 @@ _GPU_SKIP = _gpu_skip_reason() pytestmark = [ pytest.mark.skipif( - os.environ.get("VLLM_USE_V2_MODEL_RUNNER", "0") != "1", - reason="VLLM_USE_V2_MODEL_RUNNER=1 required", + os.environ.get("VLLM_USE_V2_MODEL_RUNNER", "1") == "0", + reason="V2 model runner disabled via VLLM_USE_V2_MODEL_RUNNER=0", ), pytest.mark.skipif(_GPU_SKIP is not None, reason=_GPU_SKIP or ""), ]