From d35eb6c44071ea806018841c490f0d2f3219c485 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Nicol=C3=B2=20Lucchesi?= Date: Fri, 7 Aug 2026 01:51:10 +0200 Subject: [PATCH] [CI] Exclude KV-connector subtree from broad source dependencies (#51046) Signed-off-by: NickLucche --- .buildkite/test_areas/basic_correctness.yaml | 1 + .buildkite/test_areas/benchmarks.yaml | 1 + .buildkite/test_areas/distributed.yaml | 12 ++++++++++++ .buildkite/test_areas/engine.yaml | 1 + .buildkite/test_areas/entrypoints.yaml | 9 +++++++++ .buildkite/test_areas/expert_parallelism.yaml | 1 + .buildkite/test_areas/misc.yaml | 6 ++++++ .buildkite/test_areas/models_basic.yaml | 3 +++ .buildkite/test_areas/models_language.yaml | 7 +++++++ .buildkite/test_areas/models_multimodal.yaml | 11 +++++++++++ .buildkite/test_areas/pytorch.yaml | 5 +++++ .buildkite/test_areas/rust_frontend.yaml | 1 + .buildkite/test_areas/weight_loading.yaml | 1 + 13 files changed, 59 insertions(+) diff --git a/.buildkite/test_areas/basic_correctness.yaml b/.buildkite/test_areas/basic_correctness.yaml index 6f0c22cc8b..d0dac3a112 100644 --- a/.buildkite/test_areas/basic_correctness.yaml +++ b/.buildkite/test_areas/basic_correctness.yaml @@ -8,6 +8,7 @@ steps: device: h200_18gb source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/basic_correctness/test_basic_correctness - tests/basic_correctness/test_cpu_offload - tests/basic_correctness/test_mem.py diff --git a/.buildkite/test_areas/benchmarks.yaml b/.buildkite/test_areas/benchmarks.yaml index 034fb58840..ccf3aa9cb8 100644 --- a/.buildkite/test_areas/benchmarks.yaml +++ b/.buildkite/test_areas/benchmarks.yaml @@ -8,6 +8,7 @@ steps: device: h200_18gb source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/benchmarks/ commands: - pytest -v -s benchmarks/ diff --git a/.buildkite/test_areas/distributed.yaml b/.buildkite/test_areas/distributed.yaml index 3f3c2c06b8..96cf519fdd 100644 --- a/.buildkite/test_areas/distributed.yaml +++ b/.buildkite/test_areas/distributed.yaml @@ -9,6 +9,7 @@ steps: num_devices: 2 source_file_dependencies: - vllm/distributed + - "!vllm/distributed/kv_transfer/" - tests/distributed commands: - pytest -v -s distributed/test_comm_ops.py @@ -23,6 +24,7 @@ steps: num_devices: 2 source_file_dependencies: - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/executor/ - vllm/worker/worker_base.py @@ -46,6 +48,7 @@ steps: - image-build-amd source_file_dependencies: - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/executor/ - vllm/worker/worker_base.py @@ -63,6 +66,7 @@ steps: source_file_dependencies: - vllm/compilation/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/executor/ - vllm/worker/worker_base.py @@ -83,6 +87,7 @@ steps: num_devices: 2 source_file_dependencies: - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/executor/ - vllm/worker/worker_base.py @@ -106,6 +111,7 @@ steps: num_devices: 4 source_file_dependencies: - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - tests/distributed/test_torchrun_example.py - tests/distributed/test_torchrun_example_moe.py - examples/rl/ @@ -138,6 +144,7 @@ steps: num_devices: 4 source_file_dependencies: - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - tests/v1/distributed - tests/v1/engine/test_engine_core_client.py - tests/distributed/test_utils @@ -159,6 +166,7 @@ steps: num_devices: 4 source_file_dependencies: - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - tests/distributed/test_pynccl - tests/distributed/test_events - tests/distributed/test_symm_mem_allreduce.py @@ -182,6 +190,7 @@ steps: - examples/features/torchrun/torchrun_dp_example_offline.py - vllm/config/parallel.py - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/v1/engine/llm_engine.py - vllm/v1/executor/uniproc_executor.py - vllm/v1/worker/gpu_worker.py @@ -200,6 +209,7 @@ steps: num_devices: 4 source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" commands: # NOTE: don't test llama model here, it seems hf implementation is buggy # see https://github.com/vllm-project/vllm/pull/5689 for details @@ -247,6 +257,7 @@ steps: optional: true # TODO: revert once infra issue solved source_file_dependencies: - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/executor/ - vllm/model_executor/models/ @@ -262,6 +273,7 @@ steps: num_devices: 4 source_file_dependencies: - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/executor/ - vllm/model_executor/models/ diff --git a/.buildkite/test_areas/engine.yaml b/.buildkite/test_areas/engine.yaml index cb311ab6a1..4cb661ac16 100644 --- a/.buildkite/test_areas/engine.yaml +++ b/.buildkite/test_areas/engine.yaml @@ -98,6 +98,7 @@ steps: - vllm/compilation/ - vllm/config/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/envs.py - vllm/forward_context.py diff --git a/.buildkite/test_areas/entrypoints.yaml b/.buildkite/test_areas/entrypoints.yaml index 47c0db109c..d36f98bd67 100644 --- a/.buildkite/test_areas/entrypoints.yaml +++ b/.buildkite/test_areas/entrypoints.yaml @@ -22,6 +22,7 @@ steps: working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/entrypoints/llm commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn @@ -43,6 +44,7 @@ steps: working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/entrypoints/serve - tests/entrypoints/scale_out commands: @@ -65,6 +67,7 @@ steps: working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/entrypoints/openai - tests/entrypoints/test_chat_utils commands: @@ -85,6 +88,7 @@ steps: working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/entrypoints/openai - tests/entrypoints/test_chat_utils commands: @@ -106,6 +110,7 @@ steps: working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/tool_use - tests/entrypoints/tool_parsers - tests/entrypoints/anthropic @@ -130,6 +135,7 @@ steps: working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/entrypoints/openai/responses commands: - pytest -v -s entrypoints/openai/responses @@ -141,6 +147,7 @@ steps: working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/entrypoints/speech_to_text commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn @@ -153,6 +160,7 @@ steps: working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/entrypoints/multimodal commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn @@ -165,6 +173,7 @@ steps: working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/entrypoints/pooling commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn diff --git a/.buildkite/test_areas/expert_parallelism.yaml b/.buildkite/test_areas/expert_parallelism.yaml index 1d1609d46b..89c9fddf64 100644 --- a/.buildkite/test_areas/expert_parallelism.yaml +++ b/.buildkite/test_areas/expert_parallelism.yaml @@ -47,6 +47,7 @@ steps: num_devices: 4 source_file_dependencies: - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/executor/ - vllm/compilation/ diff --git a/.buildkite/test_areas/misc.yaml b/.buildkite/test_areas/misc.yaml index cb5bc921aa..8c80fead29 100644 --- a/.buildkite/test_areas/misc.yaml +++ b/.buildkite/test_areas/misc.yaml @@ -9,6 +9,7 @@ steps: source_file_dependencies: - vllm/config/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/inputs/ - vllm/logger.py @@ -45,6 +46,7 @@ steps: source_file_dependencies: - vllm/config/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/entrypoints/pooling/ - vllm/inputs/ @@ -101,6 +103,7 @@ steps: source_file_dependencies: - vllm/config/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/inputs/ - vllm/lora/ @@ -161,6 +164,7 @@ steps: source_file_dependencies: - vllm/config/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/inputs/ - vllm/model_executor/ @@ -228,6 +232,7 @@ steps: source_file_dependencies: - vllm/config/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/inputs/ - vllm/model_executor/ @@ -287,6 +292,7 @@ steps: - vllm/assets/ - vllm/config/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/inputs/ - vllm/model_executor/ diff --git a/.buildkite/test_areas/models_basic.yaml b/.buildkite/test_areas/models_basic.yaml index e231e6d90f..b3d922a5e9 100644 --- a/.buildkite/test_areas/models_basic.yaml +++ b/.buildkite/test_areas/models_basic.yaml @@ -8,6 +8,7 @@ steps: device: h200_18gb source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/test_initialization.py - tests/models/registry.py commands: @@ -35,6 +36,7 @@ steps: timeout_in_minutes: 35 source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/test_terratorch.py - tests/models/transformers/test_backend.py - tests/models/test_registry.py @@ -82,6 +84,7 @@ steps: timeout_in_minutes: 20 source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/test_utils.py - tests/models/test_vision.py - tests/models/test_adapters.py diff --git a/.buildkite/test_areas/models_language.yaml b/.buildkite/test_areas/models_language.yaml index 2d3f14d171..1509cae7b2 100644 --- a/.buildkite/test_areas/models_language.yaml +++ b/.buildkite/test_areas/models_language.yaml @@ -8,6 +8,7 @@ steps: device: h200_18gb source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/language commands: # Test standard language models, excluding a subset of slow tests @@ -60,6 +61,7 @@ steps: timeout_in_minutes: 65 source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/language/generation commands: # Install fast path packages for testing against transformers @@ -88,6 +90,7 @@ steps: timeout_in_minutes: 65 source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/language/generation commands: - uv pip install --system --no-build-isolation 'git+https://github.com/state-spaces/mamba@v2.3.0' @@ -101,6 +104,7 @@ steps: optional: true source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/language/generation commands: # Install fast path packages for testing against transformers @@ -115,6 +119,7 @@ steps: optional: true source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/language/generation_ppl_test commands: - pytest -v -s models/language/generation_ppl_test @@ -126,6 +131,7 @@ steps: optional: true source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/language/pooling commands: - pytest -v -s models/language/pooling -m 'not core_model' @@ -144,6 +150,7 @@ steps: optional: true source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/language/pooling_mteb_test commands: - pytest -v -s models/language/pooling_mteb_test diff --git a/.buildkite/test_areas/models_multimodal.yaml b/.buildkite/test_areas/models_multimodal.yaml index 55dc870d23..4a84abd5ba 100644 --- a/.buildkite/test_areas/models_multimodal.yaml +++ b/.buildkite/test_areas/models_multimodal.yaml @@ -8,6 +8,7 @@ steps: device: h200_18gb source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/multimodal commands: - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "qwen2" @@ -26,6 +27,7 @@ steps: device: h200_18gb source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/multimodal commands: - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "qwen3 or gemma" @@ -45,6 +47,7 @@ steps: timeout_in_minutes: 40 source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/multimodal commands: - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "not qwen2 and not qwen3 and not gemma" @@ -62,6 +65,7 @@ steps: timeout_in_minutes: 75 source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/multimodal commands: - pytest -v -s models/multimodal -m core_model --ignore models/multimodal/generation/test_common.py --ignore models/multimodal/generation/test_ultravox.py --ignore models/multimodal/generation/test_qwen2_5_vl.py --ignore models/multimodal/generation/test_qwen2_vl.py --ignore models/multimodal/generation/test_whisper.py --ignore models/multimodal/generation/test_mm_prefix_lm.py --ignore models/multimodal/generation/test_memory_leak.py --ignore models/multimodal/generation/test_vit_cudagraph.py --ignore models/multimodal/processing @@ -83,6 +87,7 @@ steps: timeout_in_minutes: 125 source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/multimodal - tests/models/registry.py device: cpu-medium @@ -96,6 +101,7 @@ steps: device: h200_18gb source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/multimodal - tests/models/registry.py commands: @@ -132,6 +138,7 @@ steps: optional: true source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/multimodal/generation - tests/models/multimodal/test_mapping.py commands: @@ -168,6 +175,7 @@ steps: optional: true source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/multimodal/generation commands: - pytest -v -s models/multimodal/generation/test_common.py -m 'split(group=0) and not core_model' @@ -178,6 +186,7 @@ steps: optional: true source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/multimodal/generation commands: - pytest -v -s models/multimodal/generation/test_common.py -m 'split(group=1) and not core_model' @@ -188,6 +197,7 @@ steps: device: h200_18gb source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/multimodal/pooling commands: - pytest -v -s models/multimodal/pooling -m 'not core_model' @@ -200,4 +210,5 @@ steps: - image-build-amd source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/models/multimodal/pooling diff --git a/.buildkite/test_areas/pytorch.yaml b/.buildkite/test_areas/pytorch.yaml index 91235d7d8e..b119e981f0 100644 --- a/.buildkite/test_areas/pytorch.yaml +++ b/.buildkite/test_areas/pytorch.yaml @@ -13,6 +13,7 @@ steps: - vllm/compilation/ - vllm/config/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/env_override.py - vllm/envs.py @@ -53,6 +54,7 @@ steps: - vllm/compilation/ - vllm/config/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/env_override.py - vllm/envs.py @@ -85,6 +87,7 @@ steps: - vllm/compilation/ - vllm/config/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/env_override.py - vllm/envs.py @@ -118,6 +121,7 @@ steps: - vllm/compilation/ - vllm/config/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/env_override.py - vllm/envs.py @@ -157,6 +161,7 @@ steps: - vllm/compilation/ - vllm/config/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/env_override.py - vllm/envs.py diff --git a/.buildkite/test_areas/rust_frontend.yaml b/.buildkite/test_areas/rust_frontend.yaml index 47cb37f36d..e9e66c46c0 100644 --- a/.buildkite/test_areas/rust_frontend.yaml +++ b/.buildkite/test_areas/rust_frontend.yaml @@ -102,6 +102,7 @@ steps: source_file_dependencies: - rust/ - vllm/distributed/ + - "!vllm/distributed/kv_transfer/" - vllm/engine/ - vllm/executor/ - vllm/v1/engine/ diff --git a/.buildkite/test_areas/weight_loading.yaml b/.buildkite/test_areas/weight_loading.yaml index 2e05e5b5ba..f587b6210e 100644 --- a/.buildkite/test_areas/weight_loading.yaml +++ b/.buildkite/test_areas/weight_loading.yaml @@ -10,6 +10,7 @@ steps: optional: true source_file_dependencies: - vllm/ + - "!vllm/distributed/kv_transfer/" - tests/weight_loading commands: - bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models.txt