# In this file, you can add more tests to run either by adding a new step or # adding a new command to an existing step. See different options here for examples. # This script will be feed into Jinja template in `test-template-aws.j2` at # https://github.com/vllm-project/buildkite-ci/blob/main/scripts/test-template-aws.j2 # to generate the final pipeline yaml file. # Documentation # label(str): the name of the test. emojis allowed. # fast_check(bool): whether to run this on each commit on the fastcheck pipeline. # torch_nightly(bool): whether to run this on vllm against the torch nightly pipeline. # fast_check_only(bool): run this test on the fastcheck pipeline only # optional(bool): never run this test by default (i.e. need to unblock manually) unless it's a scheduled nightly run. # soft_fail(bool): allow this step to fail without failing the entire pipeline (useful for flaky or experimental tests). # command(str): the single command to run for tests. incompatible with commands. # commands(list): the list of commands to run for the test. incompatible with command. # mirror_hardwares(list): the list of hardware to run the test on as well. currently only supports [amdexperimental] # num_gpus(int): override the number of GPUs for the test. defaults to 1 GPU. currently supports 2,4. # num_nodes(int): whether to simulate multi-node setup by launching multiple containers on one host, # in this case, commands must be specified. the first command runs on the first host, the second # command runs on the second host. # timeout_in_minutes(int): sets a timeout for the step in minutes. if not specified, uses the default timeout. # parallelism(int): number of parallel jobs to run for this step. enables test sharding using $$BUILDKITE_PARALLEL_JOB # and $$BUILDKITE_PARALLEL_JOB_COUNT environment variables. # working_dir(str): specify the place where the command should execute, default to /vllm-workspace/tests # source_file_dependencies(list): the list of prefixes to opt-in the test for, if empty, the test will always run. # When adding a test # - If the test belongs to an existing group, add it there # - If the test is short, add to any existing step # - If the test takes more than 10min, then it is okay to create a new step. # Note that all steps execute in parallel. ##################################################################################################################################### # # # README # # # ##################################################################################################################################### # # # IMPORTANT: # # * Currently AMD CI has MI250 agents, MI325 agents, and MI355 agents. All upcoming feature improvements are tracked in: # # https://github.com/vllm-project/vllm/issues/34994 # # # #-----------------------------------------------------------------------------------------------------------------------------------# # # # NOTES: # # * [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`. # # * [Entrypoints Integration (LLM)]: # # - {`pytest -v -s entrypoints/llm/test_generate.py`}: It needs a clean process # # - {`pytest -v -s entrypoints/offline_mode`}: Needs to avoid interference with other tests # # * [Engine / Engine (1 GPU) / e2e Scheduling / e2e Core / V1 e2e / Spec Decode / V1 Sample + Logits / V1 Core + KV + Metrics]: # # - Previously a single "V1 Test e2e + engine" step, now split across multiple groups. # # - V1 e2e (2/4 GPUs) uses 4 GPUs but is scheduled on 8-GPU machines for stability. See: # # https://github.com/vllm-project/vllm/pull/31040 # # * [V1 Sample + Logits / V1 Core + KV + Metrics / V1 others (CPU)]: # # - Previously a single "V1 others" step, now split to avoid interference. # # - Integration test for streaming correctness (requires special branch for __harness__ lib). # # * [V1 others (CPU)]: Split the tests to avoid interference # # * [PyTorch Compilation Unit Tests]: Run unit tests defined directly under `compile/`, not including subdirectories, which # # are usually heavier tests covered elsewhere. Use `find` to launch multiple instances # # of pytest so that they do not suffer from: # # https://github.com/vllm-project/vllm/issues/28965 # # * [PyTorch Fullgraph Smoke Test]: Run smoke tests under fullgraph directory, except `test_full_graph.py` as it is a heavy # # test that is covered in other steps. Use `find` to launch multiple instances of pytest # # so that they do not suffer from: https://github.com/vllm-project/vllm/issues/28965 # # * [PyTorch Fullgraph]: # # - Limit to no custom ops to reduce running time. Wrap with quotes to escape yaml and avoid starting `-k` string # # with a `-` # # - Old E2E tests such as: # # ```bash # # pytest -v -s compile/distributed/test_fusions_e2e.py -k 'TRITON and not +quant_fp8 and not Llama-4' # # ``` # # were removed in https://github.com/vllm-project/vllm/pull/33293 in favor of new tests in `fusions_e2e`. We # # avoid replicating the new jobs in this file as it's deprecated. # # * [Basic Models Tests (Extra Initialization) %N]: Only when vLLM model source is modified - test initialization of a # # large subset of supported models (the complement of the small subset in # # the above test.) Also run if model initialization test file is modified. # # * [Language Models Tests (Extra Standard) %N]: Shard slow subset of standard language models tests. Only run when model # # source is modified, or when specified test files are modified. # # * [Language Models Tests (Hybrid) %N]: Install fast path packages for testing against transformers (mamba, conv1d) and to # # run plamo2 model in vLLM. # # * [Language Models Test (Extended Generation)]: Install fast path packages for testing against transformers (mamba, conv1d) # # and to run plamo2 model in vLLM. # # * [Multi-Modal Models (Standard) 1-4]: # # - Do NOT remove `VLLM_WORKER_MULTIPROC_METHOD=spawn` setting as ROCm requires this for certain models to function. # # * [Transformers Nightly Models]: Whisper needs `VLLM_WORKER_MULTIPROC_METHOD=spawn` to avoid deadlock. # # * [Plugin Tests (2 GPUs)]: # # - {`pytest -v -s entrypoints/openai/test_oot_registration.py`}: It needs a clean process # # - {`pytest -v -s models/test_oot_registration.py`}: It needs a clean process # # - {`pytest -v -s plugins/lora_resolvers`}: Unit tests for in-tree lora resolver plugins # # * [LoRA TP (Distributed)]: # # - There is some Tensor Parallelism related processing logic in LoRA that requires multi-GPU testing for validation. # # - {`pytest -v -s -x lora/test_gptoss_tp.py`}: Disabled for now because MXFP4 backend on non-cuda platform doesn't support # # LoRA yet. # # * [Distributed Tests (NxGPUs)(HW-TAG)]: Don't test llama model here, it seems hf implementation is buggy. See: # # https://github.com/vllm-project/vllm/pull/5689 # # * [Distributed Tests (NxGPUs)(HW-TAG)]: Some old E2E tests were removed in https://github.com/vllm-project/vllm/pull/33293 # # in favor of new tests in fusions_e2e. We avoid replicating the new jobs in # # this file as it's deprecated. # # # ##################################################################################################################################### steps: ##################################################################################################################################### # # # MI250 test definitions ( currently the test set is completely mirrored // TBD which tests are to be routed there ultimately) # # # ##################################################################################################################################### - label: Pytorch Nightly Dependency Override Check # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true soft_fail: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - requirements/nightly_torch_test.txt - vllm/platforms/rocm.py commands: - bash standalone_tests/pytorch_nightly_dependency.sh - label: Async Engine, Inputs, Utils, Worker # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/detokenizer - tests/multimodal - tests/utils_ commands: - pytest -v -s detokenizer - pytest -v -s -m 'not cpu_test' multimodal - pytest -v -s utils_ - label: Async Engine, Inputs, Utils, Worker, Config (CPU) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true no_gpu: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/test_inputs.py - tests/test_outputs.py - tests/test_pooling_params.py - tests/test_ray_env.py - tests/multimodal - tests/renderers - tests/standalone_tests/lazy_imports.py - tests/tokenizers_ - tests/tool_parsers - tests/transformers_utils - tests/config commands: - python3 standalone_tests/lazy_imports.py - pytest -v -s test_inputs.py - pytest -v -s test_outputs.py - pytest -v -s test_pooling_params.py - pytest -v -s test_ray_env.py - pytest -v -s -m 'cpu_test' multimodal - pytest -v -s renderers - pytest -v -s tokenizers_ - pytest -v -s tool_parsers - pytest -v -s transformers_utils - pytest -v -s config - label: Python-only Installation # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - tests/standalone_tests/python_only_compile.sh - setup.py - vllm/platforms/rocm.py commands: - bash standalone_tests/python_only_compile.sh - label: Basic Correctness # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/basic_correctness/test_basic_correctness - tests/basic_correctness/test_cpu_offload - tests/basic_correctness/test_cumem.py commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s basic_correctness/test_cumem.py - pytest -v -s basic_correctness/test_basic_correctness.py - pytest -v -s basic_correctness/test_cpu_offload.py - label: Entrypoints Unit Tests # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 fast_check: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/entrypoints - tests/entrypoints/ - vllm/platforms/rocm.py commands: - pytest -v -s entrypoints/openai/tool_parsers - 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) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/llm - tests/entrypoints/offline_mode commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/llm --ignore=entrypoints/llm/test_generate.py --ignore=entrypoints/llm/test_collective_rpc.py - pytest -v -s entrypoints/llm/test_generate.py - pytest -v -s entrypoints/offline_mode - label: Entrypoints Integration (API Server 2) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/rpc - tests/entrypoints/serve/instrumentator - tests/tool_use commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/serve/instrumentator - PYTHONPATH=/vllm-workspace pytest -v -s entrypoints/rpc - pytest -v -s tool_use - label: Entrypoints Integration (Responses API) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/openai/responses commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/openai/responses - label: EPLB Algorithm # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/eplb - tests/distributed/test_eplb_algo.py - vllm/platforms/rocm.py commands: - pytest -v -s distributed/test_eplb_algo.py - label: EPLB Execution # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_4 num_gpus: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/eplb - tests/distributed/test_eplb_execute.py - tests/distributed/test_eplb_spec_decode.py - vllm/platforms/rocm.py commands: - pytest -v -s distributed/test_eplb_execute.py - pytest -v -s distributed/test_eplb_spec_decode.py - label: Elastic EP Scaling Test # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_4 num_gpus: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/ - vllm/engine/ - vllm/executor/ - vllm/compilation/ - tests/distributed/ - vllm/platforms/rocm.py commands: - pytest -v -s distributed/test_elastic_ep.py - label: Metrics, Tracing (2 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 num_gpus: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1/tracing commands: - "pip install \ 'opentelemetry-sdk>=1.26.0' \ 'opentelemetry-api>=1.26.0' \ 'opentelemetry-exporter-otlp>=1.26.0' \ 'opentelemetry-semantic-conventions-ai>=0.4.1'" - pytest -v -s v1/tracing - label: Regression # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/test_regression commands: - pip install modelscope - pytest -v -s test_regression.py - label: Engine # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/engine - tests/test_sequence - tests/test_config - tests/test_logger - tests/test_vllm_port commands: - pytest -v -s engine test_sequence.py test_config.py test_logger.py test_vllm_port.py - label: Engine (1 GPU) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/ - tests/v1/engine/ - vllm/platforms/rocm.py commands: - pytest -v -s v1/engine/test_preprocess_error_handling.py - pytest -v -s v1/engine --ignore v1/engine/test_preprocess_error_handling.py - label: e2e Scheduling (1 GPU) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/ - tests/v1/e2e/general/ - vllm/platforms/rocm.py commands: - pytest -v -s v1/e2e/general/test_async_scheduling.py - label: e2e Core (1 GPU) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/ - tests/v1/e2e/general/ - vllm/platforms/rocm.py commands: - pytest -v -s v1/e2e/general --ignore v1/e2e/general/test_async_scheduling.py - label: Spec Decode Speculators + MTP # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/spec_decode/ - vllm/v1/worker/gpu/spec_decode/ - vllm/model_executor/model_loader/ - vllm/v1/sample/ - vllm/model_executor/layers/ - vllm/transformers_utils/configs/speculators/ - tests/v1/e2e/spec_decode/ - vllm/platforms/rocm.py commands: - pytest -v -s v1/e2e/spec_decode -k "speculators or mtp_correctness" - label: Spec Decode Ngram + Suffix # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/spec_decode/ - vllm/v1/worker/gpu/spec_decode/ - vllm/model_executor/model_loader/ - vllm/v1/sample/ - vllm/model_executor/layers/ - tests/v1/e2e/spec_decode/ - vllm/platforms/rocm.py commands: - pytest -v -s v1/e2e/spec_decode -k "ngram or suffix" - label: Spec Decode Draft Model # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/spec_decode/ - vllm/v1/worker/gpu/spec_decode/ - vllm/model_executor/model_loader/ - vllm/v1/sample/ - vllm/model_executor/layers/ - tests/v1/e2e/spec_decode/ - vllm/platforms/rocm.py commands: - pytest -v -s v1/e2e/spec_decode -k "draft_model or no_sync or batch_inference" - label: V1 e2e (2 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1/e2e commands: - pytest -v -s v1/e2e/spec_decode/test_spec_decode.py -k "tensor_parallelism" - label: V1 Sample + Logits # TBD timeout_in_minutes: 60 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1/sample - tests/v1/logits_processors - tests/v1/test_oracle.py - tests/v1/test_request.py - tests/v1/test_outputs.py commands: - pytest -v -s v1/sample - pytest -v -s v1/logits_processors - pytest -v -s v1/test_oracle.py - pytest -v -s v1/test_request.py - pytest -v -s v1/test_outputs.py - label: V1 Core + KV + Metrics # TBD timeout_in_minutes: 60 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1/core - tests/v1/executor - tests/v1/kv_offload - tests/v1/worker - tests/v1/kv_connector/unit - tests/v1/metrics - tests/entrypoints/openai/correctness/test_lmeval.py commands: - uv pip install --system -r /vllm-workspace/requirements/kv_connectors_rocm.txt - pytest -v -s -m 'not cpu_test' v1/core - pytest -v -s v1/executor - pytest -v -s v1/kv_offload - pytest -v -s v1/worker - pytest -v -s -m 'not cpu_test' v1/kv_connector/unit - pytest -v -s -m 'not cpu_test' v1/metrics - pip install -U git+https://github.com/robertgshaw2-redhat/lm-evaluation-harness.git@streaming-api - pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine - label: V1 Speculative Decoding (slow) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/spec_decode/ - vllm/model_executor/models/ - vllm/v1/attention/ - vllm/model_executor/layers/ - tests/v1/spec_decode/ - vllm/platforms/rocm.py commands: - pytest -v -s -m 'slow_test' v1/spec_decode/test_eagle.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_extract_hidden_states.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_max_len.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_mtp.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_ngram.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_speculators_eagle3.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_tree_attention.py - label: V1 attention (H100-MI250) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/config/attention.py - vllm/model_executor/layers/attention - vllm/v1/attention - tests/v1/attention - vllm/_aiter_ops.py - vllm/envs.py - vllm/platforms/rocm.py commands: - pytest -v -s v1/attention - label: V1 others (CPU) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 no_gpu: true optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1 commands: - pytest -v -s -m 'cpu_test' v1/core - pytest -v -s v1/structured_output - pytest -v -s v1/test_serial_utils.py - pytest -v -s -m 'cpu_test' v1/kv_connector/unit - pytest -v -s -m 'cpu_test' v1/metrics - label: Examples # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/examples" source_file_dependencies: - vllm/entrypoints - vllm/multimodal - examples/ - vllm/platforms/rocm.py commands: - pip install tensorizer # Basic - python3 basic/offline_inference/chat.py --attention-backend TRITON_ATTN - 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 - python3 basic/offline_inference/classify.py - python3 basic/offline_inference/embed.py - python3 basic/offline_inference/score.py # Multi-modal models - python3 offline_inference/audio_language.py --seed 0 - python3 offline_inference/vision_language.py --seed 0 - python3 offline_inference/vision_language_multi_image.py --seed 0 - python3 offline_inference/encoder_decoder_multimodal.py --model-type whisper --seed 0 # Pooling models - python3 pooling/embed/vision_embedding_offline.py --seed 0 # Features demo - python3 offline_inference/prefix_caching.py - python3 offline_inference/llm_engine_example.py - python3 others/tensorize_vllm_model.py --model facebook/opt-125m serialize --serialized-directory /tmp/ --suffix v1 && python3 others/tensorize_vllm_model.py --model facebook/opt-125m deserialize --path-to-tensors /tmp/vllm/facebook/opt-125m/v1/model.tensors - python3 offline_inference/spec_decode.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 - python3 offline_inference/spec_decode.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: Platform Tests (CUDA) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/cuda commands: - pytest -v -s cuda/test_cuda_context.py - pytest -v -s cuda/test_platform_no_cuda_init.py - label: Samplers Test # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/model_executor/layers - vllm/sampling_metadata.py - vllm/v1/sample/ - vllm/beam_search.py - tests/samplers - tests/conftest.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - pytest -v -s samplers - label: LoRA %N # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 parallelism: 4 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/lora - tests/lora - vllm/platforms/rocm.py commands: - pytest -v -s lora --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --ignore=lora/test_chatglm3_tp.py --ignore=lora/test_llama_tp.py --ignore=lora/test_llm_with_multi_loras.py --ignore=lora/test_olmoe_tp.py --ignore=lora/test_deepseekv2_tp.py --ignore=lora/test_gptoss_tp.py --ignore=lora/test_qwen3moe_tp.py - label: PyTorch Compilation Unit Tests # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 torch_nightly: true optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/compilation/ - vllm/model_executor/layers/ - vllm/v1/worker/ - vllm/v1/attention/ - vllm/v1/cudagraph_dispatcher.py - vllm/config/compilation.py - csrc/ - tests/compile - vllm/platforms/rocm.py commands: - "find compile/ -maxdepth 1 -name 'test_*.py' -print0 | xargs -0 -n1 -I{} pytest -s -v '{}'" - label: PyTorch Fullgraph Smoke Test # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/compilation/ - vllm/model_executor/ - vllm/v1/attention/ - vllm/config/compilation.py - csrc/ - tests/compile - vllm/platforms/rocm.py commands: - "find compile/fullgraph/ -name 'test_*.py' -not -name 'test_full_graph.py' -exec pytest -s -v {} \\\\;" - label: PyTorch Fullgraph # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/compilation/ - vllm/model_executor/ - vllm/v1/attention/ - vllm/config/compilation.py - csrc/ - tests/compile - vllm/platforms/rocm.py commands: - pytest -v -s compile/fullgraph/test_full_graph.py -k 'not test_fp8_kv_scale_compile' - label: Cudagraph # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - tests/v1/cudagraph - vllm/v1/cudagraph_dispatcher.py - vllm/config/compilation.py - vllm/compilation - vllm/platforms/rocm.py commands: - pytest -v -s v1/cudagraph/test_cudagraph_dispatch.py - pytest -v -s v1/cudagraph/test_cudagraph_mode.py - label: Kernels Core Operation Test # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/ - tests/kernels/core - tests/kernels/test_top_k_per_row.py - tests/kernels/test_concat_mla_q.py - vllm/model_executor/layers/rotary_embedding/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - pytest -v -s kernels/core kernels/test_top_k_per_row.py - label: Kernels Mamba Test # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/mamba/ - tests/kernels/mamba - vllm/model_executor/layers/mamba/ops - vllm/platforms/rocm.py commands: - pytest -v -s kernels/mamba - label: Kernels Helion Test # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/utils/import_utils.py - tests/kernels/helion/ - vllm/platforms/rocm.py commands: - pip install helion==0.3.3 - pytest -v -s kernels/helion/ - label: Model Executor # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/engine/arg_utils.py - vllm/config/model.py - vllm/model_executor - tests/model_executor - tests/entrypoints/openai/completion/test_tensorizer_entrypoint.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - apt-get update && apt-get install -y curl libsodium23 - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s model_executor -m '(not slow_test)' - pytest -v -s entrypoints/openai/completion/test_tensorizer_entrypoint.py - label: Benchmarks # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/.buildkite" source_file_dependencies: - benchmarks/ - vllm/platforms/rocm.py commands: - bash scripts/run-benchmarks.sh - label: Benchmarks CLI Test # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/benchmarks/ commands: - pytest -v -s benchmarks/ - label: OpenAI API correctness # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/ - vllm/entrypoints/openai/ - vllm/model_executor/models/whisper.py - vllm/model_executor/layers/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py - vllm/model_executor/model_loader/ commands: - bash ../tools/install_torchcodec_rocm.sh || exit 1 - pytest -s entrypoints/openai/correctness/ - label: Basic Models Tests (Initialization) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/test_initialization.py - tests/models/registry.py commands: - pytest -v -s models/test_initialization.py::test_can_initialize_small_subset - label: Basic Models Tests (Extra Initialization) %N # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 torch_nightly: true parallelism: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/model_executor/models/ - vllm/model_executor/layers/ - tests/models/test_initialization.py - tests/models/registry.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - pytest -v -s models/test_initialization.py -k 'not test_can_initialize_small_subset' --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --shard-id=$$BUILDKITE_PARALLEL_JOB - label: Basic Models Tests (Other) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/test_terratorch.py - tests/models/test_transformers.py - tests/models/test_registry.py commands: - pytest -v -s models/test_terratorch.py models/test_transformers.py models/test_registry.py - label: Basic Models Test (Other CPU) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 no_gpu: true optional: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/test_utils.py - tests/models/test_vision.py commands: - pytest -v -s models/test_utils.py models/test_vision.py - label: Language Models Tests (Extra Standard) %N # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 torch_nightly: true parallelism: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/model_executor/layers/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - tests/models/language/pooling/test_embedding.py - tests/models/language/generation/test_common.py - tests/models/language/pooling/test_classification.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - pip freeze | grep -E 'torch' - export TORCH_NCCL_BLOCKING_WAIT=1 - pytest -v -s models/language -m 'core_model and slow_test' --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --shard-id=$$BUILDKITE_PARALLEL_JOB - label: Language Models Test (PPL) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/language/generation_ppl_test commands: - pytest -v -s models/language/generation_ppl_test - label: Language Models Test (Extended Pooling) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/language/pooling commands: - pytest -v -s models/language/pooling -m 'not core_model' - label: Language Models Test (MTEB) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/language/pooling_mteb_test commands: - pytest -v -s models/language/pooling_mteb_test - label: Multi-Modal Processor (CPU) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 no_gpu: true optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal - tests/models/registry.py commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/processing --ignore models/multimodal/processing/test_tensor_schema.py - label: Multi-Modal Accuracy Eval (Small Models) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 optional: true working_dir: "/vllm-workspace/.buildkite/lm-eval-harness" source_file_dependencies: - vllm/multimodal/ - vllm/inputs/ - vllm/v1/core/ - vllm/platforms/rocm.py - vllm/model_executor/model_loader/ commands: - pytest -s -v test_lm_eval_correctness.py --config-list-file=configs/models-mm-small.txt --tp-size=1 - label: "Multi-Modal Models (Standard) 1: qwen2" # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "qwen2" - pytest -v -s models/multimodal/generation/test_ultravox.py -m core_model - label: "Multi-Modal Models (Standard) 2: qwen3 + gemma" # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "qwen3 or gemma" - pytest -v -s models/multimodal/generation/test_qwen2_5_vl.py -m core_model - label: "Multi-Modal Models (Standard) 3: llava + qwen2_vl" # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "not qwen2 and not qwen3 and not gemma" - pytest -v -s models/multimodal/generation/test_qwen2_vl.py -m core_model - label: "Multi-Modal Models (Standard) 4: other + whisper" # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation - tests/models/multimodal/test_mapping.py commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - 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/processing - cd .. && VLLM_WORKER_MULTIPROC_METHOD=spawn pytest -v -s tests/models/multimodal/generation/test_whisper.py -m core_model - label: Multi-Modal Models (Extended Generation 1) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation - tests/models/multimodal/test_mapping.py commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation -m 'not core_model' --ignore models/multimodal/generation/test_common.py - pytest -v -s models/multimodal/test_mapping.py - label: Multi-Modal Models (Extended Generation 2) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m 'split(group=0) and not core_model' - label: Multi-Modal Models (Extended Generation 3) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m 'split(group=1) and not core_model' - label: Multi-Modal Models (Extended Pooling) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/pooling commands: - pytest -v -s models/multimodal/pooling -m 'not core_model' - label: Distributed Comm Ops # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 num_gpus: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed - tests/distributed - vllm/platforms/rocm.py commands: - pytest -v -s distributed/test_comm_ops.py - pytest -v -s distributed/test_shm_broadcast.py - pytest -v -s distributed/test_shm_buffer.py - pytest -v -s distributed/test_shm_storage.py - label: Distributed DP Tests (2 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_2 num_gpus: 2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/ - vllm/engine/ - vllm/executor/ - vllm/worker/worker_base.py - vllm/v1/engine/ - vllm/v1/worker/ - tests/v1/distributed - tests/entrypoints/openai/test_multi_api_servers.py - vllm/platforms/rocm.py commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - TP_SIZE=1 DP_SIZE=2 pytest -v -s v1/distributed/test_async_llm_dp.py - 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/openai/test_multi_api_servers.py - label: Distributed Compile + RPC Tests (2 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_2 num_gpus: 2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/compilation/ - vllm/distributed/ - vllm/engine/ - vllm/executor/ - vllm/worker/worker_base.py - vllm/v1/engine/ - vllm/v1/worker/ - tests/compile/fullgraph/test_basic_correctness.py - tests/compile/test_wrapper.py - tests/entrypoints/llm/test_collective_rpc.py - vllm/platforms/rocm.py commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - pytest -v -s entrypoints/llm/test_collective_rpc.py - pytest -v -s ./compile/fullgraph/test_basic_correctness.py - pytest -v -s ./compile/test_wrapper.py - label: Distributed Torchrun + Shutdown Tests (2 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_2 num_gpus: 2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/ - vllm/engine/ - vllm/executor/ - vllm/worker/worker_base.py - vllm/v1/engine/ - vllm/v1/worker/ - tests/distributed/ - tests/v1/shutdown - tests/v1/worker/test_worker_memory_snapshot.py - vllm/platforms/rocm.py commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - VLLM_TEST_SAME_HOST=1 torchrun --nproc-per-node=4 distributed/test_same_node.py | grep 'Same node test passed' - VLLM_TEST_SAME_HOST=1 VLLM_TEST_WITH_DEFAULT_DEVICE_SET=1 torchrun --nproc-per-node=4 distributed/test_same_node.py | grep 'Same node test passed' - CUDA_VISIBLE_DEVICES=0,1 pytest -v -s v1/shutdown - pytest -v -s v1/worker/test_worker_memory_snapshot.py - label: Distributed Model Tests (2 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_2 num_gpus: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/model_executor/model_loader/sharded_state_loader.py - vllm/model_executor/models/ - vllm/model_executor/layers/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - tests/basic_correctness/ - tests/model_executor/model_loader/test_sharded_state_loader.py - tests/models/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - TARGET_TEST_SUITE=L4 pytest basic_correctness/ -v -s -m 'distributed(num_gpus=2)' - CUDA_VISIBLE_DEVICES=0,1 pytest -v -s model_executor/model_loader/test_sharded_state_loader.py -m '(not slow_test)' - pytest models/test_transformers.py -v -s -m 'distributed(num_gpus=2)' - pytest models/language -v -s -m 'distributed(num_gpus=2)' - pytest models/multimodal -v -s -m 'distributed(num_gpus=2)' --ignore models/multimodal/generation/test_whisper.py - VLLM_WORKER_MULTIPROC_METHOD=spawn pytest models/multimodal/generation/test_whisper.py -v -s -m 'distributed(num_gpus=2)' - label: Plugin Tests (2 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_2 num_gpus: 2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/plugins/ - tests/plugins/ - vllm/platforms/rocm.py commands: # BEGIN: platform plugin and general plugin tests, all the code in-between runs on dummy platform - pip install -e ./plugins/vllm_add_dummy_platform - pytest -v -s plugins_tests/test_platform_plugins.py - pip uninstall vllm_add_dummy_platform -y # END: platform plugin tests # BEGIN: `io_processor` plugins test, all the code in between uses the `prithvi_io_processor` plugin - pip install -e ./plugins/prithvi_io_processor_plugin - pytest -v -s plugins_tests/test_io_processor_plugins.py - pip uninstall prithvi_io_processor_plugin -y # END: `io_processor` plugins test # BEGIN: `bge_m3_sparse io_processor` test - pip install -e ./plugins/bge_m3_sparse_plugin - pytest -v -s plugins_tests/test_bge_m3_sparse_io_processor_plugins.py - pip uninstall bge_m3_sparse_plugin -y # END: `bge_m3_sparse io_processor` test # BEGIN: `stat_logger` plugins test - pip install -e ./plugins/vllm_add_dummy_stat_logger - pytest -v -s plugins_tests/test_stats_logger_plugins.py - pip uninstall dummy_stat_logger -y # END: `stat_logger` plugins test # BEGIN: other tests - pytest -v -s plugins_tests/test_scheduler_plugins.py - pip install -e ./plugins/vllm_add_dummy_model - pytest -v -s distributed/test_distributed_oot.py - pytest -v -s entrypoints/openai/chat_completion/test_oot_registration.py - pytest -v -s models/test_oot_registration.py - pytest -v -s plugins/lora_resolvers # END: other tests - label: Pipeline + Context Parallelism (4 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_4 num_gpus: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/ - vllm/engine/ - vllm/executor/ - vllm/model_executor/models/ - vllm/model_executor/layers/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - tests/distributed/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - pytest -v -s distributed/test_pp_cudagraph.py - pytest -v -s distributed/test_pipeline_parallel.py - label: Ray Dependency Compatibility Check # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_1 working_dir: "/" source_file_dependencies: - requirements/ - setup.py - vllm/platforms/rocm.py commands: - bash /vllm-workspace/.buildkite/scripts/check-ray-compatibility.sh - label: Distributed NixlConnector PD accuracy (4 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_4 num_gpus: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/kv_transfer/kv_connector/v1/nixl_connector.py - tests/v1/kv_connector/nixl_integration/ - vllm/platforms/rocm.py commands: - uv pip install --system -r /vllm-workspace/requirements/kv_connectors_rocm.txt - ROCM_ATTN=1 bash v1/kv_connector/nixl_integration/config_sweep_accuracy_test.sh - label: DP EP Distributed NixlConnector PD accuracy tests (4 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_4 num_gpus: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/kv_transfer/kv_connector/v1/nixl_connector.py - tests/v1/kv_connector/nixl_integration/ - vllm/platforms/rocm.py commands: - uv pip install --system -r /vllm-workspace/requirements/kv_connectors_rocm.txt - DP_EP=1 ROCM_ATTN=1 bash v1/kv_connector/nixl_integration/config_sweep_accuracy_test.sh - label: NixlConnector PD + Spec Decode acceptance (2 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_2 num_gpus: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/kv_transfer/kv_connector/v1/nixl_connector.py - vllm/v1/worker/kv_connector_model_runner_mixin.py - tests/v1/kv_connector/nixl_integration/ - vllm/platforms/rocm.py commands: - uv pip install --system -r /vllm-workspace/requirements/kv_connectors_rocm.txt - ROCM_ATTN=1 bash v1/kv_connector/nixl_integration/spec_decode_acceptance_test.sh - label: CrossLayer KV layout Distributed NixlConnector PD accuracy tests (4 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_4 num_gpus: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/kv_transfer/kv_connector/v1/nixl_connector.py - tests/v1/kv_connector/nixl_integration/ - vllm/platforms/rocm.py commands: - uv pip install --system -r /vllm-workspace/requirements/kv_connectors_rocm.txt - CROSS_LAYERS_BLOCKS=True ROCM_ATTN=1 bash v1/kv_connector/nixl_integration/config_sweep_accuracy_test.sh - label: Hyrbid SSM NixlConnector PD accuracy tests (4 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi250_4 num_gpus: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/kv_transfer/kv_connector/v1/nixl_connector.py - tests/v1/kv_connector/nixl_integration/ - vllm/platforms/rocm.py commands: - uv pip install --system -r /vllm-workspace/requirements/kv_connectors_rocm.txt - HYBRID_SSM=1 ROCM_ATTN=1 bash v1/kv_connector/nixl_integration/config_sweep_accuracy_test.sh - label: Distributed Tests (2 GPUs)(H100-MI250) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250] agent_pool: mi325_2 num_gpus: 2 working_dir: "/vllm-workspace/" source_file_dependencies: - vllm/distributed/ - vllm/v1/distributed/ - vllm/model_executor/layers/fused_moe/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - tests/distributed/test_context_parallel.py - examples/offline_inference/data_parallel.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - pytest -v -s tests/distributed/test_context_parallel.py - VLLM_LOGGING_LEVEL=DEBUG python3 examples/offline_inference/data_parallel.py --model=Qwen/Qwen1.5-MoE-A2.7B -tp=1 -dp=2 --max-model-len=2048 --all2all-backend=allgather_reducescatter --disable-nccl-for-dp-synchronization ##################################################################################################################################### # # # gfx942 # # # ##################################################################################################################################### - label: Entrypoints Integration (LLM) # 13.1m timeout_in_minutes: 22 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/llm - tests/entrypoints/offline_mode commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/llm --ignore=entrypoints/llm/test_generate.py --ignore=entrypoints/llm/test_collective_rpc.py - pytest -v -s entrypoints/llm/test_generate.py - pytest -v -s entrypoints/offline_mode - label: Entrypoints Integration (API Server openai - Part 1) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/openai - tests/entrypoints/test_chat_utils commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/openai/chat_completion --ignore=entrypoints/openai/chat_completion/test_chat_with_tool_reasoning.py --ignore=entrypoints/openai/chat_completion/test_oot_registration.py - label: Entrypoints Integration (API Server openai - Part 2) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/openai - tests/entrypoints/test_chat_utils commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/openai/completion --ignore=entrypoints/openai/completion/test_tensorizer_entrypoint.py - pytest -v -s entrypoints/openai/speech_to_text/ - pytest -v -s entrypoints/test_chat_utils.py - label: Entrypoints Integration (API Server openai - Part 3) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/openai - tests/entrypoints/test_chat_utils commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - 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 - label: Entrypoints Integration (API Server 2) #26.9m timeout_in_minutes: 45 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/rpc - tests/entrypoints/serve/instrumentator - tests/tool_use commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/serve/instrumentator - PYTHONPATH=/vllm-workspace pytest -v -s entrypoints/rpc - pytest -v -s tool_use - label: Entrypoints Integration (Pooling) # 22.8m timeout_in_minutes: 48 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/pooling commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/pooling - label: Distributed Torchrun + Examples (4 GPUs) # TBD timeout_in_minutes: 80 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 num_gpus: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/ - tests/distributed/test_torchrun_example.py - tests/distributed/test_torchrun_example_moe.py - examples/rl/ - tests/examples/offline_inference/data_parallel.py - vllm/platforms/rocm.py commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - torchrun --nproc-per-node=4 distributed/test_torchrun_example.py - PP_SIZE=2 torchrun --nproc-per-node=4 distributed/test_torchrun_example.py - TP_SIZE=4 torchrun --nproc-per-node=4 distributed/test_torchrun_example_moe.py - PP_SIZE=2 TP_SIZE=2 torchrun --nproc-per-node=4 distributed/test_torchrun_example_moe.py - DP_SIZE=4 ENABLE_EP=1 torchrun --nproc-per-node=4 distributed/test_torchrun_example_moe.py - TP_SIZE=2 DP_SIZE=2 ENABLE_EP=1 torchrun --nproc-per-node=4 distributed/test_torchrun_example_moe.py - python3 ../examples/offline_inference/data_parallel.py --enforce-eager # rlhf examples - VLLM_ALLOW_INSECURE_SERIALIZATION=1 python3 ../examples/rl/rlhf_nccl.py - VLLM_ALLOW_INSECURE_SERIALIZATION=1 python3 ../examples/rl/rlhf_ipc.py - label: Distributed DP Tests (4 GPUs) # TBD timeout_in_minutes: 60 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 num_gpus: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/ - tests/v1/distributed - tests/v1/engine/test_engine_core_client.py - tests/distributed/test_utils - vllm/platforms/rocm.py commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - TP_SIZE=2 DP_SIZE=2 pytest -v -s v1/distributed/test_async_llm_dp.py - TP_SIZE=2 DP_SIZE=2 pytest -v -s v1/distributed/test_eagle_dp.py - TP_SIZE=2 DP_SIZE=2 pytest -v -s v1/distributed/test_external_lb_dp.py - TP_SIZE=1 DP_SIZE=4 pytest -v -s v1/distributed/test_internal_lb_dp.py - TP_SIZE=1 DP_SIZE=4 pytest -v -s v1/distributed/test_hybrid_lb_dp.py - pytest -v -s v1/engine/test_engine_core_client.py::test_kv_cache_events_dp - pytest -v -s distributed/test_utils.py - label: Distributed Compile + Comm (4 GPUs) # TBD timeout_in_minutes: 40 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 num_gpus: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/ - tests/distributed/test_pynccl - tests/distributed/test_events - tests/compile/fullgraph/test_basic_correctness.py - tests/distributed/test_symm_mem_allreduce.py - tests/distributed/test_multiproc_executor.py - vllm/platforms/rocm.py commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - pytest -v -s compile/fullgraph/test_basic_correctness.py - pytest -v -s distributed/test_pynccl.py - pytest -v -s distributed/test_events.py - pytest -v -s distributed/test_symm_mem_allreduce.py - pytest -v -s distributed/test_multiproc_executor.py::test_multiproc_executor_multi_node - label: Distributed Tests (8 GPUs)(H100-MI325) # 6.4m timeout_in_minutes: 10 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_8 num_gpus: 8 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - examples/offline_inference/torchrun_dp_example.py - vllm/config/parallel.py - vllm/distributed/ - vllm/v1/engine/llm_engine.py - vllm/v1/executor/uniproc_executor.py - vllm/v1/worker/gpu_worker.py - vllm/platforms/rocm.py commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - torchrun --nproc-per-node=8 ../examples/offline_inference/torchrun_dp_example.py --tp-size=2 --pp-size=1 --dp-size=4 --enable-ep - label: Elastic EP Scaling Test # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 num_gpus: 4 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/ - vllm/engine/ - vllm/executor/ - vllm/compilation/ - tests/distributed/ - vllm/platforms/rocm.py commands: - pytest -v -s distributed/test_elastic_ep.py - label: Engine # 11.3m timeout_in_minutes: 35 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/engine - tests/test_sequence - tests/test_config - tests/test_logger - tests/test_vllm_port commands: - pytest -v -s engine test_sequence.py test_config.py test_logger.py test_vllm_port.py - label: Engine (1 GPU) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/engine/ - tests/v1/engine/ - vllm/platforms/rocm.py commands: - pytest -v -s v1/engine/test_preprocess_error_handling.py - pytest -v -s v1/engine --ignore v1/engine/test_preprocess_error_handling.py - label: e2e Scheduling (1 GPU) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/ - tests/v1/e2e/general/ - vllm/platforms/rocm.py commands: - pytest -v -s v1/e2e/general/test_async_scheduling.py - label: e2e Core (1 GPU) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/ - tests/v1/e2e/ - vllm/platforms/rocm.py commands: - pytest -v -s v1/e2e/general --ignore v1/e2e/general/test_async_scheduling.py - label: Spec Decode Eagle # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/spec_decode/ - vllm/v1/worker/gpu/spec_decode/ - vllm/model_executor/model_loader/ - vllm/v1/sample/ - vllm/model_executor/layers/ - tests/v1/e2e/spec_decode/ - vllm/platforms/rocm.py commands: - pytest -v -s v1/e2e/spec_decode -k "eagle_correctness" - label: Spec Decode Speculators + MTP # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/spec_decode/ - vllm/v1/worker/gpu/spec_decode/ - vllm/model_executor/model_loader/ - vllm/v1/sample/ - vllm/model_executor/layers/ - vllm/transformers_utils/configs/speculators/ - tests/v1/e2e/spec_decode/ - vllm/platforms/rocm.py commands: - pytest -v -s v1/e2e/spec_decode -k "speculators or mtp_correctness" - label: Spec Decode Ngram + Suffix # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/spec_decode/ - vllm/v1/worker/gpu/spec_decode/ - vllm/model_executor/model_loader/ - vllm/v1/sample/ - vllm/model_executor/layers/ - tests/v1/e2e/spec_decode/ - vllm/platforms/rocm.py commands: - pytest -v -s v1/e2e/spec_decode -k "ngram or suffix" - label: Spec Decode Draft Model # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/spec_decode/ - vllm/v1/worker/gpu/spec_decode/ - vllm/model_executor/model_loader/ - vllm/v1/sample/ - vllm/model_executor/layers/ - tests/v1/e2e/spec_decode/ - vllm/platforms/rocm.py commands: - pytest -v -s v1/e2e/spec_decode -k "draft_model or no_sync or batch_inference" - label: V1 e2e (2 GPUs) # 7.1m timeout_in_minutes: 12 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1/e2e commands: - pytest -v -s v1/e2e/spec_decode/test_spec_decode.py -k "tensor_parallelism" - label: V1 e2e (4 GPUs) # 52.6m timeout_in_minutes: 106 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1/e2e commands: - pytest -v -s v1/e2e/spec_decode/test_spec_decode.py -k "eagle_correctness_heavy" - label: V1 e2e (4xH100-4xMI325) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 optional: true source_file_dependencies: - vllm/v1/attention/backends/utils.py - vllm/v1/worker/gpu_model_runner.py - tests/v1/e2e/test_hybrid_chunked_prefill.py commands: - pytest -v -s v1/e2e/test_hybrid_chunked_prefill.py - label: V1 Spec Decode # TBD timeout_in_minutes: 40 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1/spec_decode commands: - pytest -v -s -m 'not slow_test' v1/spec_decode - label: V1 Sample + Logits # TBD timeout_in_minutes: 40 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1/sample - tests/v1/logits_processors - tests/v1/test_oracle.py - tests/v1/test_request.py - tests/v1/test_outputs.py commands: - pytest -v -s v1/sample - pytest -v -s v1/logits_processors - pytest -v -s v1/test_oracle.py - pytest -v -s v1/test_request.py - pytest -v -s v1/test_outputs.py - label: V1 Core + KV + Metrics # TBD timeout_in_minutes: 40 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1/core - tests/v1/executor - tests/v1/kv_offload - tests/v1/worker - tests/v1/kv_connector/unit - tests/v1/metrics - tests/entrypoints/openai/correctness/test_lmeval.py commands: - uv pip install --system -r /vllm-workspace/requirements/kv_connectors_rocm.txt - pytest -v -s -m 'not cpu_test' v1/core - pytest -v -s v1/executor - pytest -v -s v1/kv_offload - pytest -v -s v1/worker - pytest -v -s -m 'not cpu_test' v1/kv_connector/unit - pytest -v -s -m 'not cpu_test' v1/metrics - pip install -U git+https://github.com/robertgshaw2-redhat/lm-evaluation-harness.git@streaming-api # - export HSA_NO_SCRATCH_RECLAIM=1 - pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine - label: V1 Speculative Decoding (slow) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/spec_decode/ - vllm/model_executor/models/ - vllm/v1/attention/ - vllm/model_executor/layers/ - tests/v1/spec_decode/ - vllm/platforms/rocm.py commands: - pytest -v -s -m 'slow_test' v1/spec_decode/test_eagle.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_extract_hidden_states.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_max_len.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_mtp.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_ngram.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_speculators_eagle3.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_tree_attention.py - label: Acceptance Length Test (Large Models) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/spec_decode/ - vllm/model_executor/models/mlp_speculator.py - tests/v1/spec_decode/test_acceptance_length.py - vllm/platforms/rocm.py commands: - export VLLM_ALLOW_INSECURE_SERIALIZATION=1 - pytest -v -s v1/spec_decode/test_acceptance_length.py -m slow_test - label: V1 attention (H100-MI325) # 14.5m timeout_in_minutes: 40 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/config/attention.py - vllm/model_executor/layers/attention - vllm/v1/attention - tests/v1/attention - vllm/_aiter_ops.py - vllm/envs.py - vllm/platforms/rocm.py commands: - pytest -v -s v1/attention - label: Batch Invariance (H100-MI325) # 5.2m timeout_in_minutes: 12 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/attention - vllm/model_executor/layers - tests/v1/determinism/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pip install pytest-timeout pytest-forked - pytest -v -s v1/determinism/test_batch_invariance.py - pytest -v -s v1/determinism/test_rms_norm_batch_invariant.py - label: V1 others (CPU) # 10.4m timeout_in_minutes: 28 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 no_gpu: true optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1 commands: - pytest -v -s -m 'cpu_test' v1/core - pytest -v -s v1/structured_output - pytest -v -s v1/test_serial_utils.py - pytest -v -s -m 'cpu_test' v1/kv_connector/unit - pytest -v -s -m 'cpu_test' v1/metrics - label: Examples # 24.5m timeout_in_minutes: 55 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/examples" source_file_dependencies: - vllm/entrypoints - vllm/multimodal - examples/ - vllm/platforms/rocm.py commands: - pip install tensorizer # Basic - python3 basic/offline_inference/chat.py --attention-backend TRITON_ATTN - 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 - python3 basic/offline_inference/classify.py - python3 basic/offline_inference/embed.py - python3 basic/offline_inference/score.py # Multi-modal models - python3 offline_inference/audio_language.py --seed 0 - python3 offline_inference/vision_language.py --seed 0 - python3 offline_inference/vision_language_multi_image.py --seed 0 - python3 offline_inference/encoder_decoder_multimodal.py --model-type whisper --seed 0 # Pooling models - python3 pooling/embed/vision_embedding_offline.py --seed 0 # Features demo - python3 offline_inference/prefix_caching.py - python3 offline_inference/llm_engine_example.py - python3 others/tensorize_vllm_model.py --model facebook/opt-125m serialize --serialized-directory /tmp/ --suffix v1 && python3 others/tensorize_vllm_model.py --model facebook/opt-125m deserialize --path-to-tensors /tmp/vllm/facebook/opt-125m/v1/model.tensors - python3 offline_inference/spec_decode.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 - python3 offline_inference/spec_decode.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: Platform Tests (CUDA) # 5.0m timeout_in_minutes: 9 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/cuda commands: - pytest -v -s cuda/test_cuda_context.py - pytest -v -s cuda/test_platform_no_cuda_init.py - label: PyTorch Compilation Passes Unit Tests # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/compile/passes commands: - pytest -s -v compile/passes --ignore compile/passes/distributed - label: Kernels Core Operation Test # 26.8m timeout_in_minutes: 38 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/ - tests/kernels/core - tests/kernels/test_top_k_per_row.py - tests/kernels/test_concat_mla_q.py - vllm/model_executor/layers/rotary_embedding/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - pytest -v -s kernels/core kernels/test_top_k_per_row.py - label: Kernels Attention Test %N # 17.7m timeout_in_minutes: 28 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 parallelism: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/attention/ - vllm/v1/attention - vllm/model_executor/layers/attention - tests/kernels/attention - vllm/_aiter_ops.py - vllm/envs.py - vllm/platforms/rocm.py commands: - pytest -v -s kernels/attention --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT - label: Kernels Quantization Test %N # 15.2m timeout_in_minutes: 24 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 parallelism: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/quantization/ - vllm/model_executor/layers/quantization - tests/kernels/quantization - tests/kernels/quantization/test_rocm_skinny_gemms.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py - vllm/model_executor/kernels/ commands: - pytest -v -s kernels/quantization --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT - label: Kernels MoE Test %N # TBD timeout_in_minutes: 19 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 parallelism: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/quantization/cutlass_w8a8/moe/ - csrc/moe/ - tests/kernels/moe - vllm/model_executor/layers/fused_moe/ - vllm/distributed/device_communicators/ - vllm/envs.py - vllm/config - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - pytest -v -s kernels/moe --ignore=kernels/moe/test_modular_oai_triton_moe.py --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT - pytest -v -s kernels/moe/test_modular_oai_triton_moe.py --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT - label: Kernels FP8 MoE Test # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/moe/ - csrc/quantization/w8a8/cutlass/moe/ - vllm/model_executor/layers/fused_moe/ - tests/kernels/moe/test_deepep_moe.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py - vllm/envs.py commands: - pytest -v -s kernels/moe/test_deepep_moe.py - label: ROCm AITER Ops Test # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/_aiter_ops.py - vllm/envs.py - vllm/platforms/rocm.py - tests/rocm/aiter/ - vllm/v1/attention/backends/mla/rocm_aiter_mla.py - vllm/v1/attention/selector.py commands: - pytest -v -s rocm/aiter/ - label: Benchmarks # 8.2m timeout_in_minutes: 20 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/.buildkite" source_file_dependencies: - benchmarks/ - vllm/platforms/rocm.py commands: - bash scripts/run-benchmarks.sh - label: Quantization # 36.1m timeout_in_minutes: 60 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/ - vllm/model_executor/layers/quantization - vllm/_aiter_ops.py - vllm/platforms/rocm.py - tests/quantization commands: # temporary install here since we need nightly, will move to requirements/test.in # after torchao 0.12 release, and pin a working version of torchao nightly here # since torchao nightly is only compatible with torch nightly currently # https://github.com/pytorch/ao/issues/2919, we'll have to skip new torchao tests for now # we can only upgrade after this is resolved # TODO(jerryzh168): resolve the above comment - uv pip install --system torchao==0.17.0 - uv pip install --system conch-triton-kernels - VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization/ --ignore quantization/test_blackwell_moe.py - label: Language Models Tests (Standard) # 22.8m timeout_in_minutes: 38 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/language commands: - pip freeze | grep -E 'torch' - pytest -v -s models/language -m 'core_model and (not slow_test)' - label: Language Models Tests (Hybrid) %N # 34.9m timeout_in_minutes: 55 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 torch_nightly: true parallelism: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/language/generation commands: - uv pip install --system --no-build-isolation 'git+https://github.com/AndreasKaratzas/mamba@fix-rocm-7.0-warp-size-constexpr' - uv pip install --system --no-build-isolation 'git+https://github.com/Dao-AILab/causal-conv1d@v1.5.2' - pytest -v -s models/language/generation -m hybrid_model --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --shard-id=$$BUILDKITE_PARALLEL_JOB - label: Language Models Test (Extended Generation) # 32.2m timeout_in_minutes: 55 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/language/generation commands: - uv pip install --system --no-build-isolation 'git+https://github.com/AndreasKaratzas/mamba@fix-rocm-7.0-warp-size-constexpr' - uv pip install --system --no-build-isolation 'git+https://github.com/Dao-AILab/causal-conv1d@v1.5.2' - pytest -v -s models/language/generation -m '(not core_model) and (not hybrid_model)' - label: Multi-Modal Processor # 1h 42m timeout_in_minutes: 138 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal - tests/models/registry.py commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/processing/test_tensor_schema.py - label: "Multi-Modal Models (Standard) 1: qwen2" # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 torch_nightly: true optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "qwen2" - pytest -v -s models/multimodal/generation/test_ultravox.py -m core_model - label: "Multi-Modal Models (Standard) 2: qwen3 + gemma" # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "qwen3 or gemma" - pytest -v -s models/multimodal/generation/test_qwen2_5_vl.py -m core_model - label: "Multi-Modal Models (Standard) 3: llava + qwen2_vl" # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 torch_nightly: true optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation - tests/models/multimodal/test_mapping.py commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "not qwen2 and not qwen3 and not gemma" - pytest -v -s models/multimodal/generation/test_qwen2_vl.py -m core_model - label: "Multi-Modal Models (Standard) 4: other + whisper" # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 torch_nightly: true optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - 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/processing - cd .. && VLLM_WORKER_MULTIPROC_METHOD=spawn pytest -v -s tests/models/multimodal/generation/test_whisper.py -m core_model - label: Multi-Modal Models (Extended Generation 1) # 1h 2m timeout_in_minutes: 106 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation - tests/models/multimodal/test_mapping.py commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation -m 'not core_model' --ignore models/multimodal/generation/test_common.py - pytest -v -s models/multimodal/test_mapping.py - label: Multi-Modal Models (Extended Generation 2) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m 'split(group=0) and not core_model' - label: Multi-Modal Models (Extended Generation 3) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m 'split(group=1) and not core_model' - label: Multi-Modal Models (Extended Pooling) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/pooling commands: - pytest -v -s models/multimodal/pooling -m 'not core_model' - label: Quantized Models Test # 21.4m timeout_in_minutes: 38 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/model_executor/layers/quantization - vllm/_aiter_ops.py - vllm/platforms/rocm.py - tests/models/quantization - vllm/model_executor/model_loader/ commands: - pytest -v -s models/quantization - label: Transformers Nightly Models # 50.9m timeout_in_minutes: 102 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/vllm-workspace/" source_file_dependencies: - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/multimodal/ - vllm/model_executor/layers/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py - tests/models/ - examples/ commands: - pip install --upgrade git+https://github.com/huggingface/transformers - pytest -v -s tests/models/test_initialization.py - pytest -v -s tests/models/test_transformers.py - pytest -v -s tests/models/multimodal/processing/ - pytest -v -s tests/models/multimodal/test_mapping.py - python3 examples/basic/offline_inference/chat.py - python3 examples/offline_inference/vision_language.py --model-type qwen2_5_vl - VLLM_WORKER_MULTIPROC_METHOD=spawn python3 examples/offline_inference/audio_language.py --model-type whisper - label: Quantized MoE Test (B200-MI325) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/" source_file_dependencies: - tests/quantization/test_gfx3xx_moe.py - vllm/model_executor/models/deepseek_v2.py - vllm/model_executor/models/gpt_oss.py - vllm/model_executor/models/llama4.py - vllm/model_executor/layers/fused_moe - vllm/model_executor/layers/quantization/compressed_tensors - vllm/model_executor/layers/quantization/modelopt.py - vllm/model_executor/layers/quantization/mxfp4.py - vllm/v1/attention/backends/triton_attn.py - vllm/v1/attention/backends/rocm_attn.py - vllm/v1/attention/backends/rocm_aiter_fa.py - vllm/v1/attention/backends/mla/ - vllm/v1/attention/selector.py - vllm/model_executor/layers/layernorm.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py - vllm/model_executor/model_loader/ commands: - pytest -s -v tests/quantization/test_gfx3xx_moe.py - label: Distributed DP Tests (2 GPUs) # 56.1m timeout_in_minutes: 102 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_2 num_gpus: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/ - vllm/engine/ - vllm/executor/ - vllm/worker/worker_base.py - vllm/v1/engine/ - vllm/v1/worker/ - tests/v1/distributed - tests/entrypoints/openai/test_multi_api_servers.py - vllm/platforms/rocm.py commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - TP_SIZE=1 DP_SIZE=2 pytest -v -s v1/distributed/test_async_llm_dp.py - 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/openai/test_multi_api_servers.py - label: Distributed Compile + RPC Tests (2 GPUs) # 56.1m timeout_in_minutes: 102 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_2 num_gpus: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/compilation/ - vllm/distributed/ - vllm/engine/ - vllm/executor/ - vllm/worker/worker_base.py - vllm/v1/engine/ - vllm/v1/worker/ - tests/compile/fullgraph/test_basic_correctness.py - tests/compile/test_wrapper.py - tests/entrypoints/llm/test_collective_rpc.py - vllm/platforms/rocm.py commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - pytest -v -s entrypoints/llm/test_collective_rpc.py - pytest -v -s ./compile/fullgraph/test_basic_correctness.py - pytest -v -s ./compile/test_wrapper.py - label: Distributed Torchrun + Shutdown Tests (2 GPUs) # 56.1m timeout_in_minutes: 102 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_2 num_gpus: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/ - vllm/engine/ - vllm/executor/ - vllm/worker/worker_base.py - vllm/v1/engine/ - vllm/v1/worker/ - tests/distributed/ - tests/v1/shutdown - tests/v1/worker/test_worker_memory_snapshot.py - vllm/platforms/rocm.py commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - VLLM_TEST_SAME_HOST=1 torchrun --nproc-per-node=4 distributed/test_same_node.py | grep 'Same node test passed' - VLLM_TEST_SAME_HOST=1 VLLM_TEST_WITH_DEFAULT_DEVICE_SET=1 torchrun --nproc-per-node=4 distributed/test_same_node.py | grep 'Same node test passed' - CUDA_VISIBLE_DEVICES=0,1 pytest -v -s v1/shutdown - pytest -v -s v1/worker/test_worker_memory_snapshot.py - label: Distributed Model Tests (2 GPUs) # 19.3m timeout_in_minutes: 38 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_2 num_gpus: 2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/model_executor/model_loader/sharded_state_loader.py - vllm/model_executor/models/ - vllm/model_executor/layers/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py - tests/basic_correctness/ - tests/model_executor/model_loader/test_sharded_state_loader.py - tests/models/ commands: - TARGET_TEST_SUITE=L4 pytest basic_correctness/ -v -s -m 'distributed(num_gpus=2)' - CUDA_VISIBLE_DEVICES=0,1 pytest -v -s model_executor/model_loader/test_sharded_state_loader.py -m '(not slow_test)' - pytest models/test_transformers.py -v -s -m 'distributed(num_gpus=2)' - pytest models/language -v -s -m 'distributed(num_gpus=2)' - pytest models/multimodal -v -s -m 'distributed(num_gpus=2)' --ignore models/multimodal/generation/test_whisper.py - VLLM_WORKER_MULTIPROC_METHOD=spawn pytest models/multimodal/generation/test_whisper.py -v -s -m 'distributed(num_gpus=2)' - label: LoRA TP (Distributed) # 9.8m timeout_in_minutes: 18 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 num_gpus: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/lora - tests/lora - vllm/platforms/rocm.py commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - export PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True - pytest -v -s -x lora/test_chatglm3_tp.py - pytest -v -s -x lora/test_llama_tp.py - pytest -v -s -x lora/test_llm_with_multi_loras.py - pytest -v -s -x lora/test_olmoe_tp.py - pytest -v -s -x lora/test_gptoss_tp.py - pytest -v -s -x lora/test_qwen35_densemodel_lora.py - label: Weight Loading Multiple GPU # 7.5m timeout_in_minutes: 14 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_2 num_gpus: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/weight_loading commands: - bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models-amd.txt - label: Weight Loading Multiple GPU - Large Models # 12.6m timeout_in_minutes: 26 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_2 num_gpus: 2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/weight_loading commands: - bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models-large-amd.txt - label: Ray Dependency Compatibility Check # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 optional: true working_dir: "/" source_file_dependencies: - requirements/ - setup.py - vllm/platforms/rocm.py commands: - bash /vllm-workspace/.buildkite/scripts/check-ray-compatibility.sh - label: Distributed NixlConnector PD accuracy (4 GPUs) # 27.4m timeout_in_minutes: 44 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 num_gpus: 4 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/kv_transfer/kv_connector/v1/nixl_connector.py - tests/v1/kv_connector/nixl_integration/ - vllm/platforms/rocm.py commands: - uv pip install --system -r /vllm-workspace/requirements/kv_connectors_rocm.txt - ROCM_ATTN=1 bash v1/kv_connector/nixl_integration/config_sweep_accuracy_test.sh - label: CrossLayer KV layout Distributed NixlConnector PD accuracy tests (4 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 num_gpus: 4 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/kv_transfer/kv_connector/v1/nixl_connector.py - tests/v1/kv_connector/nixl_integration/ - vllm/platforms/rocm.py commands: - uv pip install --system -r /vllm-workspace/requirements/kv_connectors_rocm.txt - CROSS_LAYERS_BLOCKS=True ROCM_ATTN=1 bash v1/kv_connector/nixl_integration/config_sweep_accuracy_test.sh - label: Distributed Tests (4 GPUs)(A100-MI325) # 20.9m timeout_in_minutes: 37 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 num_gpus: 4 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - pytest -v -s distributed/test_custom_all_reduce.py - torchrun --nproc_per_node=2 distributed/test_ca_buffer_sharing.py - TARGET_TEST_SUITE=A100 pytest basic_correctness/ -v -s -m 'distributed(num_gpus=2)' - pytest -v -s -x lora/test_mixtral.py - label: Distributed Tests (2 GPUs)(H100-MI325) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_2 num_gpus: 2 working_dir: "/vllm-workspace/" source_file_dependencies: - vllm/distributed/ - vllm/v1/distributed/ - vllm/model_executor/layers/fused_moe/ - tests/v1/distributed/test_dbo.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - pytest -v -s tests/v1/distributed/test_dbo.py - label: Distributed Compile Unit Tests (2xH100-2xMI325) # 14.3m timeout_in_minutes: 32 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_2 num_gpus: 2 working_dir: "/vllm-workspace/" source_file_dependencies: - vllm/compilation/ - vllm/model_executor/layers - tests/compile/passes/distributed/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - export VLLM_TEST_CLEAN_GPU_MEMORY=1 - VLLM_TEST_CLEAN_GPU_MEMORY=1 pytest -v -s tests/compile/passes/distributed/test_async_tp.py - pytest -v -s tests/compile/passes/distributed/test_sequence_parallelism.py # TODO: this test is not supported on ROCm, there are aiter kernels for this. # - pytest -v -s tests/compile/passes/distributed/test_fusion_all_reduce.py # - pytest -v -s tests/compile/distributed/test_fusions_e2e.py::test_tp2_attn_quant_allreduce_rmsnorm # - "VLLM_TEST_CLEAN_GPU_MEMORY=1 pytest -v -s tests/compile/distributed/test_fusions_e2e.py -k 'not Llama-4'" - label: LM Eval Small Models # 13.3m timeout_in_minutes: 23 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/ - vllm/model_executor/layers/quantization - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=configs/models-small.txt - label: LM Eval Small Models (MI325) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 working_dir: "/vllm-workspace/.buildkite/lm-eval-harness" source_file_dependencies: - csrc/ - vllm/model_executor/layers/quantization - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - pytest -s -v test_lm_eval_correctness.py --config-list-file=configs/models-small-rocm.txt - label: LM Eval Small Models (B200-MI325) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/ - vllm/model_executor/layers/quantization - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=configs/models-mi3xx-fp8-and-mixed.txt - label: LM Eval Large Models (H200-MI325) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_8 optional: true num_gpus: 8 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/model_executor/layers/quantization/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/model_executor/layers/layernorm.py - csrc/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py - tests/evals/ commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=configs/models-mi3xx.txt - label: LM Eval Large Models (4 GPUs)(FP8) # 24.8m timeout_in_minutes: 42 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 num_gpus: 4 optional: true working_dir: "/vllm-workspace/.buildkite/lm-eval-harness" source_file_dependencies: - csrc/ - vllm/model_executor/layers/quantization - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - export VLLM_USE_DEEP_GEMM=0 - pytest -s -v test_lm_eval_correctness.py --config-list-file=configs/models-large-rocm-fp8.txt --tp-size=4 - label: LM Eval Large Models (4 GPUs)(A100-MI325) # 17.3m timeout_in_minutes: 27 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 num_gpus: 4 optional: true working_dir: "/vllm-workspace/.buildkite/lm-eval-harness" source_file_dependencies: - csrc/ - vllm/model_executor/layers/quantization - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -s -v test_lm_eval_correctness.py --config-list-file=configs/models-large.txt --tp-size=4 - label: ROCm LM Eval Large Models (8 Card) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_8 optional: true num_gpus: 8 working_dir: "/vllm-workspace/.buildkite/lm-eval-harness" source_file_dependencies: - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/model_executor/layers/quantization/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/model_executor/layers/layernorm.py - csrc/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -s -v test_lm_eval_correctness.py --config-list-file=configs/models-large-rocm.txt --tp-size=8 - label: GPQA Eval (GPT-OSS) (H100-MI325) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_2 num_gpus: 2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/ - vllm/model_executor/layers/quantization - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/model_executor/layers/fused_moe/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py - tests/evals/gpt_oss/ commands: - uv pip install --system 'gpt-oss[eval]==0.0.5' - pytest -s -v evals/gpt_oss/test_gpqa_correctness.py --config-list-file=configs/models-gfx942.txt - label: DeepSeek V2-Lite Accuracy # 6.7m timeout_in_minutes: 12 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 num_gpus: 4 optional: true working_dir: "/vllm-workspace" source_file_dependencies: - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/distributed/eplb - vllm/model_executor/layers/fused_moe/ - vllm/model_executor/layers/quantization/ - vllm/v1/attention/backends/ - vllm/v1/attention/backends/mla/ - vllm/v1/attention/selector.py - .buildkite/scripts/scheduled_integration_test/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - bash .buildkite/scripts/scheduled_integration_test/deepseek_v2_lite_ep_eplb.sh 0.25 200 8010 - label: DeepSeek V2-Lite Prefetch Offload Accuracy (H100-MI325) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 num_gpus: 1 working_dir: "/vllm-workspace" source_file_dependencies: - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/model_executor/layers/fused_moe/ - vllm/model_executor/layers/quantization/ - vllm/v1/attention/backends/ - vllm/v1/attention/backends/mla/ - vllm/v1/attention/selector.py - .buildkite/scripts/scheduled_integration_test/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - bash .buildkite/scripts/scheduled_integration_test/deepseek_v2_lite_prefetch_offload.sh 0.25 200 8030 - label: Qwen3-30B-A3B-FP8-block Accuracy # 6.4m timeout_in_minutes: 11 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 optional: true working_dir: "/vllm-workspace" source_file_dependencies: - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/model_executor/layers/quantization/ - vllm/distributed/eplb - vllm/model_executor/layers/fused_moe/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - .buildkite/scripts/scheduled_integration_test/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - bash .buildkite/scripts/scheduled_integration_test/qwen30b_a3b_fp8_block_ep_eplb.sh 0.8 200 8020 - label: Qwen3-Next-80B-A3B-Instruct MTP Async EPLB Accuracy # 10.9m timeout_in_minutes: 22 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_4 num_gpus: 4 optional: true working_dir: "/vllm-workspace" source_file_dependencies: - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/v1/spec_decode/ - vllm/distributed/eplb - vllm/model_executor/layers/fused_moe/ - vllm/model_executor/layers/quantization/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - .buildkite/scripts/scheduled_integration_test/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - bash .buildkite/scripts/scheduled_integration_test/qwen3_next_mtp_async_eplb.sh 0.8 1319 8040 ##### .buildkite/test_areas/compile.yaml ##### # Slowly setting up the tests so that it is also easier for the # CI team to review and upstream to the pipelinev2. # The following tests are important for vLLM IR Ops refactoring, # which affects fusion passes on ROCm. So we have to # enable them as as soon as possible. ## TODO: Enable the test in this group # # corresponds to .buildkite/test_areas/compile.yaml # - label: Fusion and Compile Unit Tests (2xB200-2xMI325) # TBD # timeout_in_minutes: 180 # mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325, tj] # agent_pool: mi325_1 # changed to 1 GPU until the fusion all reduce is enabled then only revert back to 2 GPUs # num_gpus: 1 # working_dir: "/vllm-workspace/" # source_file_dependencies: # - csrc/quantization/fp4/ # - vllm/model_executor/layers/quantization/ # - vllm/model_executor/layers/layernorm.py # - vllm/model_executor/layers/activation.py # - vllm/model_executor/layers/attention/attention.py # - vllm/v1/attention/backends/flashinfer.py # - vllm/compilation/ # TODO(luka) limit to vllm/compilation/passes # - tests/compile/test_fusion_attn.py # - tests/compile/test_silu_mul_quant_fusion.py # - tests/compile/distributed/test_fusion_all_reduce.py # - tests/compile/fullgraph/test_full_graph.py # commands: # - rocm-smi # # we run all backend tests on ROCm # # These two tests are covered in "PyTorch Compilation Passes Unit Tests" # # - "pytest -v -s tests/compile/passes/test_fusion_attn.py" # # - "pytest -v -s tests/compile/passes/test_silu_mul_quant_fusion.py" # # TODO: this test is not supported on ROCm, there are aiter kernels for this. # # - pytest -v -s tests/compile/passes/distributed/test_fusion_all_reduce.py # # TODO: find out more details # # - pytest -v -s tests/compile/fullgraph/test_full_graph.py::test_fp8_kv_scale_compile - label: Fusion E2E Quick (H100-MI325) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 num_gpus: 1 working_dir: "/vllm-workspace/" source_file_dependencies: - csrc/quantization/ - vllm/model_executor/ - vllm/v1/attention/ - vllm/compilation/ - tests/compile/fusions_e2e/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - rocm-smi # Run all models and attn backends but only Inductor partition and native custom ops - "pytest -v -s tests/compile/fusions_e2e/test_tp1_quant.py -k 'inductor_partition and not +rms_norm and not +quant_fp8'" # Different from CUDA, Qwen requires +rms_norm and +quant_fp8 as rms+quant fusion is only supported on AITER - "pytest -v -s tests/compile/fusions_e2e/test_tp1_quant.py -k 'inductor_partition and +rms_norm and +quant_fp8 and qwen3'" - label: Fusion E2E Config Sweep (H100-MI325) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi325] agent_pool: mi325_1 num_gpus: 1 working_dir: "/vllm-workspace/" source_file_dependencies: - csrc/quantization/ - vllm/compilation/ - vllm/model_executor/layers/layernorm.py - vllm/model_executor/layers/activation.py - vllm/model_executor/layers/attention/attention.py - vllm/model_executor/layers/quantization/input_quant_fp8.py - tests/compile/fusions_e2e/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - rocm-smi - pytest -v -s tests/compile/fusions_e2e/test_tp1_quant.py -k "llama-3" ## There are no ops on ROCm for these tests. ## The test still passes but the logs are not useful. ## fused ops just call torch.ops.symm_mem which ## exists in ROCm even though they don't work # - label: AsyncTP Correctness Tests (2xH100-2xMI325) # - label: Fusion E2E TP2 Quick (H100-MI325) # - label: Fusion E2E TP2 AsyncTP Config Sweep (H100-MI325) # - label: Fusion E2E TP2 (B200-MI325) # - label: Sequence Parallel Correctness Tests (2xH100-2xMI325) ##################################################################################################################################### # # # gfx950 # # # ##################################################################################################################################### - label: Entrypoints Integration (API Server openai - Part 1) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/openai - tests/entrypoints/test_chat_utils commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/openai/chat_completion --ignore=entrypoints/openai/chat_completion/test_chat_with_tool_reasoning.py --ignore=entrypoints/openai/chat_completion/test_oot_registration.py - label: Entrypoints Integration (API Server openai - Part 2) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/openai - tests/entrypoints/test_chat_utils commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/openai/completion --ignore=entrypoints/openai/completion/test_tensorizer_entrypoint.py - pytest -v -s entrypoints/openai/speech_to_text/ - pytest -v -s entrypoints/test_chat_utils.py - label: Entrypoints Integration (API Server openai - Part 3) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/openai - tests/entrypoints/test_chat_utils commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - 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 - label: Entrypoints Integration (API Server 2) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 optional: true fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/rpc - tests/entrypoints/serve/instrumentator - tests/tool_use commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/serve/instrumentator - PYTHONPATH=/vllm-workspace pytest -v -s entrypoints/rpc - pytest -v -s tool_use - label: Entrypoints Integration (Pooling) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 fast_check: true torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/entrypoints/pooling commands: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/pooling - label: Regression # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/test_regression commands: - pip install modelscope - pytest -v -s test_regression.py - label: V1 Spec Decode # TBD timeout_in_minutes: 60 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1/spec_decode commands: - pytest -v -s -m 'not slow_test' v1/spec_decode - label: V1 Sample + Logits # TBD timeout_in_minutes: 60 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1/sample - tests/v1/logits_processors - tests/v1/test_oracle.py - tests/v1/test_request.py - tests/v1/test_outputs.py commands: - pytest -v -s v1/sample - pytest -v -s v1/logits_processors - pytest -v -s v1/test_oracle.py - pytest -v -s v1/test_request.py - pytest -v -s v1/test_outputs.py - label: V1 Core + KV + Metrics # TBD timeout_in_minutes: 60 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/v1/core - tests/v1/executor - tests/v1/kv_offload - tests/v1/worker - tests/v1/kv_connector/unit - tests/v1/metrics - tests/entrypoints/openai/correctness/test_lmeval.py commands: - uv pip install --system -r /vllm-workspace/requirements/kv_connectors_rocm.txt - pytest -v -s -m 'not cpu_test' v1/core - pytest -v -s v1/executor - pytest -v -s v1/kv_offload - pytest -v -s v1/worker - pytest -v -s -m 'not cpu_test' v1/kv_connector/unit - pytest -v -s -m 'not cpu_test' v1/metrics - pip install -U git+https://github.com/robertgshaw2-redhat/lm-evaluation-harness.git@streaming-api - pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine - label: V1 Speculative Decoding (slow) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/v1/spec_decode/ - vllm/model_executor/models/ - vllm/v1/attention/ - vllm/model_executor/layers/ - tests/v1/spec_decode/ - vllm/platforms/rocm.py commands: - pytest -v -s -m 'slow_test' v1/spec_decode/test_eagle.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_extract_hidden_states.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_max_len.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_mtp.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_ngram.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_speculators_eagle3.py - pytest -v -s -m 'slow_test' v1/spec_decode/test_tree_attention.py - label: V1 attention (B200-MI355) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/config/attention.py - vllm/model_executor/layers/attention - vllm/v1/attention - tests/v1/attention - vllm/_aiter_ops.py - vllm/envs.py - vllm/platforms/rocm.py commands: - pytest -v -s v1/attention - label: Examples # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 working_dir: "/vllm-workspace/examples" source_file_dependencies: - vllm/entrypoints - vllm/multimodal - examples/ - vllm/platforms/rocm.py commands: - pip install tensorizer # Basic - python3 basic/offline_inference/chat.py --attention-backend TRITON_ATTN - 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 - python3 basic/offline_inference/classify.py - python3 basic/offline_inference/embed.py - python3 basic/offline_inference/score.py # Multi-modal models - python3 offline_inference/audio_language.py --seed 0 - python3 offline_inference/vision_language.py --seed 0 - python3 offline_inference/vision_language_multi_image.py --seed 0 - python3 offline_inference/encoder_decoder_multimodal.py --model-type whisper --seed 0 # Pooling models - python3 pooling/embed/vision_embedding_offline.py --seed 0 # Features demo - python3 offline_inference/prefix_caching.py - python3 offline_inference/llm_engine_example.py - python3 others/tensorize_vllm_model.py --model facebook/opt-125m serialize --serialized-directory /tmp/ --suffix v1 && python3 others/tensorize_vllm_model.py --model facebook/opt-125m deserialize --path-to-tensors /tmp/vllm/facebook/opt-125m/v1/model.tensors - python3 offline_inference/spec_decode.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 - python3 offline_inference/spec_decode.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: Kernels Attention Test %N # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 parallelism: 2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/attention/ - vllm/v1/attention - vllm/model_executor/layers/attention - tests/kernels/attention - vllm/_aiter_ops.py - vllm/envs.py - vllm/platforms/rocm.py commands: - pytest -v -s kernels/attention --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT - label: Kernels Quantization Test %N # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 parallelism: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/quantization/ - vllm/model_executor/layers/quantization - tests/kernels/quantization - tests/kernels/quantization/test_rocm_skinny_gemms.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py - vllm/model_executor/kernels/ commands: - pytest -v -s kernels/quantization --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT - label: Kernels MoE Test %N # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 parallelism: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/quantization/cutlass_w8a8/moe/ - csrc/moe/ - tests/kernels/moe - vllm/model_executor/layers/fused_moe/ - vllm/distributed/device_communicators/ - vllm/envs.py - vllm/config - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - pytest -v -s kernels/moe --ignore=kernels/moe/test_modular_oai_triton_moe.py --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT - pytest -v -s kernels/moe/test_modular_oai_triton_moe.py --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT - label: Kernels FP8 MoE Test # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/moe/ - csrc/quantization/w8a8/cutlass/moe/ - vllm/model_executor/layers/fused_moe/ - tests/kernels/moe/test_deepep_moe.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py - vllm/envs.py commands: - pytest -v -s kernels/moe/test_deepep_moe.py - label: Quantization # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/ - vllm/model_executor/layers/quantization - tests/quantization - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - uv pip install --system torchao==0.17.0 - uv pip install --system conch-triton-kernels - VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization/ --ignore quantization/test_blackwell_moe.py - label: Language Models Tests (Standard) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 torch_nightly: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/language commands: - pip freeze | grep -E 'torch' - pytest -v -s models/language -m 'core_model and (not slow_test)' - label: Language Models Test (Extended Generation) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/language/generation commands: - uv pip install --system --no-build-isolation 'git+https://github.com/AndreasKaratzas/mamba@fix-rocm-7.0-warp-size-constexpr' - uv pip install --system --no-build-isolation 'git+https://github.com/Dao-AILab/causal-conv1d@v1.5.2' - pytest -v -s models/language/generation -m '(not core_model) and (not hybrid_model)' - label: Language Models Test (Extended Pooling) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/language/pooling commands: - pytest -v -s models/language/pooling -m 'not core_model' - label: "Multi-Modal Models (Standard) 1: qwen2" # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 torch_nightly: true optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "qwen2" - pytest -v -s models/multimodal/generation/test_ultravox.py -m core_model - label: "Multi-Modal Models (Standard) 2: qwen3 + gemma" # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 torch_nightly: true optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "qwen3 or gemma" - pytest -v -s models/multimodal/generation/test_qwen2_5_vl.py -m core_model - label: "Multi-Modal Models (Standard) 3: llava + qwen2_vl" # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 torch_nightly: true optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation - tests/models/multimodal/test_mapping.py commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m core_model -k "not qwen2 and not qwen3 and not gemma" - pytest -v -s models/multimodal/generation/test_qwen2_vl.py -m core_model - label: "Multi-Modal Models (Standard) 4: other + whisper" # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 torch_nightly: true optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - 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/processing - cd .. && VLLM_WORKER_MULTIPROC_METHOD=spawn pytest -v -s tests/models/multimodal/generation/test_whisper.py -m core_model - label: Multi-Modal Models (Extended Generation 1) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation - tests/models/multimodal/test_mapping.py commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation -m 'not core_model' --ignore models/multimodal/generation/test_common.py - pytest -v -s models/multimodal/test_mapping.py - label: Multi-Modal Models (Extended Generation 2) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m 'split(group=0) and not core_model' - label: Multi-Modal Models (Extended Generation 3) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/generation commands: - pip install git+https://github.com/TIGER-AI-Lab/Mantis.git - pytest -v -s models/multimodal/generation/test_common.py -m 'split(group=1) and not core_model' - label: Multi-Modal Models (Extended Pooling) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/models/multimodal/pooling commands: - pytest -v -s models/multimodal/pooling -m 'not core_model' - label: Quantized Models Test # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/model_executor/layers/quantization - tests/models/quantization - vllm/_aiter_ops.py - vllm/platforms/rocm.py - vllm/model_executor/model_loader/ commands: - pytest -v -s models/quantization - label: Kernels (B200-MI355) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 working_dir: "/vllm-workspace/" source_file_dependencies: - csrc/quantization/fp4/ - csrc/attention/mla/ - csrc/quantization/cutlass_w8a8/moe/ - vllm/model_executor/layers/fused_moe/cutlass_moe.py - vllm/v1/attention/backends/triton_attn.py - vllm/v1/attention/backends/rocm_attn.py - vllm/v1/attention/backends/rocm_aiter_fa.py - vllm/v1/attention/backends/rocm_aiter_unified_attn.py - vllm/v1/attention/backends/mla/aiter_triton_mla.py - vllm/v1/attention/backends/mla/rocm_aiter_mla.py - vllm/v1/attention/selector.py - vllm/platforms/rocm.py - vllm/_aiter_ops.py commands: - rocm-smi - python3 examples/basic/offline_inference/chat.py - pytest -v -s tests/kernels/attention/test_attention_selector.py - label: Weight Loading Multiple GPU # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_2 num_gpus: 2 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/ - tests/weight_loading commands: - bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models-amd.txt - label: Weight Loading Multiple GPU - Large Models # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_2 working_dir: "/vllm-workspace/tests" num_gpus: 2 optional: true source_file_dependencies: - vllm/ - tests/weight_loading commands: - bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models-large-amd.txt - label: Ray Dependency Compatibility Check # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_1 optional: true working_dir: "/" source_file_dependencies: - requirements/ - setup.py - vllm/platforms/rocm.py commands: - bash /vllm-workspace/.buildkite/scripts/check-ray-compatibility.sh - label: Distributed NixlConnector PD accuracy (4 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_4 num_gpus: 4 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/kv_transfer/kv_connector/v1/nixl_connector.py - tests/v1/kv_connector/nixl_integration/ - vllm/platforms/rocm.py commands: - uv pip install --system -r /vllm-workspace/requirements/kv_connectors_rocm.txt - ROCM_ATTN=1 bash v1/kv_connector/nixl_integration/config_sweep_accuracy_test.sh - label: DP EP Distributed NixlConnector PD accuracy tests (4 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_4 num_gpus: 4 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/kv_transfer/kv_connector/v1/nixl_connector.py - tests/v1/kv_connector/nixl_integration/ - vllm/platforms/rocm.py commands: - uv pip install --system -r /vllm-workspace/requirements/kv_connectors_rocm.txt - DP_EP=1 ROCM_ATTN=1 bash v1/kv_connector/nixl_integration/config_sweep_accuracy_test.sh - label: NixlConnector PD + Spec Decode acceptance (2 GPUs) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_2 num_gpus: 2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/distributed/kv_transfer/kv_connector/v1/nixl_connector.py - vllm/v1/worker/kv_connector_model_runner_mixin.py - tests/v1/kv_connector/nixl_integration/ - vllm/platforms/rocm.py commands: - uv pip install --system -r /vllm-workspace/requirements/kv_connectors_rocm.txt - ROCM_ATTN=1 bash v1/kv_connector/nixl_integration/spec_decode_acceptance_test.sh - label: Distributed Tests (2 GPUs)(H100-MI355) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_2 num_gpus: 2 optional: true working_dir: "/vllm-workspace/" source_file_dependencies: - vllm/distributed/ - vllm/v1/distributed/ - vllm/model_executor/layers/fused_moe/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - tests/distributed/test_context_parallel.py - tests/v1/distributed/test_dbo.py - examples/offline_inference/data_parallel.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - export TORCH_NCCL_BLOCKING_WAIT=1 - pytest -v -s tests/distributed/test_context_parallel.py - VLLM_LOGGING_LEVEL=DEBUG python3 examples/offline_inference/data_parallel.py --model=Qwen/Qwen1.5-MoE-A2.7B -tp=1 -dp=2 --max-model-len=2048 --all2all-backend=deepep_high_throughput - pytest -v -s tests/v1/distributed/test_dbo.py - label: Distributed Compile Unit Tests (2xH100-2xMI355) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_2 num_gpus: 2 optional: true working_dir: "/vllm-workspace/" source_file_dependencies: - vllm/compilation/ - vllm/model_executor/layers - tests/compile/passes/distributed/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - export VLLM_TEST_CLEAN_GPU_MEMORY=1 - VLLM_TEST_CLEAN_GPU_MEMORY=1 pytest -v -s tests/compile/passes/distributed/test_async_tp.py - pytest -v -s tests/compile/passes/distributed/test_sequence_parallelism.py # TODO: this test is not supported on ROCm, there are aiter kernels for this. # - pytest -v -s tests/compile/passes/distributed/test_fusion_all_reduce.py # - pytest -v -s tests/compile/distributed/test_fusions_e2e.py::test_tp2_attn_quant_allreduce_rmsnorm # - "VLLM_TEST_CLEAN_GPU_MEMORY=1 pytest -v -s tests/compile/distributed/test_fusions_e2e.py -k 'not Llama-4'" - label: LM Eval Small Models (B200-MI355) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/ - vllm/model_executor/layers/quantization - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=configs/models-mi3xx-fp8-and-mixed.txt - label: LM Eval Large Models (4 GPUs)(FP8) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_4 num_gpus: 4 optional: true working_dir: "/vllm-workspace/.buildkite/lm-eval-harness" source_file_dependencies: - csrc/ - vllm/model_executor/layers/quantization - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - export VLLM_USE_DEEP_GEMM=0 - pytest -s -v test_lm_eval_correctness.py --config-list-file=configs/models-large-rocm-fp8.txt --tp-size=4 - label: GPQA Eval (GPT-OSS) (B200-MI355) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx955nightly, amdmi355] agent_pool: mi355_2 num_gpus: 2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - csrc/ - vllm/model_executor/layers/quantization - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - vllm/model_executor/layers/fused_moe/ - tests/evals/gpt_oss/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - uv pip install --system 'gpt-oss[eval]==0.0.5' - pytest -s -v evals/gpt_oss/test_gpqa_correctness.py --config-list-file=configs/models-gfx950.txt - label: Qwen3-30B-A3B-FP8-block Accuracy (B200-MI355) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_2 num_gpus: 2 working_dir: "/vllm-workspace" source_file_dependencies: - vllm/model_executor/models/ - vllm/model_executor/model_loader/ - vllm/model_executor/layers/quantization/ - vllm/model_executor/layers/fused_moe/ - vllm/distributed/eplb - vllm/v1/attention/backends/ - vllm/v1/attention/selector.py - .buildkite/scripts/scheduled_integration_test/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - bash .buildkite/scripts/scheduled_integration_test/qwen30b_a3b_fp8_block_ep_eplb.sh 0.8 200 8020 2 1 - label: Attention Benchmarks Smoke Test (B200-MI355) # TBD timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_2 num_gpus: 2 working_dir: "/vllm-workspace/" source_file_dependencies: - benchmarks/attention_benchmarks/ - vllm/v1/attention/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - python3 benchmarks/attention_benchmarks/benchmark.py --backends ROCM_ATTN ROCM_AITER_FA ROCM_AITER_UNIFIED_ATTN --batch-specs "8q1s1k" --repeats 1 --warmup-iters 1 - label: LM Eval Qwen3-5 Models (B200-MI355) # TBD timeout_in_minutes: 120 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_2 num_gpus: 2 optional: true working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/model_executor/models/qwen3_5.py - vllm/model_executor/models/qwen3_5_mtp.py - vllm/transformers_utils/configs/qwen3_5.py - vllm/transformers_utils/configs/qwen3_5_moe.py - vllm/model_executor/models/qwen.py - vllm/model_executor/models/qwen2.py - vllm/model_executor/models/qwen3.py - vllm/model_executor/models/qwen3_next.py - vllm/model_executor/models/qwen3_next_mtp.py - vllm/model_executor/layers/fla/ops/ - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=configs/models-qwen35-mi355.txt