From 78fa18910ee2fe64b2624e9358f622e1b50b14bf Mon Sep 17 00:00:00 2001 From: "Kevin H. Luu" Date: Sun, 30 Aug 2026 04:35:01 -0700 Subject: [PATCH] ci: add MIG slice size to H200 job labels (#54420) Signed-off-by: khluu --- .buildkite/test_areas/basic_correctness.yaml | 2 +- .buildkite/test_areas/benchmarks.yaml | 2 +- .buildkite/test_areas/cuda.yaml | 4 ++-- .buildkite/test_areas/engine.yaml | 6 ++--- .buildkite/test_areas/entrypoints.yaml | 22 +++++++++---------- .buildkite/test_areas/expert_parallelism.yaml | 2 +- .buildkite/test_areas/jit_monitor.yaml | 2 +- .buildkite/test_areas/kernels.yaml | 6 ++--- .buildkite/test_areas/lm_eval.yaml | 8 +++---- .buildkite/test_areas/lora.yaml | 2 +- .buildkite/test_areas/misc.yaml | 16 +++++++------- .buildkite/test_areas/model_executor.yaml | 2 +- .buildkite/test_areas/model_runner_v2.yaml | 6 ++--- .buildkite/test_areas/models_basic.yaml | 6 ++--- .buildkite/test_areas/models_language.yaml | 14 ++++++------ .buildkite/test_areas/models_multimodal.yaml | 22 +++++++++---------- .buildkite/test_areas/plugins.yaml | 4 ++-- .buildkite/test_areas/pytorch.yaml | 8 +++---- .buildkite/test_areas/quantization.yaml | 6 ++--- .buildkite/test_areas/ray_compat.yaml | 2 +- .buildkite/test_areas/rust_frontend.yaml | 8 +++---- .buildkite/test_areas/samplers.yaml | 2 +- .buildkite/test_areas/spec_decode.yaml | 18 +++++++-------- 23 files changed, 85 insertions(+), 85 deletions(-) diff --git a/.buildkite/test_areas/basic_correctness.yaml b/.buildkite/test_areas/basic_correctness.yaml index 1ce9dcbe88..43e3113eb3 100644 --- a/.buildkite/test_areas/basic_correctness.yaml +++ b/.buildkite/test_areas/basic_correctness.yaml @@ -2,7 +2,7 @@ group: Basic Correctness depends_on: - image-build steps: -- label: ":nvidia: (H200) Basic Correctness" +- label: ":nvidia: (H200 MIG 18GB) Basic Correctness" key: basic-correctness timeout_in_minutes: 68 device: h200_18gb diff --git a/.buildkite/test_areas/benchmarks.yaml b/.buildkite/test_areas/benchmarks.yaml index 94d9999b81..45a9e02fc8 100644 --- a/.buildkite/test_areas/benchmarks.yaml +++ b/.buildkite/test_areas/benchmarks.yaml @@ -2,7 +2,7 @@ group: Benchmarks depends_on: - image-build steps: -- label: ":nvidia: (H200) Benchmarks CLI" +- label: ":nvidia: (H200 MIG 18GB) Benchmarks CLI" key: benchmarks-cli-test timeout_in_minutes: 45 device: h200_18gb diff --git a/.buildkite/test_areas/cuda.yaml b/.buildkite/test_areas/cuda.yaml index a491eed274..f0b1f4af2b 100644 --- a/.buildkite/test_areas/cuda.yaml +++ b/.buildkite/test_areas/cuda.yaml @@ -2,7 +2,7 @@ group: CUDA depends_on: - image-build steps: -- label: ":nvidia: (H200) CUDA Platform" +- label: ":nvidia: (H200 MIG 18GB) CUDA Platform" key: platform-tests timeout_in_minutes: 20 device: h200_18gb @@ -30,7 +30,7 @@ steps: - vllm/ - tests/cuda -- label: ":nvidia: (H200) CUDAGraph" +- label: ":nvidia: (H200 MIG 35GB) CUDAGraph" device: h200_35gb key: cudagraph timeout_in_minutes: 30 diff --git a/.buildkite/test_areas/engine.yaml b/.buildkite/test_areas/engine.yaml index 6bd7f92ef1..d4a347fae8 100644 --- a/.buildkite/test_areas/engine.yaml +++ b/.buildkite/test_areas/engine.yaml @@ -2,7 +2,7 @@ group: Engine depends_on: - image-build steps: -- label: ":nvidia: (H200) Engine Core" +- label: ":nvidia: (H200 MIG 18GB) Engine Core" key: engine timeout_in_minutes: 30 device: h200_18gb @@ -55,7 +55,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) E2E Scheduling" +- label: ":nvidia: (H200 MIG 18GB) E2E Scheduling" key: e2e-scheduling-1-gpu timeout_in_minutes: 53 device: h200_18gb @@ -72,7 +72,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) E2E Core" +- label: ":nvidia: (H200 MIG 35GB) E2E Core" device: h200_35gb key: e2e-core-1-gpu timeout_in_minutes: 40 diff --git a/.buildkite/test_areas/entrypoints.yaml b/.buildkite/test_areas/entrypoints.yaml index 578a821c59..cabba26c26 100644 --- a/.buildkite/test_areas/entrypoints.yaml +++ b/.buildkite/test_areas/entrypoints.yaml @@ -2,7 +2,7 @@ group: Entrypoints depends_on: - image-build steps: -- label: ":nvidia: (H200) Entrypoints Unit" +- label: ":nvidia: (H200 MIG 35GB) Entrypoints Unit" device: h200_35gb key: entrypoints-unit-tests timeout_in_minutes: 40 @@ -25,7 +25,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Entrypoints Integration (LLM)" +- label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (LLM)" device: h200_35gb key: entrypoints-integration-llm timeout_in_minutes: 60 @@ -48,7 +48,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Entrypoints Integration (API Server)" +- label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (API Server)" key: entrypoints-integration-api-server device: h200_35gb timeout_in_minutes: 75 @@ -72,7 +72,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Entrypoints Integration (API Server OpenAI - Part 1)" +- label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (API Server OpenAI - Part 1)" device: h200_35gb key: entrypoints-integration-api-server-openai-part-1 timeout_in_minutes: 68 @@ -94,7 +94,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Entrypoints Integration (API Server OpenAI - Part 2)" +- label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (API Server OpenAI - Part 2)" device: h200_35gb key: entrypoints-integration-api-server-openai-part-2 timeout_in_minutes: 83 @@ -117,7 +117,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Entrypoints Integration (API Server Generate)" +- label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (API Server Generate)" device: h200_35gb key: entrypoints-integration-api-server-generate timeout_in_minutes: 50 @@ -143,7 +143,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Entrypoints Integration (Responses API)" +- label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (Responses API)" device: h200_35gb key: entrypoints-integration-responses-api timeout_in_minutes: 50 @@ -163,7 +163,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Entrypoints Integration (Speech to Text)" +- label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (Speech to Text)" device: h200_35gb key: entrypoints-integration-speech_to_text timeout_in_minutes: 45 @@ -176,7 +176,7 @@ steps: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/speech_to_text -- label: ":nvidia: (H200) Entrypoints Integration (Multimodal)" +- label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (Multimodal)" device: h200_35gb key: entrypoints-integration-multimodal timeout_in_minutes: 45 @@ -197,7 +197,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Entrypoints Integration (Pooling)" +- label: ":nvidia: (H200 MIG 35GB) Entrypoints Integration (Pooling)" device: h200_35gb key: entrypoints-integration-pooling timeout_in_minutes: 75 @@ -210,7 +210,7 @@ steps: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s entrypoints/pooling -- label: ":nvidia: (H200) OpenAI API Correctness" +- label: ":nvidia: (H200 MIG 18GB) OpenAI API Correctness" key: openai-api-correctness timeout_in_minutes: 20 device: h200_18gb diff --git a/.buildkite/test_areas/expert_parallelism.yaml b/.buildkite/test_areas/expert_parallelism.yaml index e0b931f5fe..0333b28cee 100644 --- a/.buildkite/test_areas/expert_parallelism.yaml +++ b/.buildkite/test_areas/expert_parallelism.yaml @@ -2,7 +2,7 @@ group: Expert Parallelism depends_on: - image-build steps: -- label: ":nvidia: (H200) EPLB Algorithm" +- label: ":nvidia: (H200 MIG 18GB) EPLB Algorithm" key: eplb-algorithm timeout_in_minutes: 20 device: h200_18gb diff --git a/.buildkite/test_areas/jit_monitor.yaml b/.buildkite/test_areas/jit_monitor.yaml index 941c190138..3355470c17 100644 --- a/.buildkite/test_areas/jit_monitor.yaml +++ b/.buildkite/test_areas/jit_monitor.yaml @@ -2,7 +2,7 @@ group: JIT Monitor depends_on: - image-build steps: -- label: ":nvidia: (H200) No Runtime JITs E2E" +- label: ":nvidia: (H200 MIG 35GB) No Runtime JITs E2E" key: jit-monitor-no-runtime-jit device: h200_35gb timeout_in_minutes: 45 diff --git a/.buildkite/test_areas/kernels.yaml b/.buildkite/test_areas/kernels.yaml index 49754ad81b..ae3d5da878 100644 --- a/.buildkite/test_areas/kernels.yaml +++ b/.buildkite/test_areas/kernels.yaml @@ -2,7 +2,7 @@ group: Kernels depends_on: - image-build steps: -- label: ":nvidia: (H200) vLLM IR" +- label: ":nvidia: (H200 MIG 18GB) vLLM IR" key: vllm-ir-tests timeout_in_minutes: 35 device: h200_18gb @@ -14,7 +14,7 @@ steps: - pytest -v -s tests/ir - pytest -v -s tests/kernels/ir -- label: ":nvidia: (H200) Core Operation Kernels Shard %N" +- label: ":nvidia: (H200 MIG 35GB) Core Operation Kernels Shard %N" device: h200_35gb key: kernels-core-operation-test timeout_in_minutes: 120 @@ -225,7 +225,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Mamba Kernels" +- label: ":nvidia: (H200 MIG 35GB) Mamba Kernels" device: h200_35gb key: kernels-mamba-test timeout_in_minutes: 60 diff --git a/.buildkite/test_areas/lm_eval.yaml b/.buildkite/test_areas/lm_eval.yaml index bb659bb749..d52eccf1ee 100644 --- a/.buildkite/test_areas/lm_eval.yaml +++ b/.buildkite/test_areas/lm_eval.yaml @@ -2,7 +2,7 @@ group: LM Eval depends_on: - image-build steps: -- label: ":nvidia: (H200) LM Eval Small Models" +- label: ":nvidia: (H200 MIG 35GB) LM Eval Small Models" device: h200_35gb key: lm-eval-small-models timeout_in_minutes: 45 @@ -291,7 +291,7 @@ steps: - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/humming/config-act-fp8.txt - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/humming/config-act-int8.txt -- label: ":nvidia: (H200) LM Eval TurboQuant KV Cache" +- label: ":nvidia: (H200 MIG 18GB) LM Eval TurboQuant KV Cache" key: lm-eval-turboquant-kv-cache timeout_in_minutes: 55 device: h200_18gb @@ -347,7 +347,7 @@ steps: - 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-spark.txt -- label: ":nvidia: (H200) LM Eval KV-Offload" +- label: ":nvidia: (H200 MIG 35GB) LM Eval KV-Offload" key: kv-offload-small timeout_in_minutes: 30 device: h200_35gb @@ -388,7 +388,7 @@ steps: commands: - pytest -s -v evals/gsm8k/test_gsm8k_offloading.py -k "deepseek-v4-flash" -- label: ":nvidia: (H200) MRCR Eval Small Models" +- label: ":nvidia: (H200 MIG 35GB) MRCR Eval Small Models" device: h200_35gb timeout_in_minutes: 25 source_file_dependencies: diff --git a/.buildkite/test_areas/lora.yaml b/.buildkite/test_areas/lora.yaml index 55a7f54d82..60e106c20c 100644 --- a/.buildkite/test_areas/lora.yaml +++ b/.buildkite/test_areas/lora.yaml @@ -2,7 +2,7 @@ group: LoRA depends_on: - image-build steps: -- label: ":nvidia: (H200) LoRA Shard %N" +- label: ":nvidia: (H200 MIG 35GB) LoRA Shard %N" device: h200_35gb key: lora timeout_in_minutes: 40 diff --git a/.buildkite/test_areas/misc.yaml b/.buildkite/test_areas/misc.yaml index 1d952b028c..3e0f2b9682 100644 --- a/.buildkite/test_areas/misc.yaml +++ b/.buildkite/test_areas/misc.yaml @@ -2,7 +2,7 @@ group: Miscellaneous depends_on: - image-build steps: -- label: ":nvidia: (H200) V1 Sample + Logits" +- label: ":nvidia: (H200 MIG 18GB) V1 Sample + Logits" key: v1-sample-logits timeout_in_minutes: 83 device: h200_18gb @@ -40,7 +40,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) V1 Core + KV + Metrics" +- label: ":nvidia: (H200 MIG 35GB) V1 Core + KV + Metrics" device: h200_35gb key: v1-core-kv-metrics timeout_in_minutes: 80 @@ -132,7 +132,7 @@ steps: - pytest -v -s -m 'cpu_test' v1/ec_connector/unit - pytest -v -s -m 'cpu_test' v1/metrics -- label: ":nvidia: (H200) Extract Hidden States Integration" +- label: ":nvidia: (H200 MIG 18GB) Extract Hidden States Integration" key: extract-hidden-states-integration timeout_in_minutes: 20 device: h200_18gb @@ -162,7 +162,7 @@ steps: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -v -s -m 'distributed' tests/v1/kv_connector/extract_hidden_states_integration -- label: ":nvidia: (H200) Regression" +- label: ":nvidia: (H200 MIG 18GB) Regression" key: regression timeout_in_minutes: 30 device: h200_18gb @@ -185,7 +185,7 @@ steps: - pytest -v -s test_regression.py working_dir: "/vllm-workspace/tests" # optional -- label: ":nvidia: (H200) Examples" +- label: ":nvidia: (H200 MIG 35GB) Examples" device: h200_35gb key: examples timeout_in_minutes: 40 @@ -268,7 +268,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Python-only Installation" +- label: ":nvidia: (H200 MIG 18GB) Python-only Installation" key: python-only-installation depends_on: ~ optional: true @@ -293,7 +293,7 @@ steps: - setup.py - vllm/platforms/rocm.py -- label: ":nvidia: (H200) Async Engine, Inputs, Utils, Worker" +- label: ":nvidia: (H200 MIG 35GB) Async Engine, Inputs, Utils, Worker" device: h200_35gb key: async-engine-inputs-utils-worker timeout_in_minutes: 25 @@ -440,7 +440,7 @@ steps: - pytest -v -s v1/determinism/test_cutlass_batch_invariance.py - pytest -v -s v1/determinism/test_online_batch_invariance.py -- label: ":nvidia: (H200) Acceptance Length (Large Models)" +- label: ":nvidia: (H200 MIG 35GB) Acceptance Length (Large Models)" device: h200_35gb key: acceptance-length-test-large-models timeout_in_minutes: 20 diff --git a/.buildkite/test_areas/model_executor.yaml b/.buildkite/test_areas/model_executor.yaml index 8228bbc83a..1e92462fe6 100644 --- a/.buildkite/test_areas/model_executor.yaml +++ b/.buildkite/test_areas/model_executor.yaml @@ -2,7 +2,7 @@ group: Model Executor depends_on: - image-build steps: -- label: ":nvidia: (H200) Model Executor" +- label: ":nvidia: (H200 MIG 35GB) Model Executor" device: h200_35gb key: model-executor timeout_in_minutes: 60 diff --git a/.buildkite/test_areas/model_runner_v2.yaml b/.buildkite/test_areas/model_runner_v2.yaml index e6d93d330b..1d4b6e4f71 100644 --- a/.buildkite/test_areas/model_runner_v2.yaml +++ b/.buildkite/test_areas/model_runner_v2.yaml @@ -2,7 +2,7 @@ group: Model Runner V2 depends_on: - image-build steps: -- label: ":nvidia: (H200) Model Runner V2 Core" +- label: ":nvidia: (H200 MIG 35GB) Model Runner V2 Core" device: h200_35gb key: model-runner-v2-core-tests timeout_in_minutes: 35 @@ -24,7 +24,7 @@ steps: # Temporary hack filter to exclude ngram spec decoding based tests. - pytest -v -s entrypoints/llm/test_struct_output_generate.py -k "xgrammar and not speculative_config6 and not speculative_config7 and not speculative_config8 and not speculative_config0" -- label: ":nvidia: (H200) Model Runner V2 Examples" +- label: ":nvidia: (H200 MIG 35GB) Model Runner V2 Examples" device: h200_35gb key: model-runner-v2-examples timeout_in_minutes: 35 @@ -126,7 +126,7 @@ steps: - pytest -v -s distributed/test_pp_cudagraph.py -k "not ray" - pytest -v -s v1/distributed/test_pp_dp_v2.py -- label: ":nvidia: (H200) Model Runner V2 Spec Decode" +- label: ":nvidia: (H200 MIG 35GB) Model Runner V2 Spec Decode" device: h200_35gb key: model-runner-v2-spec-decode timeout_in_minutes: 50 diff --git a/.buildkite/test_areas/models_basic.yaml b/.buildkite/test_areas/models_basic.yaml index f6d0915564..780eef68ce 100644 --- a/.buildkite/test_areas/models_basic.yaml +++ b/.buildkite/test_areas/models_basic.yaml @@ -2,7 +2,7 @@ group: Models - Basic depends_on: - image-build steps: -- label: ":nvidia: (H200) Basic Models (Initialization)" +- label: ":nvidia: (H200 MIG 18GB) Basic Models (Initialization)" key: basic-models-tests-initialization timeout_in_minutes: 25 device: h200_18gb @@ -23,7 +23,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Basic Models (Extra Initialization) Shard %N" +- label: ":nvidia: (H200 MIG 35GB) Basic Models (Extra Initialization) Shard %N" device: h200_35gb key: basic-models-tests-extra-initialization timeout_in_minutes: 100 @@ -38,7 +38,7 @@ steps: - 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 parallelism: 4 -- label: ":nvidia: (H200) Basic Models (Other)" +- label: ":nvidia: (H200 MIG 35GB) Basic Models (Other)" device: h200_35gb key: basic-models-tests-other timeout_in_minutes: 35 diff --git a/.buildkite/test_areas/models_language.yaml b/.buildkite/test_areas/models_language.yaml index be2e36c301..c55f733fd9 100644 --- a/.buildkite/test_areas/models_language.yaml +++ b/.buildkite/test_areas/models_language.yaml @@ -2,7 +2,7 @@ group: Models - Language depends_on: - image-build steps: -- label: ":nvidia: (H200) Language Models (Standard)" +- label: ":nvidia: (H200 MIG 18GB) Language Models (Standard)" key: language-models-tests-standard timeout_in_minutes: 30 device: h200_18gb @@ -23,7 +23,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Language Models (Extra Standard) Shard %N" +- label: ":nvidia: (H200 MIG 35GB) Language Models (Extra Standard) Shard %N" device: h200_35gb key: language-models-tests-extra-standard timeout_in_minutes: 40 @@ -57,7 +57,7 @@ steps: - tests/models/language/pooling/test_classification.py - vllm/_aiter_ops.py - vllm/platforms/rocm.py -- label: ":nvidia: (H200) Language Models (Hybrid) Shard %N" +- label: ":nvidia: (H200 MIG 35GB) Language Models (Hybrid) Shard %N" device: h200_35gb key: language-models-tests-hybrid timeout_in_minutes: 65 @@ -104,7 +104,7 @@ steps: - CAUSAL_CONV1D_FORCE_BUILD=TRUE uv pip install --system --no-build-isolation 'git+https://github.com/Dao-AILab/causal-conv1d@v1.6.0' - pytest -v -s models/language/generation -m hybrid_model -k 'granite-4.0-tiny-preview' -- label: ":nvidia: (H200) Language Models (Extended Generation)" +- label: ":nvidia: (H200 MIG 35GB) Language Models (Extended Generation)" device: h200_35gb key: language-models-test-extended-generation timeout_in_minutes: 80 @@ -133,7 +133,7 @@ steps: - CAUSAL_CONV1D_FORCE_BUILD=TRUE uv pip install --system --no-build-isolation 'git+https://github.com/Dao-AILab/causal-conv1d@v1.6.0' - pytest -v -s models/language/generation -m '(not core_model) and (not hybrid_model)' -- label: ":nvidia: (H200) Language Models (PPL)" +- label: ":nvidia: (H200 MIG 18GB) Language Models (PPL)" key: language-models-test-ppl timeout_in_minutes: 30 device: h200_18gb @@ -171,7 +171,7 @@ steps: - vllm/platforms/rocm.py - tests/models/language/generation_ppl_test -- label: ":nvidia: (H200) Language Models (Extended Pooling) Shard %N" +- label: ":nvidia: (H200 MIG 35GB) Language Models (Extended Pooling) Shard %N" device: h200_35gb key: language-models-test-extended-pooling timeout_in_minutes: 120 @@ -195,7 +195,7 @@ steps: commands: - pytest -v -s models/language/pooling -m 'not core_model' --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --shard-id=$$BUILDKITE_PARALLEL_JOB -- label: ":nvidia: (H200) Language Models (MTEB)" +- label: ":nvidia: (H200 MIG 18GB) Language Models (MTEB)" key: language-models-test-mteb timeout_in_minutes: 68 device: h200_18gb diff --git a/.buildkite/test_areas/models_multimodal.yaml b/.buildkite/test_areas/models_multimodal.yaml index 340c97527c..4c0c45139e 100644 --- a/.buildkite/test_areas/models_multimodal.yaml +++ b/.buildkite/test_areas/models_multimodal.yaml @@ -2,7 +2,7 @@ group: Models - Multimodal depends_on: - image-build steps: -- label: ":nvidia: (H200) Multimodal Models (Standard) 1: qwen2" +- label: ":nvidia: (H200 MIG 18GB) Multimodal Models (Standard) 1: qwen2" key: multi-modal-models-standard-1-qwen2 timeout_in_minutes: 68 device: h200_18gb @@ -22,7 +22,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Multimodal Models (Standard) 2: qwen3 + gemma" +- label: ":nvidia: (H200 MIG 18GB) Multimodal Models (Standard) 2: qwen3 + gemma" key: multi-modal-models-standard-2-qwen3-gemma timeout_in_minutes: 75 device: h200_18gb @@ -43,7 +43,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Multimodal Models (Standard) 3: llava + qwen2_vl" +- label: ":nvidia: (H200 MIG 35GB) Multimodal Models (Standard) 3: llava + qwen2_vl" device: h200_35gb key: multi-modal-models-standard-3-llava-qwen2-vl timeout_in_minutes: 40 @@ -62,7 +62,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Multimodal Models (Standard) 4: other + whisper" +- label: ":nvidia: (H200 MIG 35GB) Multimodal Models (Standard) 4: other + whisper" device: h200_35gb key: multi-modal-models-standard-4-other-whisper timeout_in_minutes: 75 @@ -99,7 +99,7 @@ steps: - pytest -v -s models/multimodal/processing --ignore models/multimodal/processing/test_tensor_schema.py --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --shard-id=$$BUILDKITE_PARALLEL_JOB parallelism: 4 -- label: ":nvidia: (H200) Multimodal Processor" +- label: ":nvidia: (H200 MIG 18GB) Multimodal Processor" key: multi-modal-processor timeout_in_minutes: 98 device: h200_18gb @@ -119,7 +119,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Multimodal Accuracy Eval (Small Models)" +- label: ":nvidia: (H200 MIG 35GB) Multimodal Accuracy Eval (Small Models)" device: h200_35gb key: multi-modal-accuracy-eval-small-models timeout_in_minutes: 30 @@ -145,7 +145,7 @@ steps: - vllm/platforms/rocm.py - vllm/model_executor/model_loader/ -- label: ":nvidia: (H200) Multimodal Models (Extended Generation 1)" +- label: ":nvidia: (H200 MIG 35GB) Multimodal Models (Extended Generation 1)" device: h200_35gb key: multi-modal-models-extended-generation-1 optional: true @@ -166,7 +166,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Multimodal Models (PPL)" +- label: ":nvidia: (H200 MIG 35GB) Multimodal Models (PPL)" device: h200_35gb key: multi-modal-models-extended-ppl optional: true @@ -184,7 +184,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Multimodal Models (Extended Generation 2) Shard %N" +- label: ":nvidia: (H200 MIG 35GB) Multimodal Models (Extended Generation 2) Shard %N" device: h200_35gb key: multi-modal-models-extended-generation-2 parallelism: 4 @@ -204,7 +204,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Multimodal Models (Extended Generation 3)" +- label: ":nvidia: (H200 MIG 35GB) Multimodal Models (Extended Generation 3)" device: h200_35gb key: multi-modal-models-extended-generation-3 optional: true @@ -223,7 +223,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Multimodal Models (Extended Pooling)" +- label: ":nvidia: (H200 MIG 18GB) Multimodal Models (Extended Pooling)" key: multi-modal-models-extended-pooling optional: true device: h200_18gb diff --git a/.buildkite/test_areas/plugins.yaml b/.buildkite/test_areas/plugins.yaml index 19fa0e4d1e..f4f5712023 100644 --- a/.buildkite/test_areas/plugins.yaml +++ b/.buildkite/test_areas/plugins.yaml @@ -62,7 +62,7 @@ steps: - vllm/platforms/rocm.py -- label: ":nvidia: (H200) GGUF Plugin" +- label: ":nvidia: (H200 MIG 18GB) GGUF Plugin" key: gguf-plugin device: h200_18gb timeout_in_minutes: 30 @@ -75,7 +75,7 @@ steps: - pip install "vllm-gguf-plugin >= 0.0.2" - pytest -v -s plugins_tests/gguf -- label: ":nvidia: (H200) BitsAndBytes Plugin" +- label: ":nvidia: (H200 MIG 18GB) BitsAndBytes Plugin" key: bitsandbytes-plugin device: h200_18gb timeout_in_minutes: 30 diff --git a/.buildkite/test_areas/pytorch.yaml b/.buildkite/test_areas/pytorch.yaml index 810d335cf0..a5e39f7b66 100644 --- a/.buildkite/test_areas/pytorch.yaml +++ b/.buildkite/test_areas/pytorch.yaml @@ -2,7 +2,7 @@ group: PyTorch depends_on: - image-build steps: -- label: ":nvidia: (H200) PyTorch Compilation" +- label: ":nvidia: (H200 MIG 35GB) PyTorch Compilation" device: h200_35gb key: pytorch-compilation-unit-tests timeout_in_minutes: 60 @@ -54,7 +54,7 @@ steps: - csrc/ - vllm/platforms/rocm.py -- label: ":nvidia: (H200) PyTorch Compilation H100 Cases" +- label: ":nvidia: (H200 MIG 18GB) PyTorch Compilation H100 Cases" key: pytorch-compilation-unit-tests-h100 timeout_in_minutes: 30 device: h200_18gb @@ -121,7 +121,7 @@ steps: commands: - pytest -s -v compile/passes --ignore compile/passes/distributed -- label: ":nvidia: (H200) PyTorch Fullgraph" +- label: ":nvidia: (H200 MIG 35GB) PyTorch Fullgraph" device: h200_35gb key: pytorch-fullgraph-test timeout_in_minutes: 90 @@ -195,7 +195,7 @@ steps: commands: - pytest -s -v compile/fullgraph/test_full_cudagraph.py -- label: ":nvidia: (H200) PyTorch Nightly Dependency Override Check" +- label: ":nvidia: (H200 MIG 18GB) PyTorch Nightly Dependency Override Check" key: pytorch-nightly-dependency-override-check # if this test fails, it means the nightly torch version is not compatible with some # of the dependencies. Please check the error message and add the package to whitelist diff --git a/.buildkite/test_areas/quantization.yaml b/.buildkite/test_areas/quantization.yaml index e659cd75f6..1d895e1f2f 100644 --- a/.buildkite/test_areas/quantization.yaml +++ b/.buildkite/test_areas/quantization.yaml @@ -2,7 +2,7 @@ group: Quantization depends_on: - image-build steps: -- label: ":nvidia: (H200) Quantization Shard %N" +- label: ":nvidia: (H200 MIG 35GB) Quantization Shard %N" device: h200_35gb key: quantization timeout_in_minutes: 40 @@ -21,7 +21,7 @@ steps: # parameter. It was not exercised by the previous L4 job. - VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization/ --ignore quantization/test_blackwell_moe.py -k 'not test_compressed_tensors_w4a8_fp8' --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT -- label: ":nvidia: (H200) Quantized Fusions" +- label: ":nvidia: (H200 MIG 35GB) Quantized Fusions" device: h200_35gb key: quantized-fusions timeout_in_minutes: 20 @@ -52,7 +52,7 @@ steps: commands: - pytest -s -v tests/quantization/test_blackwell_moe.py -- label: ":nvidia: (H200) Quantized Models" +- label: ":nvidia: (H200 MIG 35GB) Quantized Models" device: h200_35gb key: quantized-models-test timeout_in_minutes: 65 diff --git a/.buildkite/test_areas/ray_compat.yaml b/.buildkite/test_areas/ray_compat.yaml index 243db5f211..d24c38430b 100644 --- a/.buildkite/test_areas/ray_compat.yaml +++ b/.buildkite/test_areas/ray_compat.yaml @@ -2,7 +2,7 @@ group: Ray Compatibility depends_on: - image-build steps: -- label: ":nvidia: (H200) Ray Dependency Compatibility Check" +- label: ":nvidia: (H200 MIG 18GB) Ray Dependency Compatibility Check" key: ray-dependency-compatibility-check # Informational only — does not block the pipeline. # If this fails, it means the PR introduces a dependency that diff --git a/.buildkite/test_areas/rust_frontend.yaml b/.buildkite/test_areas/rust_frontend.yaml index b388b8d0f7..2f4d264c1c 100644 --- a/.buildkite/test_areas/rust_frontend.yaml +++ b/.buildkite/test_areas/rust_frontend.yaml @@ -2,7 +2,7 @@ group: Rust Frontend E2E depends_on: - image-build steps: -- label: ":nvidia: (H200) Rust Frontend OpenAI Coverage" +- label: ":nvidia: (H200 MIG 18GB) Rust Frontend OpenAI Coverage" timeout_in_minutes: 30 device: h200_18gb working_dir: "/vllm-workspace/tests" @@ -60,7 +60,7 @@ steps: - tests/v1/sample/test_logprobs_e2e.py - vllm/platforms/rocm.py -- label: ":nvidia: (H200) Rust Frontend Serve/Admin Coverage" +- label: ":nvidia: (H200 MIG 18GB) Rust Frontend Serve/Admin Coverage" timeout_in_minutes: 25 device: h200_18gb working_dir: "/vllm-workspace/tests" @@ -109,7 +109,7 @@ steps: - tests/entrypoints/serve/tokenize/test_tokenization.py - vllm/platforms/rocm.py -- label: ":nvidia: (H200) Rust Frontend Core Correctness" +- label: ":nvidia: (H200 MIG 18GB) Rust Frontend Core Correctness" timeout_in_minutes: 20 device: h200_18gb working_dir: "/vllm-workspace/tests" @@ -123,7 +123,7 @@ steps: - export VLLM_WORKER_MULTIPROC_METHOD=spawn - pytest -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine -- label: ":nvidia: (H200) Rust Frontend Tool Use" +- label: ":nvidia: (H200 MIG 35GB) Rust Frontend Tool Use" device: h200_35gb timeout_in_minutes: 25 working_dir: "/vllm-workspace/tests" diff --git a/.buildkite/test_areas/samplers.yaml b/.buildkite/test_areas/samplers.yaml index c0297d79e5..cfde3717c8 100644 --- a/.buildkite/test_areas/samplers.yaml +++ b/.buildkite/test_areas/samplers.yaml @@ -2,7 +2,7 @@ group: Samplers depends_on: - image-build steps: -- label: ":nvidia: (H200) Samplers" +- label: ":nvidia: (H200 MIG 35GB) Samplers" device: h200_35gb key: samplers-test timeout_in_minutes: 40 diff --git a/.buildkite/test_areas/spec_decode.yaml b/.buildkite/test_areas/spec_decode.yaml index 0f3a45838b..1907dc44b9 100644 --- a/.buildkite/test_areas/spec_decode.yaml +++ b/.buildkite/test_areas/spec_decode.yaml @@ -2,7 +2,7 @@ group: Spec Decode depends_on: - image-build steps: -- label: ":nvidia: (H200) V1 Spec Decode" +- label: ":nvidia: (H200 MIG 35GB) V1 Spec Decode" device: h200_35gb key: v1-spec-decode timeout_in_minutes: 40 @@ -30,7 +30,7 @@ steps: depends_on: - image-build-amd -- label: ":nvidia: (H200) Spec Decode Eagle 1: DeepSeek + Qwen" +- label: ":nvidia: (H200 MIG 35GB) Spec Decode Eagle 1: DeepSeek + Qwen" key: spec-decode-eagle-1-deepseek-qwen timeout_in_minutes: 25 device: h200_35gb @@ -57,7 +57,7 @@ steps: - tests/v1/e2e/spec_decode/ - vllm/platforms/rocm.py -- label: ":nvidia: (H200) Spec Decode Eagle 2: Llama 3 + Qwen VL + Other" +- label: ":nvidia: (H200 MIG 35GB) Spec Decode Eagle 2: Llama 3 + Qwen VL + Other" key: spec-decode-eagle-2-llama3-qwen-vl-other timeout_in_minutes: 25 device: h200_35gb @@ -96,7 +96,7 @@ steps: commands: - pytest -v -s v1/e2e/spec_decode/eagle/ -- label: ":nvidia: (H200) Spec Decode Speculators + MTP" +- label: ":nvidia: (H200 MIG 35GB) Spec Decode Speculators + MTP" key: spec-decode-speculators-mtp timeout_in_minutes: 50 device: h200_35gb @@ -141,7 +141,7 @@ steps: - pytest -v -s v1/e2e/spec_decode/speculators/ - pytest -v -s v1/e2e/spec_decode/mtp/ -- label: ":nvidia: (H200) Spec Decode N-Gram + Suffix" +- label: ":nvidia: (H200 MIG 35GB) Spec Decode N-Gram + Suffix" key: spec-decode-ngram-suffix timeout_in_minutes: 20 device: h200_35gb @@ -170,7 +170,7 @@ steps: - tests/v1/e2e/spec_decode/ - vllm/platforms/rocm.py -- label: ":nvidia: (H200) Spec Decode Draft Model" +- label: ":nvidia: (H200 MIG 18GB) Spec Decode Draft Model" key: spec-decode-draft-model timeout_in_minutes: 45 device: h200_18gb @@ -269,7 +269,7 @@ steps: - tests/v1/e2e/spec_decode/test_mtp_parallel_load.py - vllm/platforms/rocm.py -- label: ":nvidia: (H200) Spec Decode AL DFlash Nightly" +- label: ":nvidia: (H200 MIG 35GB) Spec Decode AL DFlash Nightly" key: spec-decode-dflash-nightly timeout_in_minutes: 90 device: h200_35gb @@ -305,7 +305,7 @@ steps: - tests/v1/e2e/spec_decode/ - vllm/platforms/rocm.py -- label: ":nvidia: (H200) Spec Decode AL DSpark Nightly" +- label: ":nvidia: (H200 MIG 35GB) Spec Decode AL DSpark Nightly" key: spec-decode-dspark-nightly timeout_in_minutes: 60 device: h200_35gb @@ -345,7 +345,7 @@ steps: - tests/v1/e2e/spec_decode/ - vllm/platforms/rocm.py -- label: ":nvidia: (H200) Spec Decode AL MTP + Other Acceptance Nightly" +- label: ":nvidia: (H200 MIG 35GB) Spec Decode AL MTP + Other Acceptance Nightly" key: spec-decode-mtp-other-acceptance-nightly timeout_in_minutes: 60 device: h200_35gb