mirror of
https://github.com/vllm-project/vllm.git
synced 2026-10-28 09:45:26 +08:00
[CI] Disable jobs with unlocked transitive dependencies (#63608)
Signed-off-by: khluu <khluu000@gmail.com>
This commit is contained in:
Kevin H. Luu authored and GitHub committed
1 parent
0765565f10
commit
9d48c8ab95
7 files changed
+364
-308
No files matched your search
@@ -15,28 +15,29 @@ steps:
|
||||
limit: 2
|
||||
- exit_status: -10 # Agent was lost
|
||||
limit: 2
|
||||
- label: ":intel: (BMG) Example Test"
|
||||
key: xpu-example-test
|
||||
depends_on:
|
||||
- image-build-xpu
|
||||
timeout_in_minutes: 50
|
||||
optional: true
|
||||
device: intel_gpu
|
||||
agent_tags:
|
||||
label: production
|
||||
gpu: 2+
|
||||
mem: 24+
|
||||
no_plugin: true
|
||||
env:
|
||||
REGISTRY: "public.ecr.aws/q9t5s3a7"
|
||||
REPO: "vllm-ci-test-repo"
|
||||
source_file_dependencies:
|
||||
- .buildkite/hardware_tests/intel_xpu_ci/test-intel.yaml
|
||||
- .buildkite/scripts/hardware_ci/run-intel-ci-test.sh
|
||||
commands:
|
||||
- >-
|
||||
bash .buildkite/scripts/hardware_ci/run-intel-test.sh
|
||||
'bash .buildkite/scripts/hardware_ci/run-intel-ci-test.sh example'
|
||||
# Disabled until tblib transitive dependencies are locked.
|
||||
# - label: ":intel: (BMG) Example Test"
|
||||
# key: xpu-example-test
|
||||
# depends_on:
|
||||
# - image-build-xpu
|
||||
# timeout_in_minutes: 50
|
||||
# optional: true
|
||||
# device: intel_gpu
|
||||
# agent_tags:
|
||||
# label: production
|
||||
# gpu: 2+
|
||||
# mem: 24+
|
||||
# no_plugin: true
|
||||
# env:
|
||||
# REGISTRY: "public.ecr.aws/q9t5s3a7"
|
||||
# REPO: "vllm-ci-test-repo"
|
||||
# source_file_dependencies:
|
||||
# - .buildkite/hardware_tests/intel_xpu_ci/test-intel.yaml
|
||||
# - .buildkite/scripts/hardware_ci/run-intel-ci-test.sh
|
||||
# commands:
|
||||
# - >-
|
||||
# bash .buildkite/scripts/hardware_ci/run-intel-test.sh
|
||||
# 'bash .buildkite/scripts/hardware_ci/run-intel-ci-test.sh example'
|
||||
- label: ":intel: (BMG) V1 test"
|
||||
key: xpu-v1-test
|
||||
depends_on:
|
||||
|
||||
+109
-104
@@ -134,23 +134,24 @@ steps:
|
||||
commands:
|
||||
- pytest -v -s tools/test_docker_build_metadata_args.py
|
||||
|
||||
- label: ":amd: (MI300) Torch Stable ABI Audit" # TBD
|
||||
timeout_in_minutes: 35
|
||||
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
|
||||
dind: false
|
||||
agent_pool: mi300_1
|
||||
no_gpu: true
|
||||
optional: true
|
||||
working_dir: "/vllm-workspace/"
|
||||
source_file_dependencies:
|
||||
- .buildkite/check-torch-abi.py
|
||||
- csrc/
|
||||
- cmake/
|
||||
- setup.py
|
||||
- vllm/platforms/rocm.py
|
||||
commands:
|
||||
- uv pip install --system torch-abi-audit==0.0.1
|
||||
- python3 /vllm-workspace/.buildkite/check-torch-abi.py
|
||||
# Disabled until torch-abi-audit transitive dependencies are locked.
|
||||
# - label: ":amd: (MI300) Torch Stable ABI Audit" # TBD
|
||||
# timeout_in_minutes: 35
|
||||
# mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
|
||||
# dind: false
|
||||
# agent_pool: mi300_1
|
||||
# no_gpu: true
|
||||
# optional: true
|
||||
# working_dir: "/vllm-workspace/"
|
||||
# source_file_dependencies:
|
||||
# - .buildkite/check-torch-abi.py
|
||||
# - csrc/
|
||||
# - cmake/
|
||||
# - setup.py
|
||||
# - vllm/platforms/rocm.py
|
||||
# commands:
|
||||
# - uv pip install --system torch-abi-audit==0.0.1
|
||||
# - python3 /vllm-workspace/.buildkite/check-torch-abi.py
|
||||
|
||||
#########################################################################################################################################
|
||||
# #
|
||||
@@ -207,22 +208,23 @@ steps:
|
||||
|
||||
#---------------------------------------------------------- mi250 · kernels ----------------------------------------------------------#
|
||||
|
||||
- label: ":amd: (MI250) Helion Kernels Shard %N"
|
||||
timeout_in_minutes: 45
|
||||
parallelism: 2
|
||||
mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250]
|
||||
dind: false
|
||||
agent_pool: mi250_1
|
||||
optional: true
|
||||
working_dir: "/vllm-workspace/tests"
|
||||
source_file_dependencies:
|
||||
- vllm/utils/import_utils.py
|
||||
- tools/benchmark_helion_kernels.py
|
||||
- tests/kernels/helion/
|
||||
- vllm/platforms/rocm.py
|
||||
commands:
|
||||
- pip install helion==1.4.0
|
||||
- pytest -v -s kernels/helion/ --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
|
||||
# Disabled until helion transitive dependencies are locked.
|
||||
# - label: ":amd: (MI250) Helion Kernels Shard %N"
|
||||
# timeout_in_minutes: 45
|
||||
# parallelism: 2
|
||||
# mirror_hardwares: [amdexperimental, amdproduction, amdgfx90anightly, amdmi250]
|
||||
# dind: false
|
||||
# agent_pool: mi250_1
|
||||
# optional: true
|
||||
# working_dir: "/vllm-workspace/tests"
|
||||
# source_file_dependencies:
|
||||
# - vllm/utils/import_utils.py
|
||||
# - tools/benchmark_helion_kernels.py
|
||||
# - tests/kernels/helion/
|
||||
# - vllm/platforms/rocm.py
|
||||
# commands:
|
||||
# - pip install helion==1.4.0
|
||||
# - pytest -v -s kernels/helion/ --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
|
||||
|
||||
#----------------------------------------------------- mi250 · models / language -----------------------------------------------------#
|
||||
|
||||
@@ -1841,36 +1843,37 @@ steps:
|
||||
|
||||
#------------------------------------------------------- mi300 · quantization --------------------------------------------------------#
|
||||
|
||||
- label: ":amd: (MI300) Quantization Shard %N"
|
||||
timeout_in_minutes: 115
|
||||
parallelism: 4
|
||||
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
|
||||
dind: false
|
||||
agent_pool: mi300_1
|
||||
working_dir: "/vllm-workspace/tests"
|
||||
source_file_dependencies:
|
||||
- csrc/
|
||||
- vllm/model_executor/layers/quantization
|
||||
- vllm/_aiter_ops.py
|
||||
- vllm/platforms/rocm.py
|
||||
- vllm/model_executor/layers/fused_moe/oracle/unquantized.py
|
||||
- vllm/model_executor/layers/fused_moe/unquantized_fused_moe_method.py
|
||||
- tests/quantization
|
||||
- tests/rocm/test_moe_weight_replay.py
|
||||
- tests/rocm/test_moe_padding.py
|
||||
optional: true
|
||||
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==0.18.0
|
||||
- uv pip install --system conch-triton-kernels
|
||||
- VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization/ --ignore quantization/test_blackwell_moe.py --ignore quantization/test_rocm_moe.py --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
|
||||
- if [ "$$BUILDKITE_PARALLEL_JOB" = "0" ]; then pytest -v -s rocm/test_moe_weight_replay.py; fi
|
||||
- if [ "$$BUILDKITE_PARALLEL_JOB" = "0" ]; then VLLM_ROCM_USE_AITER=1 VLLM_ROCM_USE_AITER_MOE=1 pytest -v -s rocm/test_moe_padding.py; fi
|
||||
# Disabled until torchao transitive dependencies are locked.
|
||||
# - label: ":amd: (MI300) Quantization Shard %N"
|
||||
# timeout_in_minutes: 115
|
||||
# parallelism: 4
|
||||
# mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
|
||||
# dind: false
|
||||
# agent_pool: mi300_1
|
||||
# working_dir: "/vllm-workspace/tests"
|
||||
# source_file_dependencies:
|
||||
# - csrc/
|
||||
# - vllm/model_executor/layers/quantization
|
||||
# - vllm/_aiter_ops.py
|
||||
# - vllm/platforms/rocm.py
|
||||
# - vllm/model_executor/layers/fused_moe/oracle/unquantized.py
|
||||
# - vllm/model_executor/layers/fused_moe/unquantized_fused_moe_method.py
|
||||
# - tests/quantization
|
||||
# - tests/rocm/test_moe_weight_replay.py
|
||||
# - tests/rocm/test_moe_padding.py
|
||||
# optional: true
|
||||
# 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==0.18.0
|
||||
# - uv pip install --system conch-triton-kernels
|
||||
# - VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization/ --ignore quantization/test_blackwell_moe.py --ignore quantization/test_rocm_moe.py --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
|
||||
# - if [ "$$BUILDKITE_PARALLEL_JOB" = "0" ]; then pytest -v -s rocm/test_moe_weight_replay.py; fi
|
||||
# - if [ "$$BUILDKITE_PARALLEL_JOB" = "0" ]; then VLLM_ROCM_USE_AITER=1 VLLM_ROCM_USE_AITER_MOE=1 pytest -v -s rocm/test_moe_padding.py; fi
|
||||
|
||||
- label: ":amd: (MI300) Quantized Fusions" # TBD
|
||||
timeout_in_minutes: 45
|
||||
@@ -3558,25 +3561,26 @@ steps:
|
||||
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
|
||||
- pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=configs/models-blackwell-ep-mi355.txt
|
||||
|
||||
- label: ":amd: (MI355) Qwen3.8-Flash-Next-FP8 Accuracy Eval"
|
||||
timeout_in_minutes: 105
|
||||
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
|
||||
dind: false
|
||||
agent_pool: mi355_4
|
||||
num_gpus: 4
|
||||
optional: true
|
||||
working_dir: "/vllm-workspace/tests"
|
||||
source_file_dependencies:
|
||||
- csrc/gdn/
|
||||
- tests/evals/qwen4_exp/
|
||||
- vllm/model_executor/layers/mamba/
|
||||
- vllm/models/qwen4_exp/
|
||||
- vllm/v1/attention/backends/short_conv_attn.py
|
||||
- vllm/platforms/rocm.py
|
||||
commands:
|
||||
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
|
||||
- uv pip install --system 'evalscope==1.10.0'
|
||||
- pytest -s -v evals/qwen4_exp/test_accuracy.py --config-list-file=configs/models-rocm.txt
|
||||
# Disabled until evalscope transitive dependencies are locked.
|
||||
# - label: ":amd: (MI355) Qwen3.8-Flash-Next-FP8 Accuracy Eval"
|
||||
# timeout_in_minutes: 105
|
||||
# mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
|
||||
# dind: false
|
||||
# agent_pool: mi355_4
|
||||
# num_gpus: 4
|
||||
# optional: true
|
||||
# working_dir: "/vllm-workspace/tests"
|
||||
# source_file_dependencies:
|
||||
# - csrc/gdn/
|
||||
# - tests/evals/qwen4_exp/
|
||||
# - vllm/model_executor/layers/mamba/
|
||||
# - vllm/models/qwen4_exp/
|
||||
# - vllm/v1/attention/backends/short_conv_attn.py
|
||||
# - vllm/platforms/rocm.py
|
||||
# commands:
|
||||
# - export VLLM_WORKER_MULTIPROC_METHOD=spawn
|
||||
# - uv pip install --system 'evalscope==1.10.0'
|
||||
# - pytest -s -v evals/qwen4_exp/test_accuracy.py --config-list-file=configs/models-rocm.txt
|
||||
|
||||
- label: ":amd: (MI355 DPX) LM Eval TurboQuant k8v4"
|
||||
timeout_in_minutes: 30
|
||||
@@ -5038,28 +5042,29 @@ steps:
|
||||
|
||||
#------------------------------------------------------- mi355 · quantization --------------------------------------------------------#
|
||||
|
||||
- label: ":amd: (MI355 DPX) Quantization Shard %N"
|
||||
timeout_in_minutes: 115
|
||||
parallelism: 4
|
||||
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
|
||||
dind: false
|
||||
agent_pool: mi355_dpx
|
||||
working_dir: "/vllm-workspace/tests"
|
||||
source_file_dependencies:
|
||||
- csrc/
|
||||
- vllm/model_executor/layers/quantization
|
||||
- tests/quantization
|
||||
- vllm/_aiter_ops.py
|
||||
- vllm/platforms/rocm.py
|
||||
- vllm/model_executor/layers/fused_moe/oracle/unquantized.py
|
||||
- vllm/model_executor/layers/fused_moe/unquantized_fused_moe_method.py
|
||||
- tests/rocm/test_moe_weight_replay.py
|
||||
optional: true
|
||||
commands:
|
||||
- uv pip install --system torchao==0.18.0
|
||||
- uv pip install --system conch-triton-kernels
|
||||
- VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization/ --ignore quantization/test_blackwell_moe.py --ignore quantization/test_rocm_moe.py --ignore quantization/test_gfx950_moe.py --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
|
||||
- if [ "$$BUILDKITE_PARALLEL_JOB" = "0" ]; then pytest -v -s rocm/test_moe_weight_replay.py; fi
|
||||
# Disabled until torchao transitive dependencies are locked.
|
||||
# - label: ":amd: (MI355 DPX) Quantization Shard %N"
|
||||
# timeout_in_minutes: 115
|
||||
# parallelism: 4
|
||||
# mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
|
||||
# dind: false
|
||||
# agent_pool: mi355_dpx
|
||||
# working_dir: "/vllm-workspace/tests"
|
||||
# source_file_dependencies:
|
||||
# - csrc/
|
||||
# - vllm/model_executor/layers/quantization
|
||||
# - tests/quantization
|
||||
# - vllm/_aiter_ops.py
|
||||
# - vllm/platforms/rocm.py
|
||||
# - vllm/model_executor/layers/fused_moe/oracle/unquantized.py
|
||||
# - vllm/model_executor/layers/fused_moe/unquantized_fused_moe_method.py
|
||||
# - tests/rocm/test_moe_weight_replay.py
|
||||
# optional: true
|
||||
# commands:
|
||||
# - uv pip install --system torchao==0.18.0
|
||||
# - uv pip install --system conch-triton-kernels
|
||||
# - VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization/ --ignore quantization/test_blackwell_moe.py --ignore quantization/test_rocm_moe.py --ignore quantization/test_gfx950_moe.py --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
|
||||
# - if [ "$$BUILDKITE_PARALLEL_JOB" = "0" ]; then pytest -v -s rocm/test_moe_weight_replay.py; fi
|
||||
|
||||
#--------------------------------------------------------- mi355 · samplers ----------------------------------------------------------#
|
||||
|
||||
|
||||
@@ -744,69 +744,71 @@ steps:
|
||||
commands:
|
||||
- pytest -v -s tests/kernels/test_rocm_fp32_router_gemm.py tests/kernels/test_gate_linear_rocm_dispatch.py tests/kernels/test_bf16x3_router_gemm_rocm.py
|
||||
|
||||
- label: ":nvidia: (DGX Spark) [Nightly] b12x Linear Kernels"
|
||||
key: b12x-linear-kernels-dgx-spark-nightly
|
||||
timeout_in_minutes: 60
|
||||
device: dgx-spark
|
||||
optional: true
|
||||
num_devices: 1
|
||||
depends_on:
|
||||
- arm64-image-build
|
||||
source_file_dependencies:
|
||||
- setup.py
|
||||
- vllm/config/kernel.py
|
||||
- vllm/envs.py
|
||||
- vllm/model_executor/kernels/linear/
|
||||
- vllm/model_executor/layers/fused_moe/b12x.py
|
||||
- vllm/model_executor/layers/fused_moe/config.py
|
||||
- vllm/model_executor/layers/fused_moe/oracle/mxfp4.py
|
||||
- vllm/model_executor/layers/fused_moe/oracle/nvfp4.py
|
||||
- vllm/model_executor/layers/quantization/compressed_tensors/compressed_tensors_moe/
|
||||
- vllm/model_executor/layers/quantization/inc/schemes/inc_mxfp4_moe.py
|
||||
- vllm/model_executor/layers/quantization/modelopt.py
|
||||
- vllm/model_executor/layers/quantization/mxfp4.py
|
||||
- vllm/model_executor/layers/quantization/utils/b12x_moe.py
|
||||
- vllm/model_executor/warmup/b12x_warmup.py
|
||||
- vllm/utils/b12x.py
|
||||
- tests/kernels/moe/test_b12x.py
|
||||
- tests/kernels/quantization/nvfp4_utils.py
|
||||
- tests/model_executor/kernels/test_b12x_linear.py
|
||||
- tests/model_executor/test_b12x_warmup.py
|
||||
- tests/kernels/quantization/test_block_fp8.py
|
||||
- tests/quantization/reference_mxfp4.py
|
||||
- tests/quantization/test_auto_round.py
|
||||
commands:
|
||||
- uv pip install --system b12x==1.5.1
|
||||
- pytest -v -s model_executor/kernels/test_b12x_linear.py
|
||||
model_executor/test_b12x_warmup.py
|
||||
kernels/quantization/test_block_fp8.py
|
||||
kernels/moe/test_b12x.py
|
||||
quantization/test_auto_round.py -k b12x
|
||||
# Disabled until b12x transitive dependencies are locked.
|
||||
# - label: ":nvidia: (DGX Spark) [Nightly] b12x Linear Kernels"
|
||||
# key: b12x-linear-kernels-dgx-spark-nightly
|
||||
# timeout_in_minutes: 60
|
||||
# device: dgx-spark
|
||||
# optional: true
|
||||
# num_devices: 1
|
||||
# depends_on:
|
||||
# - arm64-image-build
|
||||
# source_file_dependencies:
|
||||
# - setup.py
|
||||
# - vllm/config/kernel.py
|
||||
# - vllm/envs.py
|
||||
# - vllm/model_executor/kernels/linear/
|
||||
# - vllm/model_executor/layers/fused_moe/b12x.py
|
||||
# - vllm/model_executor/layers/fused_moe/config.py
|
||||
# - vllm/model_executor/layers/fused_moe/oracle/mxfp4.py
|
||||
# - vllm/model_executor/layers/fused_moe/oracle/nvfp4.py
|
||||
# - vllm/model_executor/layers/quantization/compressed_tensors/compressed_tensors_moe/
|
||||
# - vllm/model_executor/layers/quantization/inc/schemes/inc_mxfp4_moe.py
|
||||
# - vllm/model_executor/layers/quantization/modelopt.py
|
||||
# - vllm/model_executor/layers/quantization/mxfp4.py
|
||||
# - vllm/model_executor/layers/quantization/utils/b12x_moe.py
|
||||
# - vllm/model_executor/warmup/b12x_warmup.py
|
||||
# - vllm/utils/b12x.py
|
||||
# - tests/kernels/moe/test_b12x.py
|
||||
# - tests/kernels/quantization/nvfp4_utils.py
|
||||
# - tests/model_executor/kernels/test_b12x_linear.py
|
||||
# - tests/model_executor/test_b12x_warmup.py
|
||||
# - tests/kernels/quantization/test_block_fp8.py
|
||||
# - tests/quantization/reference_mxfp4.py
|
||||
# - tests/quantization/test_auto_round.py
|
||||
# commands:
|
||||
# - uv pip install --system b12x==1.5.1
|
||||
# - pytest -v -s model_executor/kernels/test_b12x_linear.py
|
||||
# model_executor/test_b12x_warmup.py
|
||||
# kernels/quantization/test_block_fp8.py
|
||||
# kernels/moe/test_b12x.py
|
||||
# quantization/test_auto_round.py -k b12x
|
||||
|
||||
- label: ":nvidia: (H100) Helion Kernels Shard %N"
|
||||
key: kernels-helion-test
|
||||
timeout_in_minutes: 115
|
||||
device: h100
|
||||
source_file_dependencies:
|
||||
- vllm/utils/import_utils.py
|
||||
- tests/kernels/helion/
|
||||
commands:
|
||||
- pip install helion==1.4.0
|
||||
- pytest -v -s kernels/helion/ --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
|
||||
parallelism: 5
|
||||
mirror:
|
||||
amd:
|
||||
label: ":amd: (MI250) Helion Kernels Shard %N"
|
||||
dind: false
|
||||
device: mi250_1
|
||||
timeout_in_minutes: 45
|
||||
# Explicit so the base step's count does not multiply the mirror's replicas.
|
||||
parallelism: 2
|
||||
depends_on:
|
||||
- image-build-amd
|
||||
source_file_dependencies:
|
||||
- vllm/platforms/rocm.py
|
||||
- tools/benchmark_helion_kernels.py
|
||||
# Disabled until helion transitive dependencies are locked.
|
||||
# - label: ":nvidia: (H100) Helion Kernels Shard %N"
|
||||
# key: kernels-helion-test
|
||||
# timeout_in_minutes: 115
|
||||
# device: h100
|
||||
# source_file_dependencies:
|
||||
# - vllm/utils/import_utils.py
|
||||
# - tests/kernels/helion/
|
||||
# commands:
|
||||
# - pip install helion==1.4.0
|
||||
# - pytest -v -s kernels/helion/ --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
|
||||
# parallelism: 5
|
||||
# mirror:
|
||||
# amd:
|
||||
# label: ":amd: (MI250) Helion Kernels Shard %N"
|
||||
# dind: false
|
||||
# device: mi250_1
|
||||
# timeout_in_minutes: 45
|
||||
# # Explicit so the base step's count does not multiply the mirror's replicas.
|
||||
# parallelism: 2
|
||||
# depends_on:
|
||||
# - image-build-amd
|
||||
# source_file_dependencies:
|
||||
# - vllm/platforms/rocm.py
|
||||
# - tools/benchmark_helion_kernels.py
|
||||
|
||||
|
||||
- label: ":nvidia: (H100) [Nightly] FP8 MoE Kernels"
|
||||
|
||||
@@ -1191,59 +1191,61 @@ steps:
|
||||
commands:
|
||||
- pytest -s -v evals/gpt_oss/test_gpqa_correctness.py --config-list-file=configs/models-gfx950.txt
|
||||
|
||||
- label: ":nvidia: (B200) [Nightly] Qwen3.8-Flash-Next-FP8 Accuracy Eval"
|
||||
key: accuracy-eval-qwen3-8-flash-next-fp8-b200
|
||||
timeout_in_minutes: 60
|
||||
device: b200-k8s
|
||||
optional: true
|
||||
num_devices: 4
|
||||
source_file_dependencies:
|
||||
- csrc/gdn/
|
||||
- tests/evals/qwen4_exp/
|
||||
- vllm/model_executor/layers/mamba/
|
||||
- vllm/models/qwen4_exp/
|
||||
- vllm/v1/attention/backends/short_conv_attn.py
|
||||
commands:
|
||||
- uv pip install --system 'evalscope==1.10.0'
|
||||
- pytest -s -v evals/qwen4_exp/test_accuracy.py --config-list-file=configs/models-b200.txt
|
||||
# Disabled until evalscope transitive dependencies are locked.
|
||||
# - label: ":nvidia: (B200) [Nightly] Qwen3.8-Flash-Next-FP8 Accuracy Eval"
|
||||
# key: accuracy-eval-qwen3-8-flash-next-fp8-b200
|
||||
# timeout_in_minutes: 60
|
||||
# device: b200-k8s
|
||||
# optional: true
|
||||
# num_devices: 4
|
||||
# source_file_dependencies:
|
||||
# - csrc/gdn/
|
||||
# - tests/evals/qwen4_exp/
|
||||
# - vllm/model_executor/layers/mamba/
|
||||
# - vllm/models/qwen4_exp/
|
||||
# - vllm/v1/attention/backends/short_conv_attn.py
|
||||
# commands:
|
||||
# - uv pip install --system 'evalscope==1.10.0'
|
||||
# - pytest -s -v evals/qwen4_exp/test_accuracy.py --config-list-file=configs/models-b200.txt
|
||||
|
||||
- label: ":nvidia: (H200) [Nightly] Qwen3.8-Flash-Next-FP8 Accuracy Eval"
|
||||
key: accuracy-eval-qwen3-8-flash-next-fp8-h200
|
||||
timeout_in_minutes: 60
|
||||
device: h200
|
||||
optional: true
|
||||
num_devices: 4
|
||||
source_file_dependencies:
|
||||
- csrc/gdn/
|
||||
- tests/evals/qwen4_exp/
|
||||
- vllm/model_executor/layers/mamba/
|
||||
- vllm/models/qwen4_exp/
|
||||
- vllm/v1/attention/backends/short_conv_attn.py
|
||||
commands:
|
||||
- uv pip install --system 'evalscope==1.10.0'
|
||||
- pytest -s -v evals/qwen4_exp/test_accuracy.py --config-list-file=configs/models-h200.txt
|
||||
mirror:
|
||||
amd:
|
||||
label: ":amd: (MI355) Qwen3.8-Flash-Next-FP8 Accuracy Eval"
|
||||
dind: false
|
||||
device: mi355_4
|
||||
num_devices: 4
|
||||
timeout_in_minutes: 105
|
||||
working_dir: /vllm-workspace/tests
|
||||
depends_on:
|
||||
- image-build-amd
|
||||
source_file_dependencies:
|
||||
- csrc/gdn/
|
||||
- tests/evals/qwen4_exp/
|
||||
- vllm/model_executor/layers/mamba/
|
||||
- vllm/models/qwen4_exp/
|
||||
- vllm/v1/attention/backends/short_conv_attn.py
|
||||
- vllm/platforms/rocm.py
|
||||
env:
|
||||
VLLM_WORKER_MULTIPROC_METHOD: "spawn"
|
||||
commands:
|
||||
- uv pip install --system 'evalscope==1.10.0'
|
||||
- pytest -s -v evals/qwen4_exp/test_accuracy.py --config-list-file=configs/models-rocm.txt
|
||||
# Disabled until evalscope transitive dependencies are locked.
|
||||
# - label: ":nvidia: (H200) [Nightly] Qwen3.8-Flash-Next-FP8 Accuracy Eval"
|
||||
# key: accuracy-eval-qwen3-8-flash-next-fp8-h200
|
||||
# timeout_in_minutes: 60
|
||||
# device: h200
|
||||
# optional: true
|
||||
# num_devices: 4
|
||||
# source_file_dependencies:
|
||||
# - csrc/gdn/
|
||||
# - tests/evals/qwen4_exp/
|
||||
# - vllm/model_executor/layers/mamba/
|
||||
# - vllm/models/qwen4_exp/
|
||||
# - vllm/v1/attention/backends/short_conv_attn.py
|
||||
# commands:
|
||||
# - uv pip install --system 'evalscope==1.10.0'
|
||||
# - pytest -s -v evals/qwen4_exp/test_accuracy.py --config-list-file=configs/models-h200.txt
|
||||
# mirror:
|
||||
# amd:
|
||||
# label: ":amd: (MI355) Qwen3.8-Flash-Next-FP8 Accuracy Eval"
|
||||
# dind: false
|
||||
# device: mi355_4
|
||||
# num_devices: 4
|
||||
# timeout_in_minutes: 105
|
||||
# working_dir: /vllm-workspace/tests
|
||||
# depends_on:
|
||||
# - image-build-amd
|
||||
# source_file_dependencies:
|
||||
# - csrc/gdn/
|
||||
# - tests/evals/qwen4_exp/
|
||||
# - vllm/model_executor/layers/mamba/
|
||||
# - vllm/models/qwen4_exp/
|
||||
# - vllm/v1/attention/backends/short_conv_attn.py
|
||||
# - vllm/platforms/rocm.py
|
||||
# env:
|
||||
# VLLM_WORKER_MULTIPROC_METHOD: "spawn"
|
||||
# commands:
|
||||
# - uv pip install --system 'evalscope==1.10.0'
|
||||
# - pytest -s -v evals/qwen4_exp/test_accuracy.py --config-list-file=configs/models-rocm.txt
|
||||
|
||||
- label: ":nvidia: (DGX Spark) [Nightly] GPQA Eval (GPT-OSS)"
|
||||
key: gpqa-eval-gpt-oss-spark
|
||||
|
||||
@@ -2,55 +2,56 @@ group: Quantization
|
||||
depends_on:
|
||||
- image-build
|
||||
steps:
|
||||
- label: ":nvidia: (H200 MIG 35GB) Quantization"
|
||||
device: h200_35gb
|
||||
key: quantization
|
||||
automatic_shard: true
|
||||
timeout_in_minutes: 40
|
||||
source_file_dependencies:
|
||||
- csrc/
|
||||
- vllm/model_executor/layers/quantization
|
||||
- vllm/model_executor/models/qwen3_dflash.py
|
||||
- vllm/model_executor/models/utils.py
|
||||
- vllm/transformers_utils/model_arch_config_convertor.py
|
||||
- vllm/models/deepseek_v41/quant_config.py
|
||||
- vllm/utils/humming.py
|
||||
- tests/quantization
|
||||
commands:
|
||||
# Pin torchao to the version compatible with the CI PyTorch and CUDA stack.
|
||||
- uv pip install --system torchao==0.18.0 --index-url https://download.pytorch.org/whl/cu130
|
||||
- uv pip install --system conch-triton-kernels
|
||||
# The SM90-only checkpoint currently contains a removed weight_chan_scale
|
||||
# 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 --ignore quantization/test_rocm_moe.py -k 'not test_compressed_tensors_w4a8_fp8'
|
||||
mirror:
|
||||
amd:
|
||||
label: ":amd: (MI300) Quantization"
|
||||
automatic_shard: true
|
||||
dind: false
|
||||
device: mi300_1
|
||||
timeout_in_minutes: 115
|
||||
working_dir: "/vllm-workspace/tests"
|
||||
depends_on:
|
||||
- image-build-amd
|
||||
source_file_dependencies:
|
||||
- csrc/
|
||||
- vllm/model_executor/layers/quantization
|
||||
- vllm/model_executor/models/qwen3_dflash.py
|
||||
- vllm/model_executor/models/utils.py
|
||||
- tests/quantization
|
||||
- vllm/_aiter_ops.py
|
||||
- vllm/platforms/rocm.py
|
||||
- vllm/model_executor/layers/fused_moe/oracle/unquantized.py
|
||||
- vllm/model_executor/layers/fused_moe/unquantized_fused_moe_method.py
|
||||
- tests/rocm/test_moe_weight_replay.py
|
||||
- tests/rocm/test_moe_padding.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 --ignore quantization/test_rocm_moe.py
|
||||
- pytest -v -s rocm/test_moe_weight_replay.py
|
||||
- VLLM_ROCM_USE_AITER=1 VLLM_ROCM_USE_AITER_MOE=1 pytest -v -s rocm/test_moe_padding.py
|
||||
# Disabled until torchao transitive dependencies are locked.
|
||||
# - label: ":nvidia: (H200 MIG 35GB) Quantization"
|
||||
# device: h200_35gb
|
||||
# key: quantization
|
||||
# automatic_shard: true
|
||||
# timeout_in_minutes: 40
|
||||
# source_file_dependencies:
|
||||
# - csrc/
|
||||
# - vllm/model_executor/layers/quantization
|
||||
# - vllm/model_executor/models/qwen3_dflash.py
|
||||
# - vllm/model_executor/models/utils.py
|
||||
# - vllm/transformers_utils/model_arch_config_convertor.py
|
||||
# - vllm/models/deepseek_v41/quant_config.py
|
||||
# - vllm/utils/humming.py
|
||||
# - tests/quantization
|
||||
# commands:
|
||||
# # Pin torchao to the version compatible with the CI PyTorch and CUDA stack.
|
||||
# - uv pip install --system torchao==0.18.0 --index-url https://download.pytorch.org/whl/cu130
|
||||
# - uv pip install --system conch-triton-kernels
|
||||
# # The SM90-only checkpoint currently contains a removed weight_chan_scale
|
||||
# # 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 --ignore quantization/test_rocm_moe.py -k 'not test_compressed_tensors_w4a8_fp8'
|
||||
# mirror:
|
||||
# amd:
|
||||
# label: ":amd: (MI300) Quantization"
|
||||
# automatic_shard: true
|
||||
# dind: false
|
||||
# device: mi300_1
|
||||
# timeout_in_minutes: 115
|
||||
# working_dir: "/vllm-workspace/tests"
|
||||
# depends_on:
|
||||
# - image-build-amd
|
||||
# source_file_dependencies:
|
||||
# - csrc/
|
||||
# - vllm/model_executor/layers/quantization
|
||||
# - vllm/model_executor/models/qwen3_dflash.py
|
||||
# - vllm/model_executor/models/utils.py
|
||||
# - tests/quantization
|
||||
# - vllm/_aiter_ops.py
|
||||
# - vllm/platforms/rocm.py
|
||||
# - vllm/model_executor/layers/fused_moe/oracle/unquantized.py
|
||||
# - vllm/model_executor/layers/fused_moe/unquantized_fused_moe_method.py
|
||||
# - tests/rocm/test_moe_weight_replay.py
|
||||
# - tests/rocm/test_moe_padding.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 --ignore quantization/test_rocm_moe.py
|
||||
# - pytest -v -s rocm/test_moe_weight_replay.py
|
||||
# - VLLM_ROCM_USE_AITER=1 VLLM_ROCM_USE_AITER_MOE=1 pytest -v -s rocm/test_moe_padding.py
|
||||
|
||||
- label: ":nvidia: (H200 MIG 35GB) Quantized Fusions"
|
||||
device: h200_35gb
|
||||
|
||||
@@ -12,22 +12,23 @@ steps:
|
||||
- setup.py
|
||||
commands:
|
||||
- python3 /vllm-workspace/.buildkite/check-torch-abi.py
|
||||
mirror:
|
||||
amd:
|
||||
label: ":amd: (MI300) Torch Stable ABI Audit"
|
||||
dind: false
|
||||
device: mi300_1
|
||||
timeout_in_minutes: 35
|
||||
depends_on:
|
||||
- image-build-amd
|
||||
no_gpu: true
|
||||
working_dir: /vllm-workspace/
|
||||
source_file_dependencies:
|
||||
- .buildkite/check-torch-abi.py
|
||||
- csrc/
|
||||
- cmake/
|
||||
- setup.py
|
||||
- vllm/platforms/rocm.py
|
||||
commands:
|
||||
- uv pip install --system torch-abi-audit==0.0.1
|
||||
- python3 /vllm-workspace/.buildkite/check-torch-abi.py
|
||||
# Disabled until torch-abi-audit transitive dependencies are locked.
|
||||
# mirror:
|
||||
# amd:
|
||||
# label: ":amd: (MI300) Torch Stable ABI Audit"
|
||||
# dind: false
|
||||
# device: mi300_1
|
||||
# timeout_in_minutes: 35
|
||||
# depends_on:
|
||||
# - image-build-amd
|
||||
# no_gpu: true
|
||||
# working_dir: /vllm-workspace/
|
||||
# source_file_dependencies:
|
||||
# - .buildkite/check-torch-abi.py
|
||||
# - csrc/
|
||||
# - cmake/
|
||||
# - setup.py
|
||||
# - vllm/platforms/rocm.py
|
||||
# commands:
|
||||
# - uv pip install --system torch-abi-audit==0.0.1
|
||||
# - python3 /vllm-workspace/.buildkite/check-torch-abi.py
|
||||
@@ -57,3 +57,47 @@ tests/platforms/test_cpu.py # CPU-only unit test; no job targets tests/platform
|
||||
|
||||
# tests/tools/
|
||||
tests/tools/test_config_validator.py # logic covered by the validate-config pre-commit hook; no test_areas job
|
||||
|
||||
# Temporary gaps from disabling jobs until their transitive dependencies are locked.
|
||||
tests/evals/qwen4_exp/test_accuracy.py # CI jobs disabled pending locked evalscope transitive dependencies
|
||||
tests/kernels/helion/test_autotune.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_benchmark_script.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_case_key.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_config_manager.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_dynamic_per_token_scaled_fp8_quant.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_fused_qk_norm_rope.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_helion_available.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_pattern_matching.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_per_token_group_fp8_quant.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_register.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_rms_norm_dynamic_per_token_quant.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_rms_norm_per_block_quant.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_silu_and_mul_per_block_quant.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_silu_mul_fp8.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/kernels/helion/test_utils.py # CI jobs disabled pending locked helion transitive dependencies
|
||||
tests/quantization/test_auto_gptq.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_config_utils.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_configs.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_cpu_offload.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_experts_int8.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_fp8.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_gptq_dynamic.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_humming_compile.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_humming_ignore.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_humming_mxfp4_block_fp8.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_int8_moe_oracle.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_keep_param_objects.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_lm_head.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_mixed_precision.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_modelopt.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_moe_wna16.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_online_mxfp4.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_online_shared_expert_fusion.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_quantization_config_args.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_quark.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_register_quantization_config.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_torchao.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_trtllm_nvfp4_hidden_dim_padding.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/quantization/test_turboquant.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/rocm/test_moe_padding.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
tests/rocm/test_moe_weight_replay.py # CI jobs disabled pending locked torchao transitive dependencies
|
||||
Reference in new issue
Block a user