[CI][XPU] Work around intermittent segfault in Intel XPU CI with VLLM_DISABLE_COMPILE_CACHE=1 (#51337)

Signed-off-by: Chaojun Zhang <chaojun.zhang@intel.com>
This commit is contained in:
Chaojun Zhang
2026-08-07 05:18:25 +00:00
committed by GitHub
parent c810e5ee99
commit b706fd1628
5 changed files with 10 additions and 5 deletions
+2 -1
View File
@@ -18,10 +18,11 @@ steps:
source_file_dependencies:
- vllm/
- tests/basic_correctness/test_cumem.py
# Remove VLLM_DISABLE_COMPILE_CACHE=1 once intel/intel-xpu-backend-for-triton#7682 is fixed
commands:
- >-
bash .buildkite/scripts/hardware_ci/run-intel-test.sh
'cd tests &&
export VLLM_WORKER_MULTIPROC_METHOD=spawn &&
pytest -v -s basic_correctness/test_cpu_offload.py &&
VLLM_DISABLE_COMPILE_CACHE=1 pytest -v -s basic_correctness/test_cpu_offload.py &&
pytest -v -s basic_correctness/test_mem.py::test_end_to_end'
@@ -21,6 +21,7 @@ steps:
- vllm/config/model.py
- vllm/model_executor
- tests/model_executor
# Remove VLLM_DISABLE_COMPILE_CACHE=1 once intel/intel-xpu-backend-for-triton#7682 is fixed
commands:
- >-
bash .buildkite/scripts/hardware_ci/run-intel-test.sh
@@ -30,4 +31,4 @@ steps:
export VLLM_WORKER_MULTIPROC_METHOD=spawn &&
export PYTHONFAULTHANDLER=1 &&
cd tests &&
pytest -v -s model_executor -m "not slow_test" --ignore="model_executor/layers/test_rocm_unquantized_gemm.py" --deselect="tests/model_executor/model_loader/test_reload.py::test_kv_scale_reload"'
VLLM_DISABLE_COMPILE_CACHE=1 pytest -v -s model_executor -m "not slow_test" --ignore="model_executor/layers/test_rocm_unquantized_gemm.py" --deselect="tests/model_executor/model_loader/test_reload.py::test_kv_scale_reload"'
@@ -22,13 +22,14 @@ steps:
- vllm/v1/attention/
- tests/v1/engine/test_llm_engine.py
- tests/v1/e2e/
# Remove VLLM_DISABLE_COMPILE_CACHE=1 once intel/intel-xpu-backend-for-triton#7682 is fixed
commands:
- >-
bash .buildkite/scripts/hardware_ci/run-intel-test.sh
'export VLLM_USE_V2_MODEL_RUNNER=1 &&
cd tests &&
pytest -v -s v1/engine/test_llm_engine.py -k "not test_engine_metrics" &&
pytest -v -s v1/e2e/general/test_context_length.py &&
VLLM_DISABLE_COMPILE_CACHE=1 pytest -v -s v1/e2e/general/test_context_length.py &&
ENFORCE_EAGER=1 pytest -v -s v1/e2e/general/test_async_scheduling.py -k "not ngram" &&
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" &&
pytest -v -s v1/e2e/general/test_min_tokens.py'
@@ -20,8 +20,9 @@ steps:
- vllm/model_executor/model_loader/sharded_state_loader.py
- vllm/model_executor/models/
- tests/model_executor/model_loader/test_sharded_state_loader.py
# Remove VLLM_DISABLE_COMPILE_CACHE=1 once intel/intel-xpu-backend-for-triton#7682 is fixed
commands:
- >-
bash .buildkite/scripts/hardware_ci/run-intel-test.sh
'cd tests &&
pytest -v -s model_executor/model_loader/test_sharded_state_loader.py -m "not slow_test"'
VLLM_DISABLE_COMPILE_CACHE=1 pytest -v -s model_executor/model_loader/test_sharded_state_loader.py -m "not slow_test"'
+2 -1
View File
@@ -22,8 +22,9 @@ steps:
- tests/samplers
- tests/conftest.py
- vllm/entrypoints/generate/beam_search
# Remove VLLM_DISABLE_COMPILE_CACHE=1 once intel/intel-xpu-backend-for-triton#7682 is fixed
commands:
- >-
bash .buildkite/scripts/hardware_ci/run-intel-test.sh
'cd tests &&
VLLM_XPU_USE_SAMPLER_KERNEL=1 pytest -v -s samplers'
VLLM_XPU_USE_SAMPLER_KERNEL=1 VLLM_DISABLE_COMPILE_CACHE=1 pytest -v -s samplers'