A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 23:45 UTC
[XPU][UT] increase intel xpu CI test scope (#21492)
Signed-off-by: Ma, Liangliang <liangliang.ma@intel.com>
Liangliang Ma committed
Jul 24, 2025 at 11:24 UTC
13e4ee1dc3976bbc78cbb9e406523dd7d38403d0
3 files changed
+13
-3
.buildkite/scripts/hardware_ci/run-xpu-test.sh
+9
@@ -31,4 +31,13 @@ docker run \
31
VLLM_USE_V1=1 python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager -tp 2 --distributed-executor-backend mp
32
cd tests
33
pytest -v -s v1/core
34
+ pytest -v -s v1/engine
35
+ pytest -v -s v1/sample --ignore=v1/sample/test_logprobs.py --ignore=v1/sample/test_logprobs_e2e.py
36
+ pytest -v -s v1/worker --ignore=v1/worker/test_gpu_model_runner.py
37
+ pytest -v -s v1/structured_output
38
+ pytest -v -s v1/spec_decode --ignore=v1/spec_decode/test_max_len.py --ignore=v1/spec_decode/test_eagle.py
39
+ pytest -v -s v1/kv_connector/unit --ignore=v1/kv_connector/unit/test_multi_connector.py --ignore=v1/kv_connector/unit/test_nixl_connector.py
40
+ pytest -v -s v1/test_serial_utils.py
41
+ pytest -v -s v1/test_utils.py
42
+ pytest -v -s v1/test_metrics_reader.py
43
'
docker/Dockerfile.xpu
+1
-1
@@ -47,7 +47,7 @@ FROM vllm-base AS vllm-openai
47
48
# install additional dependencies for openai api server
49
RUN --mount=type=cache,target=/root/.cache/pip \
50
- pip install accelerate hf_transfer pytest modelscope
50
+ pip install accelerate hf_transfer pytest pytest_asyncio lm_eval[api] modelscope
51
52
ENV VLLM_USAGE_SOURCE production-docker-image \
53
TRITON_XPU_PROFILE 1
tests/entrypoints/openai/correctness/test_lmeval.py
+3
-2
@@ -69,8 +69,9 @@ def run_test(more_args):
69
70
71
@pytest.mark.skipif(not current_platform.is_cuda()
72
- and not current_platform.is_tpu(),
73
- reason="V1 currently only supported on CUDA and TPU")
72
+ and not current_platform.is_tpu()
73
+ and not current_platform.is_xpu(),
74
+ reason="V1 currently only supported on CUDA, XPU and TPU")
75
def test_lm_eval_accuracy_v1_engine(monkeypatch: pytest.MonkeyPatch):
76
"""Run with the V1 Engine."""
77