biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Hardware][TPU] Add check for no additional graph compilation during runtime (#14710)

Browse Source 
Signed-off-by: Siyuan Liu <lsiyuan@google.com>
This commit is contained in:
Siyuan Liu
2025-03-20 20:05:28 -07:00
committed by GitHub
parent e588ac237c
commit b15fd2be2a
3 changed files with 32 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
.buildkite/run-tpu-v1-test.sh
View File

@@ -19,17 +19,19 @@ docker run --privileged --net host --shm-size=16G -it \
vllm-tpu /bin/bash -c "python3 -m pip install git+https://github.com/thuml/depyf.git \
&& python3 -m pip install pytest \
&& python3 -m pip install lm_eval[api]==0.4.4 \
&& export VLLM_USE_V1=1 \
&& export VLLM_XLA_CHECK_RECOMPILATION=1 \
&& echo TEST_1 \
&& VLLM_USE_V1=1 python3 /workspace/vllm/tests/tpu/test_compilation.py \
&& python3 /workspace/vllm/tests/tpu/test_compilation.py \
&& echo TEST_2 \
&& VLLM_USE_V1=1 pytest -v -s /workspace/vllm/tests/v1/tpu/test_basic.py \
&& pytest -v -s /workspace/vllm/tests/v1/tpu/test_basic.py \
&& echo TEST_3 \
&& VLLM_USE_V1=1 pytest -v -s /workspace/vllm/tests/entrypoints/llm/test_accuracy.py::test_lm_eval_accuracy_v1_engine \
&& pytest -v -s /workspace/vllm/tests/entrypoints/llm/test_accuracy.py::test_lm_eval_accuracy_v1_engine \
&& echo TEST_4 \
&& VLLM_USE_V1=1 pytest -s -v /workspace/vllm/tests/tpu/test_quantization_accuracy.py \
&& pytest -s -v /workspace/vllm/tests/tpu/test_quantization_accuracy.py \
&& echo TEST_5 \
&& VLLM_USE_V1=1 python3 /workspace/vllm/examples/offline_inference/tpu.py" \
&& python3 /workspace/vllm/examples/offline_inference/tpu.py" \
# TODO: This test fails because it uses RANDOM_SEED sampling
# && VLLM_USE_V1=1 pytest -v -s /workspace/vllm/tests/tpu/test_custom_dispatcher.py \