EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[Test] Add nightly test for DeepSeek-V3.2-Exp (#3908)

Browse Source 
### What this PR does / why we need it?
Add nightly test for DeepSeek-V3.2-Exp


### How was this patch tested?
test action：

https://github.com/vllm-project/vllm-ascend/actions/runs/19156153634/job/54757008557?pr=3908


- vLLM version: v0.11.0
- vLLM main:
83f478bb19

---------

Signed-off-by: hfadzxy <starmoon_zhang@163.com>
This commit is contained in:
zhangxinyuehfad
2025-11-11 10:29:57 +08:00
committed by GitHub
parent e384755ce1
commit b77b4f1abf
5 changed files with 228 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

55
tests/e2e/nightly/multi_node/config/models/DeepSeek-V3_2-Exp-bf16.yaml Normal file
View File

@@ -0,0 +1,55 @@
test_name: "test DeepSeek-V3.2-Exp-bf16 multi-dp"
model: "Yanguan/DeepSeek-V3.2-Exp-bf16"
num_nodes: 2
npu_per_node: 16
env_common:
VLLM_USE_MODELSCOPE: true
OMP_PROC_BIND: false
OMP_NUM_THREADS: 100
HCCL_BUFFSIZE: 1024
SERVER_PORT: 8080
VLLM_ASCEND_ENABLE_MLAPO: 0
deployment:
-
server_cmd: >
vllm serve Yanguan/DeepSeek-V3.2-Exp-bf16 \
--host 0.0.0.0
--port $SERVER_PORT
--data-parallel-address $LOCAL_IP
--data-parallel-size 2
--data-parallel-size-local 1
--data-parallel-rpc-port 13389
--tensor-parallel-size 16
--seed 1024
--enable-expert-parallel
--max-num-seqs 16
--max-model-len 17450
--max-num-batched-tokens 17450
--trust-remote-code
--no-enable-prefix-caching
--gpu-memory-utilization 0.9
--additional-config '{"ascend_scheduler_config":{"enabled":true},"torchair_graph_config":{"enabled":true,"graph_batch_sizes":[16]}}'
-
server_cmd: >
vllm serve Yanguan/DeepSeek-V3.2-Exp-bf16 \
--host 0.0.0.0
--port $SERVER_PORT
--headless
--data-parallel-size 2
--data-parallel-size-local 1
--data-parallel-start-rank 1
--data-parallel-address $MASTER_IP
--data-parallel-rpc-port 13389
--tensor-parallel-size 16
--seed 1024
--max-num-seqs 16
--max-model-len 17450
--max-num-batched-tokens 17450
--enable-expert-parallel
--trust-remote-code
--no-enable-prefix-caching
--gpu-memory-utilization 0.92
--additional-config '{"ascend_scheduler_config":{"enabled":true},"torchair_graph_config":{"enabled":true,"graph_batch_sizes":[16]}}'
benchmarks:

31
tests/e2e/nightly/multi_node/scripts/run.sh
View File

@@ -107,6 +107,32 @@ install_ais_bench() {
cd -
}
install_extra_components() {
echo "====> Installing extra components for DeepSeek-v3.2-exp-bf16"
if ! wget -q https://vllm-ascend.obs.cn-north-4.myhuaweicloud.com/vllm-ascend/a3/CANN-custom_ops-sfa-linux.aarch64.run; then
echo "Failed to download CANN-custom_ops-sfa-linux.aarch64.run"
return 1
fi
chmod +x ./CANN-custom_ops-sfa-linux.aarch64.run
./CANN-custom_ops-sfa-linux.aarch64.run --quiet
if ! wget -q https://vllm-ascend.obs.cn-north-4.myhuaweicloud.com/vllm-ascend/a3/custom_ops-1.0-cp311-cp311-linux_aarch64.whl; then
echo "Failed to download custom_ops wheel"
return 1
fi
pip install custom_ops-1.0-cp311-cp311-linux_aarch64.whl
export ASCEND_CUSTOM_OPP_PATH=/usr/local/Ascend/ascend-toolkit/latest/opp/vendors/customize:${ASCEND_CUSTOM_OPP_PATH}
export LD_LIBRARY_PATH=/usr/local/Ascend/ascend-toolkit/latest/opp/vendors/customize/op_api/lib/:${LD_LIBRARY_PATH}
source /usr/local/Ascend/ascend-toolkit/set_env.sh
EOF
rm -f CANN-custom_ops-sfa-linux.aarch64.run \
custom_ops-1.0-cp311-cp311-linux_aarch64.whl
echo "====> Extra components installation completed"
}
kill_npu_processes() {
pgrep python3 | xargs -r kill -9
pgrep VLLM | xargs -r kill -9
@@ -140,6 +166,9 @@ main() {
checkout_src
install_sys_dependencies
install_vllm
if [[ "$CONFIG_YAML_PATH" == *"DeepSeek-V3_2-Exp-bf16.yaml" ]]; then
install_extra_components
fi
install_ais_bench
cd "$WORKSPACE/source_code"
. $SRC_DIR/vllm-ascend/tests/e2e/nightly/multi_node/scripts/build_mooncake.sh
@@ -147,4 +176,4 @@ main() {
run_tests_with_log
}
main "$@"
main "$@"