gahow/aituner
0
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Extend Qwen235 smoke readiness deadline

Browse Source
This commit is contained in:
Gahow Wang
2026-07-18 10:32:21 +08:00
parent 7a437b4d91
commit 5fc0b48fa1
1 changed files with 8 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
runs/frontier-fidelity-envelope-v1/run_qwen235_v020_smoke.sh
View File

@@ -13,6 +13,7 @@ ENABLE_EXPERT_PARALLEL="${ENABLE_EXPERT_PARALLEL:?ENABLE_EXPERT_PARALLEL is requ
VENV_ROOT="${VENV_ROOT:-/tmp/wjh/venvs/vllm-0.20.0-cu129-profiler-v1}"
MODEL_ROOT="${MODEL_ROOT:-/home/admin/cpfs/wjh/models/Qwen/Qwen3-235B-A22B-FP8}"
CLIENT="${CLIENT:?CLIENT is required}"
READY_TIMEOUT_SECONDS="${READY_TIMEOUT_SECONDS:-1800}"
SERVED_MODEL="qwen3-235b-v020-smoke"
SERVER_PID=""
@@ -42,8 +43,12 @@ case "${ENABLE_EXPERT_PARALLEL}" in
false) EP_FLAG=() ;;
*) echo "ERROR: ENABLE_EXPERT_PARALLEL must be true or false" >&2; exit 1 ;;
esac
if ! [[ "${READY_TIMEOUT_SECONDS}" =~ ^[1-9][0-9]*$ ]]; then
echo "ERROR: READY_TIMEOUT_SECONDS must be a positive integer" >&2
exit 1
fi
echo "QWEN235_V020_SMOKE_LAUNCH_ECHO host=$(hostname) gpus=${CUDA_VISIBLE_DEVICES} model=${MODEL_ROOT} runtime=vLLM-0.20.0+cu129 checkpoint=FP8_compute_kv=auto config=TP${TP}_EP${ENABLE_EXPERT_PARALLEL}_MNS${MNS}_MBT${MBT} prefix=false requests=1 shape=2048x1 role=compatibility_gate_not_latency_measurement output=${OUT} expected_wall=8-20m expected_gpu_cap=$((${TP} * 1))_H20h"
echo "QWEN235_V020_SMOKE_LAUNCH_ECHO host=$(hostname) gpus=${CUDA_VISIBLE_DEVICES} model=${MODEL_ROOT} runtime=vLLM-0.20.0+cu129 checkpoint=FP8_compute_kv=auto config=TP${TP}_EP${ENABLE_EXPERT_PARALLEL}_MNS${MNS}_MBT${MBT} prefix=false requests=1 shape=2048x1 ready_timeout=${READY_TIMEOUT_SECONDS}s role=compatibility_gate_not_latency_measurement output=${OUT} expected_wall=20-35m expected_gpu_cap=$((${TP} * 35 / 60))_H20h"
date -u +START_UTC=%Y-%m-%dT%H:%M:%SZ
sha256sum "${BASH_SOURCE[0]}" "${CLIENT}" "${MODEL_ROOT}/config.json" > "${OUT}/provenance/input.sha256"
"${VENV_ROOT}/bin/vllm" --version > "${OUT}/provenance/vllm.version"
@@ -71,7 +76,8 @@ setsid "${VENV_ROOT}/bin/vllm" serve "${MODEL_ROOT}" \
"${EP_FLAG[@]}" > "${OUT}/logs/server.log" 2>&1 &
SERVER_PID=$!
READY=0
for _ in $(seq 1 180); do
READY_DEADLINE=$((SECONDS + READY_TIMEOUT_SECONDS))
while (( SECONDS < READY_DEADLINE )); do
if curl -fsS --max-time 2 "http://127.0.0.1:${SERVER_PORT}/v1/models" > "${OUT}/results/models.json" 2>/dev/null; then
READY=1
break