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

Pin DeepGEMM nvcc for Qwen235 smoke

Browse Source
This commit is contained in:
Gahow Wang
2026-07-18 11:15:22 +08:00
parent b7f9cef9c5
commit b9523cef5c
1 changed files with 7 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
runs/frontier-fidelity-envelope-v1/run_qwen235_v020_smoke.sh
View File

@@ -15,6 +15,7 @@ MODEL_ROOT="${MODEL_ROOT:-/home/admin/cpfs/wjh/models/Qwen/Qwen3-235B-A22B-FP8}"
CLIENT="${CLIENT:?CLIENT is required}" CLIENT="${CLIENT:?CLIENT is required}"
READY_TIMEOUT_SECONDS="${READY_TIMEOUT_SECONDS:-1800}" READY_TIMEOUT_SECONDS="${READY_TIMEOUT_SECONDS:-1800}"
TRTLLM_DG_CACHE_DIR="${TRTLLM_DG_CACHE_DIR:-${OUT}/deep_gemm}" TRTLLM_DG_CACHE_DIR="${TRTLLM_DG_CACHE_DIR:-${OUT}/deep_gemm}"
TRTLLM_DG_NVCC_COMPILER="${TRTLLM_DG_NVCC_COMPILER:-/usr/local/cuda/bin/nvcc}"
DEEP_GEMM_JIT_DEBUG="${DEEP_GEMM_JIT_DEBUG:-false}" DEEP_GEMM_JIT_DEBUG="${DEEP_GEMM_JIT_DEBUG:-false}"
SERVED_MODEL="qwen3-235b-v020-smoke" SERVED_MODEL="qwen3-235b-v020-smoke"
SERVER_PID="" SERVER_PID=""
@@ -49,12 +50,16 @@ if ! [[ "${READY_TIMEOUT_SECONDS}" =~ ^[1-9][0-9]*$ ]]; then
echo "ERROR: READY_TIMEOUT_SECONDS must be a positive integer" >&2 echo "ERROR: READY_TIMEOUT_SECONDS must be a positive integer" >&2
exit 1 exit 1
fi fi
if [[ ! -x "${TRTLLM_DG_NVCC_COMPILER}" ]]; then
echo "ERROR: TRTLLM_DG_NVCC_COMPILER is not executable: ${TRTLLM_DG_NVCC_COMPILER}" >&2
exit 1
fi
case "${DEEP_GEMM_JIT_DEBUG}" in case "${DEEP_GEMM_JIT_DEBUG}" in
true|false) ;; true|false) ;;
*) echo "ERROR: DEEP_GEMM_JIT_DEBUG must be true or false" >&2; exit 1 ;; *) echo "ERROR: DEEP_GEMM_JIT_DEBUG must be true or false" >&2; exit 1 ;;
esac esac
echo "QWEN235_V020_SMOKE_LAUNCH_ECHO host=$(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=${READY_TIMEOUT_SECONDS}s deep_gemm_cache=${TRTLLM_DG_CACHE_DIR} deep_gemm_jit_debug=${DEEP_GEMM_JIT_DEBUG} role=compatibility_gate_not_latency_measurement output=${OUT} expected_wall=20-35m expected_gpu_cap=$((${TP} * 35 / 60))_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 deep_gemm_cache=${TRTLLM_DG_CACHE_DIR} deep_gemm_nvcc=${TRTLLM_DG_NVCC_COMPILER} deep_gemm_jit_debug=${DEEP_GEMM_JIT_DEBUG} role=compatibility_gate_not_latency_measurement output=${OUT} expected_wall=20-35m expected_gpu_cap=$((${TP} * 35 / 60))_H20h"
date -u +START_UTC=%Y-%m-%dT%H:%M:%SZ date -u +START_UTC=%Y-%m-%dT%H:%M:%SZ
sha256sum "${BASH_SOURCE[0]}" "${CLIENT}" "${MODEL_ROOT}/config.json" > "${OUT}/provenance/input.sha256" sha256sum "${BASH_SOURCE[0]}" "${CLIENT}" "${MODEL_ROOT}/config.json" > "${OUT}/provenance/input.sha256"
"${VENV_ROOT}/bin/vllm" --version > "${OUT}/provenance/vllm.version" "${VENV_ROOT}/bin/vllm" --version > "${OUT}/provenance/vllm.version"
@@ -74,6 +79,7 @@ export HOME=/tmp/wjh
export XDG_CACHE_HOME=/tmp/wjh/.cache export XDG_CACHE_HOME=/tmp/wjh/.cache
export VLLM_CACHE_ROOT=/tmp/wjh/.cache/vllm export VLLM_CACHE_ROOT=/tmp/wjh/.cache/vllm
export TRTLLM_DG_CACHE_DIR export TRTLLM_DG_CACHE_DIR
export TRTLLM_DG_NVCC_COMPILER
if [[ "${DEEP_GEMM_JIT_DEBUG}" == true ]]; then if [[ "${DEEP_GEMM_JIT_DEBUG}" == true ]]; then
export TRTLLM_DG_JIT_DEBUG=1 export TRTLLM_DG_JIT_DEBUG=1
fi fi