55 lines
1.9 KiB
Bash
55 lines
1.9 KiB
Bash
# TP×DP matrix experiment for DeepSeek-V4-Flash on H200 (8 GPUs).
|
||
# Tests vLLM with three parallel configurations:
|
||
# TP=2, DP=4 -> 2 GPUs per replica, 4 replicas
|
||
# TP=4, DP=2 -> 4 GPUs per replica, 2 replicas
|
||
# TP=8, DP=1 -> 8 GPUs, no data parallelism
|
||
|
||
EXPERIMENT="dsv4_h200_vllm_tp_dp_matrix"
|
||
MODEL_NAME="DeepSeek-V4-Flash"
|
||
MODEL_PATH="/data/models/DeepSeek-V4-Flash"
|
||
SERVED_MODEL_NAME="deepseek-v4-flash"
|
||
|
||
VLLM_PORT="${VLLM_PORT:-30030}"
|
||
|
||
VENV_VLLM="${VENV_VLLM:-/data/user1/yy/envs/vllm}"
|
||
VENV_SGLANG="${VENV_SGLANG:-/data/user1/yy/envs/sglang}"
|
||
# The benchmark client is sglang.bench_serving, even when the backend is vLLM.
|
||
VENV_CLIENT="${VENV_CLIENT:-$VENV_SGLANG}"
|
||
|
||
export CUDA_VISIBLE_DEVICES="${CUDA_VISIBLE_DEVICES:-0,1,2,3,4,5,6,7}"
|
||
|
||
# Parallel configurations to test. Format: "TP DP"
|
||
declare -a PARALLEL_CONFIGS=(
|
||
"2 4"
|
||
"4 2"
|
||
"8 1"
|
||
)
|
||
|
||
# Matrix and concurrency rules are defined in matrix.json by default.
|
||
# This file is consumed by generate_scenarios.py.
|
||
MATRIX_FILE="${MATRIX_FILE:-${SCRIPT_DIR:-.}/matrix.json}"
|
||
MATRIX_MODE="${MATRIX_MODE:-Y}"
|
||
|
||
# Sampling density for concurrency.
|
||
# 0 = use the default heuristic in generate_scenarios.py (6-8 points).
|
||
# 2 = only test the two endpoints (low and high).
|
||
export CONCURRENCY_SAMPLES="${CONCURRENCY_SAMPLES:-2}"
|
||
|
||
# Per-scenario timeout to avoid hangs (seconds).
|
||
SCENARIO_TIMEOUT_S="${SCENARIO_TIMEOUT_S:-1800}"
|
||
|
||
# GPU memory sampling interval (seconds).
|
||
GPU_MEM_SAMPLE_INTERVAL_S="${GPU_MEM_SAMPLE_INTERVAL_S:-1}"
|
||
|
||
# Dry-run mode: if 1, only log the server args and scenario plan without starting
|
||
# any server or sending requests.
|
||
DRY_RUN="${DRY_RUN:-0}"
|
||
|
||
# Per-config scenario limit for quick smoke tests. 0 = run all generated scenarios.
|
||
GRID_LIMIT="${GRID_LIMIT:-0}"
|
||
|
||
# Docker deployment switch. If 1, vLLM is launched via Docker using the image
|
||
# below instead of the local VENV_VLLM environment.
|
||
USE_DOCKER="${USE_DOCKER:-0}"
|
||
DOCKER_IMAGE="${DOCKER_IMAGE:-vllm/vllm-openai:latest}"
|