feat: add ShareGPT dataset scenarios to SGLang vs vLLM comparison
- config.env: add SHAREGPT_DATASET, SHAREGPT_CONTEXT_LEN and SHAREGPT_SCENARIOS
- run_bench.sh: add run_sharegpt() and integrate it after Phase 1 for each backend
- parse_backend.py: support both phase{1|2} and sharegpt raw output filenames
- README.md: document ShareGPT scenarios
This commit is contained in:
yy-fighting
parent
e6d9e99b6f
commit
06b3398ad0
@ -31,6 +31,15 @@
|
||||
| 1 | 131072 | 1024 | 2 |
|
||||
| 1 | 200000 | 1024 | 1 |
|
||||
|
||||
### ShareGPT 真实对话数据（max_model_len=32k）
|
||||
|
||||
| Concurrency | ShareGPT context len | Output | Num prompts |
|
||||
|---:|---:|---:|---:|
|
||||
| 1 | 32768 | 256 | 32 |
|
||||
| 32 | 32768 | 256 | 128 |
|
||||
|
||||
使用 `datasets/ShareGPT_filtered_chat.json`，`--sharegpt-context-len 32768` 会自动过滤超长请求。
|
||||
|
||||
## 快速运行
|
||||
|
||||
```bash
|
||||
|
||||
@ -46,6 +46,16 @@ PHASE2_MAX_MODEL_LEN=210000
|
||||
PHASE2_MAX_NUM_SEQS=2
|
||||
PHASE2_MAX_RUNNING=2
|
||||
|
||||
# ShareGPT dataset test (uses the same max_model_len as Phase 1).
|
||||
SHAREGPT_DATASET="${SHAREGPT_DATASET:-/data/user1/yy/datasets/ShareGPT_filtered_chat.json}"
|
||||
SHAREGPT_CONTEXT_LEN=32768
|
||||
if [[ -z "${SHAREGPT_SCENARIOS:-}" ]]; then
|
||||
declare -a SHAREGPT_SCENARIOS=(
|
||||
"1 32768 256 32"
|
||||
"32 32768 256 128"
|
||||
)
|
||||
fi
|
||||
|
||||
# Benchmark client always runs from the sglang env so that the same
|
||||
# sglang.bench_serving version is used for both backends.
|
||||
VENV_CLIENT="${VENV_CLIENT:-$VENV_SGLANG}"
|
||||
|
||||
@ -140,16 +140,27 @@ def main() -> None:
|
||||
if backend is None:
|
||||
raise SystemExit("Could not infer backend from raw outputs")
|
||||
|
||||
# Filename: {backend}_phase{1|2}_MMDD_concurrency_inputlen_outputlen.jsonl
|
||||
pattern = re.compile(rf"^{backend}_phase(\d+)_(\d+)_(\d+)_(\d+)_(\d+)\.jsonl$")
|
||||
|
||||
scenarios = []
|
||||
for jsonl_path in sorted(raw_dir.glob(f"{backend}_phase*.jsonl")):
|
||||
m = pattern.match(jsonl_path.name)
|
||||
if not m:
|
||||
for jsonl_path in sorted(raw_dir.glob(f"{backend}_*.jsonl")):
|
||||
parts = jsonl_path.stem.split("_")
|
||||
if len(parts) < 5:
|
||||
continue
|
||||
|
||||
label = parts[1]
|
||||
if label.startswith("phase"):
|
||||
phase = label
|
||||
dataset = "random"
|
||||
elif label == "sharegpt":
|
||||
phase = "sharegpt"
|
||||
dataset = "sharegpt"
|
||||
else:
|
||||
continue
|
||||
|
||||
# The last three numeric tokens are: concurrency, input_len/output_ctx, output_len.
|
||||
try:
|
||||
concurrency, input_len, output_len = int(parts[-3]), int(parts[-2]), int(parts[-1])
|
||||
except ValueError:
|
||||
continue
|
||||
phase_num, concurrency, input_len, output_len = m.groups()
|
||||
concurrency, input_len, output_len = int(concurrency), int(input_len), int(output_len)
|
||||
|
||||
data = parse_jsonl(jsonl_path)
|
||||
if data is None:
|
||||
@ -159,11 +170,11 @@ def main() -> None:
|
||||
scenario = {
|
||||
"name": scenario_name(concurrency, input_len, output_len),
|
||||
"config": {
|
||||
"phase": f"phase{phase_num}",
|
||||
"phase": phase,
|
||||
"concurrency": concurrency,
|
||||
"input_len": input_len,
|
||||
"output_len": output_len,
|
||||
"dataset": "random",
|
||||
"dataset": dataset,
|
||||
"num_prompts": metrics["success"] + metrics["failed"],
|
||||
},
|
||||
"metrics": metrics,
|
||||
|
||||
@ -174,6 +174,60 @@ run_phase() {
|
||||
log "===== ${backend} ${phase} DONE ====="
|
||||
}
|
||||
|
||||
run_sharegpt() {
|
||||
local backend="$1"
|
||||
local phase="sharegpt"
|
||||
local result_root="${RESULT_BASE}/${RUN_ID}/${backend}"
|
||||
local raw_dir="${result_root}/raw_outputs"
|
||||
local phase_log_dir="${result_root}/logs"
|
||||
|
||||
mkdir -p "$raw_dir" "$phase_log_dir"
|
||||
|
||||
local port
|
||||
if [[ "$backend" == "sglang" ]]; then
|
||||
port="$SGLANG_PORT"
|
||||
else
|
||||
port="$VLLM_PORT"
|
||||
fi
|
||||
|
||||
log "===== ${backend} ${phase} START (context_len=${SHAREGPT_CONTEXT_LEN}) ====="
|
||||
|
||||
stop_server "$backend"
|
||||
start_server "$backend" phase1
|
||||
run_warmup "$backend" phase1
|
||||
|
||||
for scenario in "${SHAREGPT_SCENARIOS[@]}"; do
|
||||
read -r concurrency context_len output_len num_prompts <<< "$scenario"
|
||||
output_file="${raw_dir}/${backend}_${phase}_$(date '+%m%d')_${concurrency}_${context_len}_${output_len}.jsonl"
|
||||
detail_log="${phase_log_dir}/${backend}_${phase}_c${concurrency}_ctx${context_len}_o${output_len}.log"
|
||||
|
||||
log "running ${backend} ${phase} scenario: c=${concurrency} ctx=${context_len} o=${output_len} n=${num_prompts}"
|
||||
|
||||
"${VENV_CLIENT}/bin/python" -m sglang.bench_serving \
|
||||
--backend "$backend" \
|
||||
--host 127.0.0.1 \
|
||||
--port "$port" \
|
||||
--dataset-name sharegpt \
|
||||
--dataset-path "$SHAREGPT_DATASET" \
|
||||
--sharegpt-output-len "$output_len" \
|
||||
--sharegpt-context-len "$context_len" \
|
||||
--num-prompts "$num_prompts" \
|
||||
--max-concurrency "$concurrency" \
|
||||
--request-rate 10000 \
|
||||
--output-file "$output_file" \
|
||||
--output-details \
|
||||
> "$detail_log" 2>&1 || {
|
||||
log "ERROR: ${backend} ${phase} scenario c=${concurrency} ctx=${context_len} o=${output_len} failed; see ${detail_log}"
|
||||
continue
|
||||
}
|
||||
|
||||
log "finished ${backend} ${phase} scenario: output=${output_file}"
|
||||
done
|
||||
|
||||
stop_server "$backend"
|
||||
log "===== ${backend} ${phase} DONE ====="
|
||||
}
|
||||
|
||||
parse_backend() {
|
||||
local backend="$1"
|
||||
local result_root="${RESULT_BASE}/${RUN_ID}/${backend}"
|
||||
@ -245,11 +299,13 @@ write_backend_metadata vllm
|
||||
|
||||
# Run SGLang.
|
||||
run_phase sglang phase1
|
||||
run_sharegpt sglang
|
||||
run_phase sglang phase2
|
||||
parse_backend sglang
|
||||
|
||||
# Run vLLM.
|
||||
run_phase vllm phase1
|
||||
run_sharegpt vllm
|
||||
run_phase vllm phase2
|
||||
parse_backend vllm
|
||||
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user