Back to Details

perf-benchmarker

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

perf-benchmarker

perf-benchmarker

Run sequential benchmarks with strict duration rules.
Follow 
docs/perf-requirements.md
as the canonical contract.
运行遵循严格时长规则的连续基准测试。
请以
docs/perf-requirements.md
作为标准约定。

Required Rules

强制规则

  • Benchmarks MUST run sequentially (never parallel).
  • Minimum duration: 60s per run (30s only for binary search).
  • Warmup: 10s minimum before measurement.
  • Re-run anomalies.
  • 基准测试必须连续运行（绝不能并行）。
  • 最短时长：单次运行60秒（仅Binary Search场景为30秒）。
  • 预热：测量前至少预热10秒。
  • 异常情况需重新运行。

Output Format

输出格式

command: <benchmark command>
duration: <seconds>
warmup: <seconds>
results: <metrics summary>
notes: <anomalies or reruns>
command: <benchmark command>
duration: <seconds>
warmup: <seconds>
results: <metrics summary>
notes: <anomalies or reruns>

Output Contract

输出约定

Benchmarks MUST emit a JSON metrics block between markers:
PERF_METRICS_START
{"scenarios":{"low":{"latency_ms":120},"high":{"latency_ms":450}}}
PERF_METRICS_END
基准测试必须在标记之间输出JSON指标块：
PERF_METRICS_START
{"scenarios":{"low":{"latency_ms":120},"high":{"latency_ms":450}}}
PERF_METRICS_END

Constraints

约束条件

  • No short runs unless binary-search phase.
  • Do not change code while benchmarking.
  • 除非处于Binary Search阶段，否则不允许短时长运行。
  • 基准测试期间不得修改代码。