douqiaolong0528 2018-11-15 15:39
浏览 499
已采纳

可以限制Go Benchmarking中的迭代次数(b.N)吗?

I'm trying to benchmark an operation which requires expensive preparation which I'm excluding using StopTimer() and StartTimer(). Specifically, I'm benchmarking nth-item insertion into a sorted list.

Sample code:

n := 100

// Run the process b.N times
for i := 0; i < b.N; i++ {

    // Stop the timer for our expensive preparation work (inserting n-1 items)
    b.StopTimer()

    // ...

    // Insert n-1 items
    for j := 1; j < n; j++ {
        m.InsertItem(o)
    }

    // ...

    // Resume the timer
    b.StartTimer()

    // Insert the nth item
    m.InsertItem(o)

}

The problem is that Go's benchmarking heuristic is limiting b.N according to the benchmarked time, not the total time. It ends up asking for 5MM (5000000) iterations of 100th order insertion, and this takes more time than is reasonable (I want to benchmark up to the 10 millionth item insertion).

Is there a way to specify a maximum b.N for a specific benchmark in Go's benchmark tool? I did not find anything in the docs myself.

  • 写回答

1条回答 默认 最新

  • dongzhimin2231 2018-11-15 16:22
    关注

    -benchtime flag can be used to specify the time for which you want to run the benchmark(source). Although I haven't tried it, providing a lower value of time than 1 sec should do the trick.

    go test -bench=. -benchtime=10s

    本回答被题主选为最佳回答 , 对您是否有帮助呢?
    评论

报告相同问题?

悬赏问题

  • ¥15 linux驱动,linux应用,多线程
  • ¥20 我要一个分身加定位两个功能的安卓app
  • ¥15 基于FOC驱动器,如何实现卡丁车下坡无阻力的遛坡的效果
  • ¥15 IAR程序莫名变量多重定义
  • ¥15 (标签-UDP|关键词-client)
  • ¥15 关于库卡officelite无法与虚拟机通讯的问题
  • ¥15 目标检测项目无法读取视频
  • ¥15 GEO datasets中基因芯片数据仅仅提供了normalized signal如何进行差异分析
  • ¥100 求采集电商背景音乐的方法
  • ¥15 数学建模竞赛求指导帮助