尝试并行化的速度不够快

Question

我阅读了Go的并发模型，还看到了并发性和并行性之间的区别。为了测试并行执行，我编写了以下程序。

package main

import (

    "fmt"

    "runtime"

    "time"

)

const count = 1e8

var buffer [count]int

func main() {

    fmt.Println("GOMAXPROCS: ", runtime.GOMAXPROCS(0))

    // Initialise with dummy value

    for i := 0; i < count; i++ {

        buffer[i] = 3

    }

    // Sequential operation

    now := time.Now()

    worker(0, count-1)

    fmt.Println("sequential operation: ", time.Since(now))

    // Attempt to parallelize

    ch := make(chan int, 1)

    now = time.Now()

    go func() {

        worker(0, (count/2)-1)

        ch <- 1

    }()

    worker(count/2, count-1)

    <-ch

    fmt.Println("parallel operation: ", time.Since(now))

}

func worker(start int, end int) {

    for i := start; i <= end; i++ {

        task(i)

    }

}

func task(index int) {

    buffer[index] = 2 * buffer[index]

}

但问题是：结果不是很令人愉快。

GOMAXPROCS:  8

sequential operation:  206.85ms

parallel operation:  169.028ms

使用戈鲁丁确实可以加快速度，但还不够。我预计它的速度会接近两倍。我的代码和/或理解有什么问题？我怎样才能接近两倍的速度？

Answer 1

并行化功能强大，但很难在如此小的计算负载下看到。下面是一些结果差异较大的示例代码：package mainimport (&nbsp; &nbsp; "fmt"&nbsp; &nbsp; "math"&nbsp; &nbsp; "runtime"&nbsp; &nbsp; "time")func calctest(nCPU int) {&nbsp; &nbsp; fmt.Println("Routines:", nCPU)&nbsp; &nbsp; ch := make(chan float64, nCPU)&nbsp; &nbsp; startTime := time.Now()&nbsp; &nbsp; a := 0.0&nbsp; &nbsp; b := 1.0&nbsp; &nbsp; n := 100000.0&nbsp; &nbsp; deltax := (b - a) / n&nbsp; &nbsp; stepPerCPU := n / float64(nCPU)&nbsp; &nbsp; for start := 0.0; start < n; {&nbsp; &nbsp; &nbsp; &nbsp; stop := start + stepPerCPU&nbsp; &nbsp; &nbsp; &nbsp; go f(start, stop, a, deltax, ch)&nbsp; &nbsp; &nbsp; &nbsp; start = stop&nbsp; &nbsp; }&nbsp; &nbsp; integral := 0.0&nbsp; &nbsp; for i := 0; i < nCPU; i++ {&nbsp; &nbsp; &nbsp; &nbsp; integral += <-ch&nbsp; &nbsp; }&nbsp; &nbsp; fmt.Println(time.Now().Sub(startTime))&nbsp; &nbsp; fmt.Println(deltax * integral)}func f(start, stop, a, deltax float64, ch chan float64) {&nbsp; &nbsp; result := 0.0&nbsp; &nbsp; for i := start; i < stop; i++ {&nbsp; &nbsp; &nbsp; &nbsp; result += math.Sqrt(a + deltax*(i+0.5))&nbsp; &nbsp; }&nbsp; &nbsp; ch <- result}func main() {&nbsp; &nbsp; nCPU := runtime.NumCPU()&nbsp; &nbsp; calctest(nCPU)&nbsp; &nbsp; fmt.Println("")&nbsp; &nbsp; calctest(1)}这是我得到的结果：Routines: 8853.181µsRoutines: 12.031358ms