Go 高性能 - 打印调用堆栈

package main

import (
    "fmt"
    "runtime"
)

func main() {
    buf := make([]byte, 1024)
    n := runtime.Stack(buf, true)

    fmt.Printf("%s\n", buf[:n])
}
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.

$ go run main.go

# 输出如下 (你的输出代码路径应该和这里的不一样)
goroutine 1 [running]:
main.main()
    /home/codes/go-high-performance/main.go:10 +0x45
...
1.
2.
3.
4.
5.
6.
7.

package performance

import (
    "runtime"
    "testing"
)

func Benchmark_StackDump(b *testing.B) {
    for i := 0; i < b.N; i++ {
        buf := make([]byte, 1024)
        n := runtime.Stack(buf, true)

        _ = buf[:n]
    }
}
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.

# 运行 1000 次，统计内存分配
$ go test -run='^$' -bench=. -count=1 -benchtime=1000x -benchmem > slow.txt
1.
2.

package main

import (
    "fmt"
    "runtime"
)

func main() {
    for i := 0; ; i++ {
        if _, file, line, ok := runtime.Caller(i); ok {
            fmt.Printf("file: %s, line: %d\n", file, line)
        } else {
            break
        }
    }
}
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.

$ go run main.go

# 输出如下 (你的输出代码路径应该和这里的不一样)
file: /home/codes/go-high-performance/main.go, line: 10
file: /usr/local/go/src/runtime/proc.go, line: 250
file: /usr/local/go/src/runtime/asm_amd64.s, line: 1594
...
1.
2.
3.
4.
5.
6.
7.

package main

import (
    "fmt"
    "runtime"
    "strconv"
    "strings"
)

func main() {
    pcs := make([]uintptr, 16)
    n := runtime.Callers(0, pcs)

    frames := runtime.CallersFrames(pcs[:n])

    var sb strings.Builder
    for {
        frame, more := frames.Next()

        sb.WriteString(frame.Function)
        sb.WriteByte('\n')
        sb.WriteByte('\t')
        sb.WriteString(frame.File)
        sb.WriteByte(':')
        sb.WriteString(strconv.Itoa(frame.Line))
        sb.WriteByte('\n')

        if !more {
            break
        }
    }

    fmt.Println(sb.String())
}
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.

$ go run main.go

# 输出如下 (你的输出代码路径应该和这里的不一样)
runtime.Callers
        /usr/local/go/src/runtime/extern.go:247
main.main
        /home/codes/go-high-performance/main.go:12
runtime.main
        /usr/local/go/src/runtime/proc.go:250
runtime.goexit
        /usr/local/go/src/runtime/asm_amd64.s:1594
...
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.

package performance

import (
    "runtime"
    "strconv"
    "strings"
    "testing"
)

func stackDump() string {
    pcs := make([]uintptr, 16)
    n := runtime.Callers(0, pcs)

    frames := runtime.CallersFrames(pcs[:n])

    var buffer strings.Builder
    for {
        frame, more := frames.Next()

        buffer.WriteString(frame.Function)
        buffer.WriteByte('\n')
        buffer.WriteByte('\t')
        buffer.WriteString(frame.File)
        buffer.WriteByte(':')
        buffer.WriteString(strconv.Itoa(frame.Line))
        buffer.WriteByte('\n')

        if !more {
            break
        }
    }

    return buffer.String()
}

func Benchmark_StackDump(b *testing.B) {
    for i := 0; i < b.N; i++ {
        _ = stackDump()
    }
}
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.

# 运行 1000 次，统计内存分配
$ go test -run='^$' -bench=. -count=1 -benchtime=1000x -benchmem > fast.txt
1.
2.

$ benchstat -alpha=100 fast.txt slow.txt 

# 输出如下
name          old time/op    new time/op    delta
_StackDump-8    2.28µs ± 0%   68.89µs ± 0%  +2926.85%  (p=1.000 n=1+1)

name          old alloc/op   new alloc/op   delta
_StackDump-8    1.36kB ± 0%    1.02kB ± 0%    -24.71%  (p=1.000 n=1+1)

name          old allocs/op  new allocs/op  delta
_StackDump-8      12.0 ± 0%       1.0 ± 0%    -91.67%  (p=1.000 n=1+1)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.