found the answer with the help of author os repo, disassemble the code, fast use XORQ, the safe one use XORL but actually xor byte by byte. the fast convert the oprand to uintptr. so it can use XOR 4/8 byte one time
在 2019年3月19日星期二 UTC+8上午10:36:21，zhou yu写道： > > *Why the XOR between*(*[]uintptr)(unsafe.Pointer(&dst)) is fast then > []byte?* > > I find a fast xor lib https://github.com/templexxx/xor, and write the > benchmark below, > > the result show that BenchmarkFastXORWords is 5x faster than BenchmarkS > afeXORBytes on my computer > > I can't figure out why? Anyone can help me. THX > > package main > > import ( > "math/rand" > "testing" > "unsafe" > ) > > const wordSize = int(unsafe.Sizeof(uintptr(0))) > > var e = make([]byte, 160) > var f = make([]byte, 160) > var g = make([]byte, 160) > > func init() { > for i := 0; i < 160; i++ { > rand.Seed(22) > e[i] = byte(rand.Intn(256)) > f[i] = byte(rand.Intn(256)) > } > } > > func BenchmarkSafeXORBytes(b *testing.B) { > for i := 0; i < b.N; i++ { > safeXORBytes(g, e, f) > } > } > > func BenchmarkFastXORWords(b *testing.B) { > for i := 0; i < b.N; i++ { > fastXORWords(g, e, f) > } > } > > func safeXORBytes(dst, a, b []byte) { > n := len(a) > ex := n % 8 > for i := 0; i < ex; i++ { > dst[i] = a[i] ^ b[i] > } > > for i := ex; i < n; i += 8 { > _dst := dst[i : i+8] > _a := a[i : i+8] > _b := b[i : i+8] > _dst[0] = _a[0] ^ _b[0] > _dst[1] = _a[1] ^ _b[1] > _dst[2] = _a[2] ^ _b[2] > _dst[3] = _a[3] ^ _b[3] > > _dst[4] = _a[4] ^ _b[4] > _dst[5] = _a[5] ^ _b[5] > _dst[6] = _a[6] ^ _b[6] > _dst[7] = _a[7] ^ _b[7] > } > } > > func fastXORWords(dst, a, b []byte) { > dw := *(*[]uintptr)(unsafe.Pointer(&dst)) > aw := *(*[]uintptr)(unsafe.Pointer(&a)) > bw := *(*[]uintptr)(unsafe.Pointer(&b)) > n := len(b) / wordSize > ex := n % 8 > for i := 0; i < ex; i++ { > dw[i] = aw[i] ^ bw[i] > } > > for i := ex; i < n; i += 8 { > _dw := dw[i : i+8] > _aw := aw[i : i+8] > _bw := bw[i : i+8] > _dw[0] = _aw[0] ^ _bw[0] > _dw[1] = _aw[1] ^ _bw[1] > _dw[2] = _aw[2] ^ _bw[2] > _dw[3] = _aw[3] ^ _bw[3] > _dw[4] = _aw[4] ^ _bw[4] > _dw[5] = _aw[5] ^ _bw[5] > _dw[6] = _aw[6] ^ _bw[6] > _dw[7] = _aw[7] ^ _bw[7] > } > } > > > -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
