Die*_*91b 5 c++ gcc vectorization sse2
下一个代码(使用 -O3 -ffast-math):
#include <cmath>
float a[4];
void sin1() {
for(unsigned i = 0; i < 4; i++) a[i] = sinf(a[i]);
}
Run Code Online (Sandbox Code Playgroud)
编译 sinf 的矢量化版本 (_ZGVbN4v_sinf):
sin1():
sub rsp, 8
movaps xmm0, XMMWORD PTR a[rip]
call _ZGVbN4v_sinf
movaps XMMWORD PTR a[rip], xmm0
add rsp, 8
ret
Run Code Online (Sandbox Code Playgroud)
但是当我使用 c++ 版本的 sinf (std::sin) 时,不会发生矢量化:
void sin2() {
for(unsigned i = 0; i < 4; i++) a[i] = std::sin(a[i]);
}
sin2():
sub rsp, 8
movss xmm0, DWORD PTR a[rip]
call sinf
movss DWORD PTR a[rip], xmm0
movss xmm0, DWORD PTR a[rip+4]
call sinf
movss DWORD PTR a[rip+4], xmm0
movss xmm0, DWORD PTR a[rip+8]
call sinf
movss DWORD PTR a[rip+8], xmm0
movss xmm0, DWORD PTR a[rip+12]
call sinf
movss DWORD PTR a[rip+12], xmm0
add rsp, 8
ret
Run Code Online (Sandbox Code Playgroud)