📄 memcpy.cpp
字号:
// memcpy.cpp : Defines the entry point for the console application.
//
#include <windows.h>
#include <stdio.h>
#include <conio.h> // getch
char *tbuf = NULL;
void memfill(void *dst, int n32, unsigned long i)
{
__asm {
movq mm0, n32
punpckldq mm0, mm0
mov edi, dst
loopwrite:
movntq 0[edi], mm0
movntq 8[edi], mm0
//movntq 16[edi], mm0
//movntq 24[edi], mm0
//movntq 32[edi], mm0
//movntq 40[edi], mm0
//movntq 48[edi], mm0
//movntq 56[edi], mm0
add edi, 16
sub i, 2
jg loopwrite
emms
}
}
void mem4(void *dst, void *src, int nbytes)
{
__asm {
mov esi, src
mov ecx, nbytes
mov ebx, ecx
shr ebx, 11 // 2048 bytes at a time
mov edi, dst
loop2k: // Copy 2k into temporary buffer
push edi
mov edi, tbuf
mov ecx, 2048
shr ecx, 6
loopMemToL1:
prefetchnta 64[ESI] // Prefetch next loop, non-temporal
prefetchnta 96[ESI]
movq mm1, 0[ESI] // Read in source data
movq mm2, 8[ESI]
movq mm3, 16[ESI]
movq mm4, 24[ESI]
movq mm5, 32[ESI]
movq mm6, 40[ESI]
movq mm7, 48[ESI]
movq mm0, 56[ESI]
movq 0[EDI], mm1 // Store into L1
movq 8[EDI], mm2
movq 16[EDI], mm3
movq 24[EDI], mm4
movq 32[EDI], mm5
movq 40[EDI], mm6
movq 48[EDI], mm7
movq 56[EDI], mm0
add esi, 64
add edi, 64
dec ecx
jnz loopMemToL1
pop edi // Now copy from L1 to system memory
push esi
mov esi, tbuf
mov ecx, 2048
shr ecx, 6
loopL1ToMem:
movq mm1, 0[ESI] // Read in source data from L1
movq mm2, 8[ESI]
movq mm3, 16[ESI]
movq mm4, 24[ESI]
movq mm5, 32[ESI]
movq mm6, 40[ESI]
movq mm7, 48[ESI]
movq mm0, 56[ESI]
movntq 0[EDI], mm1 // Non-temporal stores
movntq 8[EDI], mm2
movntq 16[EDI], mm3
movntq 24[EDI], mm4
movntq 32[EDI], mm5
movntq 40[EDI], mm6
movntq 48[EDI], mm7
movntq 56[EDI], mm0
add esi, 64
add edi, 64
dec ecx
jnz loopL1ToMem
pop esi // Do next 2k block
dec ebx
jnz loop2k
emms
}
}
void mem3(void *dst, void *src, int nbytes)
{
_asm {
mov esi, src
mov edi, dst
mov ecx, nbytes
shr ecx, 6 // 64 bytes per iteration
loop1:
prefetchnta 64[ESI] // Prefetch next loop, non-temporal
prefetchnta 96[ESI]
movq mm1, 0[ESI] // Read in source data
movq mm2, 8[ESI]
movq mm3, 16[ESI]
movq mm4, 24[ESI]
movq mm5, 32[ESI]
movq mm6, 40[ESI]
movq mm7, 48[ESI]
movq mm0, 56[ESI]
movntq 0[EDI], mm1 // Non-temporal stores
movntq 8[EDI], mm2
movntq 16[EDI], mm3
movntq 24[EDI], mm4
movntq 32[EDI], mm5
movntq 40[EDI], mm6
movntq 48[EDI], mm7
movntq 56[EDI], mm0
add esi, 64
add edi, 64
dec ecx
jnz loop1
emms
}
}
void mem2(void *dst, void *src, int nbytes)
{
_asm {
mov esi, src
mov edi, dst
mov ecx, nbytes
shr ecx, 6 // 64 bytes per iteration
loop1:
movq mm1, 0[ESI] // Read in source data
movq mm2, 8[ESI]
movq mm3, 16[ESI]
movq mm4, 24[ESI]
movq mm5, 32[ESI]
movq mm6, 40[ESI]
movq mm7, 48[ESI]
movq mm0, 56[ESI]
movntq 0[EDI], mm1 // Non-temporal stores
movntq 8[EDI], mm2
movntq 16[EDI], mm3
movntq 24[EDI], mm4
movntq 32[EDI], mm5
movntq 40[EDI], mm6
movntq 48[EDI], mm7
movntq 56[EDI], mm0
add esi, 64
add edi, 64
dec ecx
jnz loop1
emms
}
}
void mem1(void *dst, void *src, int nbytes)
{
_asm {
mov esi, src
mov edi, dst
mov ecx, nbytes
shr ecx, 6 // 64 bytes per iteration
loop1:
movq mm1, 0[ESI] // Read in source data
movq mm2, 8[ESI]
movq mm3, 16[ESI]
movq mm4, 24[ESI]
movq mm5, 32[ESI]
movq mm6, 40[ESI]
movq mm7, 48[ESI]
movq mm0, 56[ESI]
movq 0[EDI], mm1 // Write to dstination
movq 8[EDI], mm2
movq 16[EDI], mm3
movq 24[EDI], mm4
movq 32[EDI], mm5
movq 40[EDI], mm6
movq 48[EDI], mm7
movq 56[EDI], mm0
add esi, 64
add edi, 64
dec ecx
jnz loop1
emms
}
}
#define size 32768 * 1024
int main(int argc, char* argv[])
{
char *foo = new char[size];
char *foo2 = new char[size];
// warm me up
mem1(foo2, foo, size);
mem1(foo2, foo, size);
tbuf = new char[2048];
LARGE_INTEGER s1, s2, f;
::QueryPerformanceFrequency(&f);
double el;
::QueryPerformanceCounter(&s1);
mem1(foo2, foo, size);
::QueryPerformanceCounter(&s2);
el = s2.QuadPart - s1.QuadPart;
el /= double(f.QuadPart);
printf("SGI ex1: %fms = %fmb/sec\n", el*1000, float(size) / 1024 / el / 1024);
::QueryPerformanceCounter(&s1);
mem2(foo2, foo, size);
::QueryPerformanceCounter(&s2);
el = s2.QuadPart - s1.QuadPart;
el /= double(f.QuadPart);
printf("SGI ex2: %fms = %fmb/sec\n", el*1000, float(size) / 1024 / el / 1024);
::QueryPerformanceCounter(&s1);
mem3(foo2, foo, size);
::QueryPerformanceCounter(&s2);
el = s2.QuadPart - s1.QuadPart;
el /= double(f.QuadPart);
printf("SGI ex3: %fms = %fmb/sec\n", el*1000, float(size) / 1024 / el / 1024);
::QueryPerformanceCounter(&s1);
mem4(foo2, foo, size);
::QueryPerformanceCounter(&s2);
el = s2.QuadPart - s1.QuadPart;
el /= double(f.QuadPart);
printf("SGI ex4: %fms = %fmb/sec\n", el*1000, float(size) / 1024 / el / 1024);
::QueryPerformanceCounter(&s1);
memcpy(foo2, foo, size);
::QueryPerformanceCounter(&s2);
el = s2.QuadPart - s1.QuadPart;
el /= double(f.QuadPart);
printf("\nmemcpy %fms = %fmb/sec\n\n", el*1000, float(size) / 1024 / el / 1024);
for (int i = 0; i < 4; i++) {
::QueryPerformanceCounter(&s1);
memfill(foo2, 0, size/8);
::QueryPerformanceCounter(&s2);
el = s2.QuadPart - s1.QuadPart;
el /= double(f.QuadPart);
printf("memfill %fms = %fmb/sec\n", el*1000, float(size) / 1024 / el / 1024);
}
::QueryPerformanceCounter(&s1);
memset(foo2, 0, size);
::QueryPerformanceCounter(&s2);
el = s2.QuadPart - s1.QuadPart;
el /= double(f.QuadPart);
printf("\nmemset %fms = %fmb/sec\n\n", el*1000, float(size) / 1024 / el / 1024);
printf("Press a key to exit...");
_getch();
return 0;
}
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -