Sleep(10);
SetPriorityClass(hprocess, REALTIME_PRIORITY_CLASS);
QueryPerformanceFrequency((LARGE_INTEGER*)&freq);
QueryPerformanceCounter((LARGE_INTEGER*)&start);

func();

QueryPerformanceCounter((LARGE_INTEGER*)&end);
SetPriorityClass(hprocess, oldclass);

return (int)(end-start);
}

main()
{
printf("func1 = %d\n", prof(func1));
printf("func2 = %d\n", prof(func2));
printf("func3 = %d\n", prof(func3));
printf("func4 = %d\n", prof(func4));
}

結果
func1 = 6685
func2 = 9706
func3 = 6426
func4 = 10277

xchg が一番早い。xorが一番遅い。