Printable View
As you see the compiler has generated code which is using SSE instead of FPU while the /arch:IA32 is set. Any idea why?Code:int c =0;
000000013FD223B5 mov dword ptr [rsp+20h],0
float f=0;
000000013FD223BD xorps xmm0,xmm0
000000013FD223C0 movss dword ptr [rsp+24h],xmm0
int a = c + f;
000000013FD223C6 cvtsi2ss xmm0,dword ptr [rsp+20h]
000000013FD223CC addss xmm0,dword ptr [rsp+24h]
000000013FD223D2 cvttss2si eax,xmm0
000000013FD223D6 mov dword ptr [rsp+28h],eax
Random guess from /arch (x86)
Tim S.Quote:
/arch only affects code generation for native functions. When you use /clr to compile, /arch has no effect on code generation for managed functions.
Thanks for sharing your opinion but when you see the assembly it is not CL. The answer is here.Quote:
Originally Posted by stahta01
For the people NOT to waste their time when using the 64 bit compiler the option to generate 32 bit without SSE does NOT work is the only real likely answer I saw.Quote:
Originally Posted by siavoshkc
Tim S.
Incorrect. Compiler ignores /arch:ia32 when is creating x64 executive. That is because there is no x64 CPU with no SSE in real world.Quote:
Originally Posted by stahta01
This is 64-bit code. SSE is part of x64 architecture. There is nothing unusual here.