topic SSE in Intel® Integrated Performance Primitives

SSE

send4chandu — Sat, 07 Aug 2004 04:01:55 GMT

Hi ,

I am trying to run a c++ code in VC++ 6.0 environment (with processor pack installed). My code contains few assembly language instrucions to execute SSE. My code is as follows.

#include

#define ARRSIZE 400

void main()
{

__declspec(align(16)) float a[ARRSIZE], b[ARRSIZE], c[ARRSIZE];

_asm
{
push esi;
push edi;
mov edi, a;
mov esi, b;
mov edx, c;
mov ecx, 100;
loop:
movaps xmm0, [edi];
movups xmm1, [esi];
mulps xmm0, xmm1;
movups [edx], xmm0;
add edi, 16;
add esi, 16;
add edx, 16;
dec ecx;
jnz loop;
pop edi;
pop esi;
}

}

When I run the above code, I am getting following 3 errors.

--------------------Configuration: mul - Win32 Debug--------------------
Compiling...
mul.cpp
c:program filesmicrosoft visual studiovc98includemmintrin.h(29) : error C2146: syntax error : missing ';' before identifier '_m_from_int'

c:program filesmicrosoft visual studiovc98includemmintrin.h(29) : error C2501: '__m64' : missing storage-class or type specifiers

c:program filesmicrosoft visual studiovc98includemmintrin.h(29) : fatal error C1004: unexpected end of file found

Error executing cl.exe.

mul.obj - 3 error(s), 0 warning(s)

-------------------------------------------------------------------------------------------

Can any one can solve my problem.

-Thanks in advance.

-Ravi Akkenapally

Re: SSE

Vladimir_Dudnik — Mon, 09 Aug 2004 16:53:20 GMT

Ravi,

could you try this with Intel compiler?

Regards,

Vladimir

Re: SSE

ids-removed222 — Thu, 19 Aug 2004 09:15:22 GMT

I changed movaps to movups , it pass

but new problem emerge at end

#define ARRSIZE 4
void main1()
{
//__declspec(align(16)) float a[ARRSIZE], b[ARRSIZE], c[ARRSIZE];
float * a = new float[ARRSIZE];
float * b = new float[ARRSIZE];
float * c = new float[ARRSIZE];

_asm
{
push esi;
push edi;
push edx
mov edi, a;
mov esi, b;
mov edx, c;
mov ecx, ARRSIZE;
L1:
movups xmm0, [edi];
movups xmm1, [esi];
mulps xmm0, xmm1;
movups [edx], xmm0;
add edi, 16;
add esi, 16;
add edx, 16;
;dec ecx;
;jnz loop1;
loop L1
pop edx;
pop edi;
pop esi;

}
delete [] a;
delete [] b;

delete [] c; // only wrong here

}