SSE code, wrong result

[VBNick]

hi, I put together a small app to try using SEE in assembly, but I am getting the wrong output.
these are the .cpp file and the .asm file

#define WIN32_LEAN_AND_MEAN
#include <iostream>
using namespace std;

extern "C" int *myFunc(float *a, float *b, float *result);

int main(int argc, char *argv[])
{
	float inA[] = {1.0f, 2.0f, 3.0f, 4.0f};
	float inB[] = {1.0f, 2.0f, 3.0f, 4.0f};
	float ret[4];

	myFunc(inA, inB, ret);

	cout << ret[0] << endl;
	cout << ret[1] << endl;
	cout << ret[2] << endl;
	cout << ret[3] << endl;

	system("pause");
	
	return 0;
}

.586P
.XMM
.MODEL FLAT, C
.STACK
.DATA
.CODE

myFunc PROC a:DWORD, b:DWORD, result:DWORD

movups xmm0, [a]
movups xmm1, [b]
addps xmm0, xmm1
movups [result], xmm0
ret

myFunc ENDP

END

I was expecting to get the result

2
4
6
8

but instead got

-1.07374e+008
-1.07374e+008
-1.07374e+008
-1.07374e+008

can anyone help me out?

[tesuji]

Hi

your indirect address modes are wrong. Unfortunately, you won't get any error message. That's life in assembly programming.

To fast show you the problem in principle I copied your assembly code inline a C program:

void myFunc(float *a, float *b, float *result)   ; *myFunc not necessary but ok
{
__asm {
        mov edi, a;          ; <--- mov via 32bit reg only!
	movups xmm0, [edi]   ; <--- indirect adress mode [a] not allowed!

        mov edi, b;
        movups xmm1, [edi]   ; <--- dito, indirect adress mode [b] not allowed!

        addps xmm0, xmm1

        mov edi, result
        movups [edi], xmm0   ; <--- dito, indirect adress mode [result] not allowed! 
      }
}
int main(int argc, char *argv[])
{
	float inA[] = {1.0f, 2.0f, 3.0f, 4.0f};
	float inB[] = {1.0f, 2.0f, 3.0f, 4.0f};
	float ret[4];

	myFunc(inA, inB, ret);


	cout << ret[0] << endl; // only first element of vector is calculated
        // result is
        // 2

	//cout << ret[1] << endl;  
	//cout << ret[2] << endl;
	//cout << ret[3] << endl;

	system("pause");
	
	return 0;
}

Now it's your turn to program the loop for adding all four elements of the vectors.

-- tesu

[VBNick]

Thanks! works now. But why can't I dereference the argument? is there another way to pass the arguments in so that I don't need the extra copy?