Can't have function prototype with 4 __m128 parameters?

Started by
2 comments, last by Zoner 11 years, 9 months ago
Trying to copy code from Game Engine Architecture by Jason Gregory..

I have a .h file, SSEMatrixMultiplier, where I tried to put a function prototype:

__m128 mulVectorMatrixAttempt2(__m128 v, __m128 Mrow1, __m128 Mrow2, __m128 Mrow3, __m128 Mrow4);


and a .cpp file, where I try to implement the function:


__m128 mulVectorMatrixAttempt2(__m128 v, __m128 Mrow1, __m128 Mrow2, __m128 Mrow3, __m128 Mrow4)
{
__m128 xMrow1 = _mm_mul_ps(_mm_replicate_x_ps(v), Mrow1);
__m128 yMrow2 = _mm_mul_ps(_mm_replicate_y_ps(v), Mrow2);
__m128 zMrow3 = _mm_mul_ps(_mm_replicate_z_ps(v), Mrow3);
__m128 wMrow4 = _mm_mul_ps(_mm_replicate_w_ps(v), Mrow4);

__m128 result = _mm_add_ps(xMrow1, yMrow2);
result = _mm_add_ps(result, zMrow3);
result = _mm_add_ps(result, wMrow4);
return result;
}


But the compiler gives me the following errors:

Error 1 error C2719: 'Mrow3': formal parameter with __declspec(align('16')) won't be aligned c:\...simd sse math test\simd sse math test\ssematrixmultiplier.h 20
Error 3 error C2719: 'Mrow3': formal parameter with __declspec(align('16')) won't be aligned c:\...simd sse math test\simd sse math test\ssematrixmultiplier.cpp 4
Error 2 error C2719: 'Mrow4': formal parameter with __declspec(align('16')) won't be aligned c:\...simd sse math test\simd sse math test\ssematrixmultiplier.h 20
Error 4 error C2719: 'Mrow4': formal parameter with __declspec(align('16')) won't be aligned c:\...simd sse math test\simd sse math test\ssematrixmultiplier.cpp 4[/quote]

Anyone know what's up?
Advertisement
I'm surprised you only get errors for those two parameters as the MSDN page for the error C2719 (http://msdn.microsoft.com/en-us/library/373ak2y1(v=VS.71).aspx) states that __align is not allowed on function parameters so the first 3 should have also tripped the message.

An MSDN Social answer (http://social.msdn.microsoft.com/Forums/en-US/vclanguage/thread/85e53fd8-9345-42d2-bfa1-1ce4ee4e4c68) recommends passing them as pointers instead; you might be able to get away with const ref as well.
Thanks phantom!
When compiling for 32 bits you need to pass SIMD data by reference or pointer. 64 bit ABI allows them to be passed by value at a language level, but if you compile on both targets you need to do it the 32 bit way.

The simple reason is 32 bit ABI does not correctly align the stack to 16 bytes. You may ask, what about local variables then?. Functions that have __m128 variables as local variables cause the compiler to generate additional code to align the stack so they can be stored there.

Note that even in x64, __m128 variables are not passed via xmm registers. They will be written to the stack and passed by reference behind the scenes. However your code will compile when you write it to pass by value. scalar floats and doubles (i.e. stuff not using __m128 as their data type) DO get passed in xmm registers, but the ABI does not handle SIMD data types. Weird I know but thats the way it is speced at the moment. The way of dealing with this problem is to forceinline all the code passing by value, but that has some rather practical limits.
http://www.gearboxsoftware.com/

This topic is closed to new replies.

Advertisement