Question

I’ve tried using the following command:

 __m128i b = _mm_set_epi32 (y, y, x, x);

Where y and x are ints.

Where I run the debugger I see that b is of type: unsigned __int64[2]

I intended b to be 4 integers of 32 bits each (I think that’s what they say here:
http://msdn.microsoft.com/en-us/library/019beekt.aspx )

Do you know what’s wring with my code?

Thanks

Answer 1

I don’t see a problem here – a 128-bit integer is represented by two 64-bit integers, which is the largest native type outside of SSE registers a 64-bit Intel CPU can handle.

Your debugger is likely showing the “real” type, not the typedefed or macro’d type.