Question

From Intel’s article:

The integer format conversions are commonly used in imaging and video applications. For example, they can be used when converting RGBA from four bytes to four floats prior to computation on a pixel. One SSE4 convert instruction can do the same thing as four SIMD instructions did previously, as shown.

SSE2: 
pmovd xmm0, m32 
pxor xmm7, xmm7 
punpcklbw xmm0, xmm7 
punpcklwd xmm0, xmm7 
cvtdq2ps xmm0, xmm0

SSE4: 
pmovzxbd xmm0, m32 
cvtdq2ps xmm0, xmm0

Can someone show me a complete example please? Thanks a lot!

Answer 1

Because pmovzxbd does only work on 32 bit memory operands or 128 bit sse registers, but not on general purpose registers, you need to insert some type conversions or an explicit load from GPR to SSE.

__m128 convert_RGBA_to_float(float* rgba) {
    return  _mm_cvtepi32_ps( _mm_cvtepu8_epi32 ( *(__m128i*)rgba) );
}