Fast vectorized conversion from RGB to BGRA

Question

In a follow-up to some previous questions on converting RGB to RGBA, and ARGB to BGR, I would like to speed up a RGB to BGRA conversion with SSE

Answer 1

I personally found that implementing the following gave me the best result for converting BGR-24 to ARGB-32.

This code runs at about 8.8ms on an image whereas the 128-bit vectorization code presented above came in at 14.5ms per image.

void PixelFix(u_int32_t *buff,unsigned char *diskmem)
{
    int i,j;
    int picptr, srcptr;
    int w = 1920;
    int h = 1080;

    for (j=0; j

Previously, I had been using this routine (about 13.2ms per image). Here, buff is an unsigned char*.

for (j=0; j

Running a 2012 MacMini 2.6ghz/i7.