I'm new to mmx technology and
would like to speed up my application
by simply replacing the C++ memcpy
function with a faster mmx one.
AMD provides a copy/paste c++
and assembler version of the memcpy
function optimized for AMD.
I couldn't find one at the intel site
nor in the web. I found many documents
about mmx, I downloaded the IPP and MKL
but none of them offers a copy/paste