Newer Intel and AMD processors have hardware support for implementing AES using the instruction set AES-NI (instructions AESENC, AESKEYGENASSIST etc). Do you know of any clean example implementation using these instructions? In assembly (x86-64 or 32) or using wmmintrin.h (I use gcc and gas, but nasm is ok too).
I found a couple of implementations but they were cluttered with ifdefs and hard to follow macros. It would be nice to begin with a clean and easy to read implementation of, say, encrypting just one block using AES 128 and then build on from there.