I've also added SSE4.1 instructions to my Instruction Reference. SSE4.2 instructions will be added very shortly.

http://mattst88.no-ip.com:8181/programming/asmref/

The latest YASM, 0.6.1, has support for SSE 4.1 and 4.2, if that helps any (for comparisons, etc).

Yes, a good set of commands. On idea, they should be useful. There is DivX codecs (version 6.6.1), wrotten by means of SSE4, under tests a gain of speed in comparison with the same realization, but by means of SSE2, about 70 %. It would be necessary to try test most the code, rewrite one on SSE4. I hope, that shortly there will be version FASM with their support. Though it's need to think still, whether have the given commands a urgency because the processors supporting them, it is NOT ENOUGH.

If CRC32 can (and it hopefully will) outperform an optimized crc32 procedure then it'll be a great boost to security and integrity checks as a whole. -- In the next 3-5 years when SSE4 becomes as common as SSE/2 is now -- but I'd hope we'd have a better architecture all together by then --

DotProduct is a step closer to giving CPU's the same capabilities as GPU's. Just need more SIMD ALU's.

Finally being able to multiply packed Dwords and Qwords is great, although it's kind of late by a decade.