Still 105 MB/sec not bad :)

Hello,

With my fix included it still performs quite well.

105 MB/sec compared to the old performance of 45 MB/sec.

Still double as fast easily !

Even more like 2.x times faster WIEEEEEEEEEEEEE.

Now me can go convert other stuff like AES and maybe even crc32 as well.

And then later implement some HasSSE2 and then me be done with it
hehehehe.

Bye,
  Skybuck.