155714SkrisRIPEMD-160 255714Skrishttp://www.esat.kuleuven.ac.be/~bosselae/ripemd160.html 355714Skris 455714SkrisThis is my implementation of RIPEMD-160. The pentium assember is a little 555714Skrisoff the pace since I only get 1050 cycles, while the best is 1013. 655714SkrisI have a few ideas for how to get another 20 or so cycles, but at 7194206Ssimonthis point I will not bother right now. I believe the trick will be 855714Skristo remove my 'copy X array onto stack' until inside the RIP1() finctions the 955714Skrisfirst time round. To do this I need another register and will only have one 1055714Skristemporary one. A bit tricky.... I can also cleanup the saving of the 5 words 1155714Skrisafter the first half of the calculation. I should read the origional 1255714Skrisvalue, add then write. Currently I just save the new and read the origioal. 1355714SkrisI then read both at the end. Bad. 1455714Skris 1555714Skriseric (20-Jan-1998) 16