Date: Sun, 12 May 2013 11:00:45 -0400 From: <jfoug@....net> To: john-dev@...ts.openwall.com Cc: Alexander Cherepanov <cherepan@...me.ru> Subject: Re: dynamic_1300 selftest FAILED (get_hash(1)) in bleeding It should also be much faster, since it now properly uses SSE2. The prior version, even though it appears to be SSE, was in fact not doing SSE at all. Like i said, the md5_overwrite_raw() dyna primative was written specifically for phpass, where there was a little additional code added to make things run 'properly'. As a side note, I will be posting some additional primatives to dyan shortly. These allow setting lengths of 20, and 40. I also made changes to the all of the length setters, to place a 0x80 in any sse2 buffers at the proper location. These 0x80 bytes were only being placed into the sse buffers for certain lengths. ---- Alexander Cherepanov <cherepan@...me.ru> wrote: > Thanks, it works now. > > -- > Alexander Cherepanov > > On 2013-05-12 03:04, jfoug@....net wrote: > > dyna_1300 was not a format I wrote. It was not working correctly. It used a function that was probably not a good one to 'export' for general usage. That function was written specifically for phpass. > > > > I fixed the format. It now works properly, and the speed went from 1300 to 5500 once ir properly is using SSE now.
Powered by blists - more mailing lists
Powered by Openwall GNU/*/Linux - Powered by OpenVZ