Date: Tue, 15 Mar 2011 17:09:18 -0500 From: "jfoug" <jfoug@....net> To: <john-users@...ts.openwall.com> Subject: RE: starting and 'running' the md5-gen format john -test=5 -for=md5_gen(1064) Benchmarking: md5_gen(1064) md5_gen(1064): sha1($p) SSE2 [128x1 (MD5_Go)]... DONE Raw: 6753K c/s john -test=5 -for=raw-sha1 Benchmarking: Raw SHA-1 SSE2 [raw-sha1 SSE2]... DONE Raw: 7954K c/s john_oldersse2 -test=5 -for=raw-sha1 Benchmarking: Raw SHA-1 SSE2 [raw-sha1 SSE2]... DONE Raw: 7456K c/s I have 'some' SSE startup into the 'generic' (md5_gen) code. It is not perfect, and will likely need to be optimized more. However, it now will use SSE, if the build is SSE. I also have sped up the 'raw-sha1' in SSE/MMX if using the .S code (x32), as can be seen above. The test with john_oldersse2 is timing of 'current' released john. The SHA1 - SSE code in the generic type does not yet do intrinsic. I will have to do some additional code to get that working. Jim.
Powered by blists - more mailing lists
Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.