Date: Tue, 5 Jun 2012 11:37:49 +0530 From: SAYANTAN DATTA <std2048@...il.com> To: john-dev@...ts.openwall.com Subject: Re: GPU formats vs Test Suite On Tue, Jun 5, 2012 at 1:30 AM, magnum <john.magnum@...hmail.com> wrote: > Lukas, Sayantan, > > I just ran the Test Suite again, against all GPU formats (that it > supports) on bull. Here are the ones that failed: > > form=phpass-opencl guesses: 225 time: 0:00:02:23 : > Expected count(s) (1470) [!!!FAILED!!!] > > Note that phpass-cuda passes exact same test, max length is 15. > > form=mscash-cuda guesses: 0 unk unk : Expected > count(s) (1430) [!!!FAILED!!!] > > form=mscash2-opencl guesses: 0 unk unk : Expected > count(s) (1500) [!!!FAILED!!!] > > form=mscash2-cuda guesses: 0 unk unk : Expected > count(s) (1479) [!!!FAILED!!!] > > These three do no seem to crack a single password IRL. The actual fault > *may* be in the TS but most of these passed the tests in February so I > think you should have a look at this. The easiest way to grab TS on bull is > from ~magnum/TestSuite_1.12-2.tar.**bz2 - unpack it in your john > directory. > > To avoid a -test=0 run (that won't work well with GPU formats) within TS, > run it with -noprelim option. Eg: > > $ ./jtrts.pl -noprelim ssha-opencl > > The -passthru option can be used to select platform: > > $ ./jtrts.pl -noprelim -passthru "-plat=1" opencl > > magnum > Hi magnum, I found that the output of openssl version of pbkdf2 in mscash2(cpu version) do not matches with the SSE version of pbkdf2 for salts greater than 19 charachters. Since mscash2-opencl is based on openssl version of pbkdf2, so I have limited the max salt length to 19 charcters only. But the file /test/MSCash2_tst.in contains salts upto 22 character. I think this is the main source of problem. Regards, Sayantan Content of type "text/html" skipped
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.