[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CA+TsHUBsK6CObweE+=LSAjAjfeRWr5Pa=AVBg-iZvmc1DCk2Yw@mail.gmail.com>
Date: Sun, 27 Jan 2013 09:56:20 +0530
From: Sayantan Datta <std2048@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: mscash2 speed (was: Proposed optimizations to pwsafe)
On Sun, Jan 27, 2013 at 6:16 AM, magnum <john.magnum@...hmail.com> wrote:
> Still, I get <75K real-life figures from latest oclHashCat and >100K from
> latest JtR. Not sure why. That is on Cat 12.8.
>
> magnum
>
>
> On 27 Jan, 2013, at 1:33 , Solar Designer <solar@...nwall.com> wrote:
>
> > On Sun, Jan 27, 2013 at 01:08:46AM +0100, Lukas Odzioba wrote:
> >> 2013/1/27 magnum <john.magnum@...hmail.com>:
> >>> Maybe that is why Atom does not list DCC2 benchmarks...
> >>
> >> Earlier he listed them:
> >>
> http://web.archive.org/web/20120103091803/http://hashcat.net/oclhashcat-plus/
> >
> > This has speed for DCC2 on 7970 with Catalyst 12.4, and for GTX 570 o/c
> > just like ours:
> >
> >
> http://web.archive.org/web/20120606061027/http://hashcat.net/oclhashcat-plus/
> >
> > 107.6k c/s and 31604 c/s, respectively.
> >
> > JtR's current speeds are:
> >
> > $ ./john -te -fo=mscash2-opencl -pla=1
> > OpenCL platform 1: AMD Accelerated Parallel Processing, 2 device(s).
> > Device 0: Tahiti (AMD Radeon HD 7900 Series)
> > Optimal Work Group Size:128
> > Kernel Execution Speed (Higher is better):1.556560
> > Benchmarking: M$ Cache Hash 2 (DCC2) PBKDF2-HMAC-SHA-1 [OpenCL]... DONE
> > Raw: 103232 c/s real, 5079K c/s virtual
> >
> > $ ./john -te -fo=mscash2-opencl
> > OpenCL platform 0: NVIDIA CUDA, 1 device(s).
> > Device 0: GeForce GTX 570
> > Optimal Work Group Size:128
> > Kernel Execution Speed (Higher is better):0.475342
> > Benchmarking: M$ Cache Hash 2 (DCC2) PBKDF2-HMAC-SHA-1 [OpenCL]... DONE
> > Raw: 30340 c/s real, 30567 c/s virtual
> >
> > $ ./john -te -fo=mscash2-cuda
> > Benchmarking: M$ Cache Hash 2 (DCC2) PBKDF2-HMAC-SHA-1 [CUDA]... DONE
> > Raw: 13714 c/s real, 13644 c/s virtual
> >
> > The CUDA build needs compile-time tuning of THREADS and BLOCKS, which I
> > did not do here...
> >
> > Alexander
> >
>
> Yes, the last time I saw it was 107.6k c/s for 7970 with cat 12.6.
Maybe hashcat is artificially limiting their speed to reduce temps and
improve responsiveness.
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.
