Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sat, 15 Dec 2012 00:17:14 +0100
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: Benchmark versus incremental mode (speed)

On 14 Dec, 2012, at 23:19 , Claudio André <claudioandre.br@...il.com> wrote:
> Magnum, all
> 
> Why benchmark shows a speed so different from real cracking (even now)? Is there anything i missed? Should i do something in a different way?
> If possible, could you please give me any hint or advice?

It should be a lot better since I merged that self-test timing patch a week ago or so. Maybe we see problems now that were hidden before.

> First point, which one is the real number? Please take a look in [1] and [2].
> My test vector is not that bad comparing to --incremental mode. Or is it?

Per definition, the real crack gives the real figure. If they differ it may be due to things like test vector length not matching the IRL run (in case the format suffers significantly from length) or if the IRL hash does not use 5000 iterations. If that's not it, I'm not sure.

> I need some advice since auto-tunning code have to show "the" speed.
> 
> [1]:
> claudio@...l:~/unstable/src$ ../run/john -fo:sha256crypt-opencl ~/sha256 -inc --max-run-time=600
> OpenCL platform 0: NVIDIA CUDA, 1 device(s).
> Using device 0: GeForce GTX 570
> Building the kernel, this could take a while
> Local work size (LWS) 128, global work size (GWS) 23040
> guesses: 0  time: 0:00:11:40 0.00%  c/s: 19753  trying: tui1194 - sidgsb
> Session stopped (max run-time reached)

Why do I not see any "Loaded x hashes with y different salts" in the above?

> [2]:
> claudio@...l:~/claudio/src$ ../run/john -fo:sha256crypt-opencl -t
> OpenCL platform 0: NVIDIA CUDA, 1 device(s).
> Using device 0: GeForce GTX 570
> Building the kernel, this could take a while
> Local work size (LWS) 128, global work size (GWS) 23040
> Benchmarking: sha256crypt (rounds=5000) [OpenCL]... DONE
> Raw:    13552 c/s real, 13592 c/s virtual

What happens if you change BENCHMARK_LENGH to 0 and re-run this benchmark?

magnum

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ