Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 19 Mar 2012 01:25:10 +0400
From: Solar Designer <>
Subject: Re: Initial Observation of OpenCL and CUDA

myrice, magnum -

On Mon, Mar 19, 2012 at 01:47:13AM +0800, myrice wrote:
> I tested it built with linux-x86-cuda and linux-x86-opencl. My environment
> is:
> CPU: Intel(R) Core(TM) i7-2600 CPU @ 3.40GHz
> GPU: GeForce GTX 580
> OS: Ubuntu Server 10.04.3 LTS 32bit
> CUDA: cuda toolkit 4.1

Oh, 32-bit OS is pretty unusual for the GPU stuff.  This may make your
testing more valuable to us. ;-)

> I refer page:

You might also want to see further messages in that thread (click
"thread-next" many times).

> Here are the results:
> myrice@...128-desktop:~/GSoC/magnum-jumbo$ -te=0
> Warning: doing quick benchmarking - the performance numbers will be
> inaccurate
> Benchmarking: Traditional DES [128/128 BS SSE2]... DONE
> Many salts:     4339K c/s real, 4339K c/s virtual
> Only one salt:  4160K c/s real, 4160K c/s virtual
> Benchmarking: BSDI DES (x725) [128/128 BS SSE2]... DONE
> Many salts:     153600 c/s real, 153600 c/s virtual
> Only one salt:  153600 c/s real, 153600 c/s virtual
> Benchmarking: FreeBSD MD5 [SSE2i 4x]... Segmentation fault

It sounds like you ran into the same problem that I did, which magnum
has since reported as fixed.  My guess is that maybe the fix was only
applied for 64-bit make targets.  magnum?

Meanwhile, you may run the actual CUDA and OpenCL tests directly.  The

./john | egrep 'cuda|opencl'

will tell you the names of "formats" that use CUDA or OpenCL.  Then you
may test those with:

./john --test --format=phpass-cuda

and so on.



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.