[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4F854354.6020709@linuxasylum.net>
Date: Wed, 11 Apr 2012 10:39:48 +0200
From: Samuele Giovanni Tonon <samu@...uxasylum.net>
To: john-dev@...ts.openwall.com
Subject: Re: OpenCL tests on HD 7970
On 04/11/2012 06:18 AM, Solar Designer wrote:
> magnum, Lukas, Sayantan, Samuele -
>
> Here's a full set of OpenCL tests for magnum-jumbo on my 7970.
> (I ran these one by one because we do not yet have "format destructors"
> and because a segfault or the like would prevent further tests.)
These are really usefoul! thanks!
> user@...l:~/john/magnum-jumbo/src$ ../run/john -te -pla=1 -fo=ssha-opencl
> OpenCL platform 1: AMD Accelerated Parallel Processing, 2 device(s).
> Using device 0: Tahiti
> Max Group Work Size 256 Optimal local work size 64
> (to avoid this test on next run, put "ssha_LWS = 64" in john.conf, section [Options:OpenCL])
> Local work size (LWS) 64, Keys per crypt (KPC) 4194304
> Benchmarking: Netscape LDAP SSHA OPENCL [salted SHA-1]... DONE
> Many salts: 58798K c/s real, 74898K c/s virtual
> Only one salt: 41943K c/s real, 46603K c/s virtual
i think that if you set KPC to 8388608 (which is 4194304 * 2) you will
see an improvements in the c/s real for many salts.
if that's what happen than i have some thoughts to share about it
Cheers
Samuele
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.
