Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20150806121446.GC17245@openwall.com>
Date: Thu, 6 Aug 2015 15:14:46 +0300
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: PHC: Argon2 on GPU

Agnieszka, magnum -

On Tue, Aug 04, 2015 at 12:10:33AM +0200, Agnieszka Bielec wrote:
> none@...e ~/Desktop/rr_coal_2/run $ ./john --test --format=argon2i-opencl
> Benchmarking: argon2i-opencl [Blake2 OpenCL]...
> memory per hash : 1000.00 kB
> Device 0: GeForce GTX 960M
> using different password for benchmarking
> DONE
> Speed for cost 1 (t) of 3, cost 2 (m) of 1000, cost 3 (l) of 1
> Many salts:     2133 c/s real, 2118 c/s virtual
> Only one salt:  2118 c/s real, 2133 c/s virtual

Agnieszka - why are you testing this on your small GPU, and at 1 MB?
Why not tune it (both 2d and 2i) the same as we did for the yescrypt vs.
Lyra2 test, so likely 1.5 MB?  I am interested in results that would be
directly comparable to those you had for yescrypt and Lyra2.

> again the same problem as with yescrypt, it's auotuned properly after
> I divided MEM_SIZE / 4 here:
> 
> opencl_init_auto_setup(SEED, 0, NULL,
>         warn, 4, self, create_clobj, release_clobj, MEM_SIZE/4, 0);

magnum - please comment on this, or have you commented on the issue in
another john-dev thread already?

Thanks,

Alexander

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.