Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sat, 5 Jan 2013 09:15:05 +0530
From: Dhiru Kholia <dhiru.kholia@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: BLAKE2 (was: New plugin load order magic)

On Sat, Jan 5, 2013 at 6:58 AM, Solar Designer <solar@...nwall.com> wrote:
> On Sat, Jan 05, 2013 at 01:10:43AM +0100, magnum wrote:
> the whole addition of BLAKE2 may need
> discussion.  If we add it at all, then we should have not only BLAKE2b,
> but also BLAKE2s.  The format names should probably be blake2b (not
> blake2-512) and blake2s.  Alternatively, we can use just blake2 (for
> BLAKE2b aka BLAKE2) and blake2s (for BLAKE2s).

Using name blake2 (for BLAKE2b aka BLAKE2) and blake2s (for BLAKE2s)
sounds good.

> Then, Dhiru included the slow reference implementation of BLAKE2b
> instead of the SIMD implementations.  I think we should include all SIMD
> implementations that are available from BLAKE2 team, plus the reference
> implementations for non-x86 and for older x86 CPUs.

I can try doing this but I am not familiar with SIMD business.

> This becomes some effort for unclear purpose.  Publicity for JtR?
> Being ready just in case?  Anyway, I don't mind this being done.  The
> current low speeds that we have for BLAKE2b are not good publicity.

Speed is not that bad on 64-bit machines.

$ ../run/john -fo:raw-blake2-512 -t # 1 core of AMD FX-8120
Benchmarking: BLAKE2 512 [32/64]... DONE
Raw:	2337K c/s real, 2337K c/s virtual

-- 
Dhiru

Powered by blists - more mailing lists

Your e-mail address:

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