Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20060729205614.GA7296@openwall.com>
Date: Sun, 30 Jul 2006 00:56:14 +0400
From: Solar Designer <solar@...nwall.com>
To: john-users@...ts.openwall.com
Subject: Re: Optimized for Cell BE

On Sat, Jul 29, 2006 at 11:26:57AM -0700, James Evans wrote:
> Hi -- I'm just curious to see if anyone has thoughts
> about running JtR on the new Cell Processor.

This topic was briefly raised in here before, but it was not seriously
discussed.

> IBM just released a new SDK for it,
> http://www.alphaworks.ibm.com/tech/cellsw?open&S_TACT=105AGX16&S_CMP=DWPA

Interesting stuff.

> I would take a guess that JtR could be optimized to
> run very quickly using the Cell's architecture,

Yes.

> ... it is going to be hard to write code which can
> fully exploit the Cell.

That's true.

> However, I think it may be easier in the case of JtR

Well, one potential way to do it might be to treat SPEs as a special
case of network nodes, once distributed processing support is added.

I wish I had the time to play with this.  Unfortunately, I don't expect
to find the time unless there's commercial demand.  Maybe I could
approach some IBM folks suggesting that this could be a great example of
performance improvement with the Cell.

> (I did see in the code some
> mention of AltiVec, but that would only exploit the
> core of the Cell rather than its SPEs).

Correct.

-- 
Alexander Peslyak <solar at openwall.com>
GPG key ID: B35D3598  fp: 6429 0D7E F130 C13E C929  6447 73C3 A290 B35D 3598
http://www.openwall.com - bringing security into open computing environments

-- 
To unsubscribe, e-mail john-users-unsubscribe@...ts.openwall.com and reply
to the automated confirmation request that will be sent to you.

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.