Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAAufJG6YcBFgD=gT-o6=PaQQYcVM3FVrk6LDtr=QaTFCELsDxA@mail.gmail.com>
Date: Fri, 3 May 2013 11:06:40 -0300
From: Claudio André <claudioandre.br@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: Mulit-gpu using claudio's interfaces

2013/1/19 Sayantan Datta <std2048@...il.com>

> Hi Claudio,magnum,
>


Just to let you know. I always had in my mind that core has to change in
order to create a proper multi-gpu stuff. Now that it did:

Suppose the user do something like this.
$ .john -dev=1,4

The main idea is that:
- it will do the same a fork=2 does
- one process will deal with dev=1 and another with dev=4 (I suppose the
real situation is that JtR has the main process plus only one forked in
this case . But this doesn't matter).
- my expectation is that just a few new lines would be necessary since fork
code is going to do the trick and common-opencl.c was developed with this
in mind.
- almost all opencl formats will start working at once. I expect problems
on Sayantan's 'avoidance' of  shared opencl code.
- so, my expectation is that his code will stay running as they do now.

This is not going to be done anytime soon (by me), but it is not a hard
task if anyone volunteers.

Claudio

Content of type "text/html" skipped

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.