Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Mon, 18 Apr 2016 13:04:25 +0000
From: Francois Gaudreault <fgaudreault@...ecure.ca>
To: "john-users@...ts.openwall.com" <john-users@...ts.openwall.com>
Subject: Re: openmpi/opencl best practices

Hi Dan,

Have you looked at VCL to achieve the OpenCL clustering? While I did not use it myself, I’ve seen some success stories and howto on the web. I believe the VCL will allow you to have a heterogeneous setup. Not sure if the cluster will work with the Laptop + EC2 instance connectivity, I assume that thing will use multicast to manage the node memberships.

FG





On 2016-04-18, 3:00 AM, "Dan Tentler" <dan@...nlabs.com> wrote:

>Hey Folks, 
>
> Some of you may remember a thread I started back last October or so - I
>lost a passsword for a .dmg image and I've been messing around trying to
>crack it ever since. I finally got jtr compiled with opencl (thanks
>Solar!) and I now have one running copy on ec2 and one running copy on a
>laptop with a decent card in it. At this point I'd like to start
>combining all of the machines I have with video cards to try and attack
>this hash, and I have a few questions: 
>
>1) last time I tried this, it was an endeavor in getting mpich setup,
>using mpirun and all that - is that still the same, or does jtr now have
>some sort of built-in functionality to do distribution? is there a
>best-practice here? 
>
>2) can you distribute between different cards/technologies? i.e. some
>hosts with just cpu, some hosts with nvidia and other hosts with amd
>cards? or does everything have to be completely uniform? 
>
>3) is there windows support for cuda/opencl yet? It would be amazing to
>have a cracking pool that's cross platform. 
>
>Thanks in advance! 
>
>-Dan 
>
>  

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.