Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Mon, 17 Feb 2014 18:49:22 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: sha512crypt-opencl format fails in xeon phi coprocesor

Claudio,

On Mon, Feb 17, 2014 at 11:02:48AM -0300, Claudio Andr? wrote:
> >I'm trying to use the sha512crypt-opencl kernel with a xeon phi coprocesor.
> >When using the host device, it works smoothly, but using the MIC
> >(coprocesor) it fails to build the kernel cryptsha512_kernel_DEFAULT.cl.
> 
> I have no idea if it works/worked on MIC.

It doesn't, and I think it didn't.

d1a.log.gz downloadable from here:

http://permalink.gmane.org/gmane.comp.security.openwall.john.devel/9545

gives this:

=== Testing sha512crypt-opencl on MIC ===
Device 1: Intel(R) Many Integrated Core Acceleration Card 
Exit code: 139

139 means segfault.

Now that I've tried bleeding-jumbo on our 5110P, I am getting results
similar to what Mr Dupito posted.

I suggest that we create you an account on our machine with the 5110P,
and you debug the issue.

> Anyway, could you please try this?
> 
> $ _TYPE=66 ../run/john  -form=sha512crypt-opencl  -t  --verbosity=5
> $ _TYPE=0 ../run/john  -form=sha512crypt-opencl  -t  --verbosity=5
> 
> To avoid unnecessary Spam, you can send it in private.

Note that Mr Dupito was testing with john-1.7.9-jumbo-7, whereas we may
only care to get bleeding-jumbo to work on more hardware.

Alexander

Powered by blists - more mailing lists

Your e-mail address:

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