Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Mon, 17 Feb 2014 21:12:52 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Cc: Claudio Andr? <claudioandre.br@...il.com>
Subject: Re: sha512crypt-opencl format fails in xeon phi coprocesor

Claudio,

Somehow your recent replies each start a new thread.  I am CC'ing you on
this reply in case this makes it easier for you to reply to the existing
thread.

On Mon, Feb 17, 2014 at 01:55:00PM -0300, Claudio Andr? wrote:
> Intel OpenCL driver/compiler creativity!?!?!

Of course.

> That said, seems to me OpenCL and MIC are not going to be a production 
> quality target for a while. I might solve the problem with raw formats, 
> but I'm not confident things are going to be much better than now.

I'm not confident either.

The primary way to program for MIC is via its native intrinsics (and
then we get to choose between fully native build vs. OpenMP offload).
I think we may achieve pretty good performance at SHA-512 and
sha512crypt when using the native intrinsics.

Alexander

Powered by blists - more mailing lists

Your e-mail address:

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