Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 13 Feb 2013 10:34:22 +0100
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: GPU driver versions (was: remaining DMG issues)

On 13 Feb, 2013, at 10:12 , Solar Designer <solar@...nwall.com> wrote:

> On Tue, Feb 12, 2013 at 09:23:06PM +0100, magnum wrote:
>> I think we should generally just recommend "latest stable" OpenCL drivers.
> 
> "Stable" as defined by whom?

I meant release versions, non-beta.
There's no such thing as a stable AMD driver :)


> I think it is important for JtR to work well with the same driver
> versions that atom recommends for hashcat.  Not only those
> recommendations are not arbitrary and are closely relevant for this kind
> of use, but also it makes users' life easier when they can run both
> hashcat and JtR on the same systems, without having to swap drivers.
> It assists JtR vs. hashcat benchmarks, too.

Yes, I think it's important to test the version Atom requires (currently 13.1). I'm not sure we should generally recommend it, although right now we can. Before new-years eve, Hashcat required a Catalyst version so old that it was simply not installable on late distros. AFAIK this was for reasons that does not apply to JtR (Jumbo worked fine on versions between 12.8 and 13.1 too). I think the blocker bugs were in something low-level Atom uses that we do not. For mostly the same reasons, Hashcat will not support a new GPU product released tomorrow (until Atom codes/builds support for it, and releases a new version) while any old GPU-aware Jumbo would support it without even a re-build.

magnum

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.