[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <7df5d7f426d3b127d56a535563e9da39@smtp.hushmail.com>
Date: Wed, 10 Jul 2013 04:21:03 +0200
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: mscash2 multi-device -test=0 (was: bug: GPU use in CPU-only formats)
On 9 Jul, 2013, at 11:04 , magnum <john.magnum@...hmail.com> wrote:
> On 8 Jul, 2013, at 17:36 , Sayantan Datta <std2048@...il.com> wrote:
>> On Friday 05 July 2013 05:23 AM, magnum wrote:
>>> I think that "decreasing order" requirement should be taken care of by the format, not the user. When using --device=gpu you select all GPU devices available but on Bull they happen to come in increasing order of performance.
>>
>> Fixed both of the issues in bleeding.
>
> This now works fine on my laptop too, using -t. For some reason -t=0 doesn't work unless given a specific order:
Sayantan, please have a look at commit 04da5ba. I fixed several issues. In my opinion the code was totally bogus - it queried the wrong things, did stuff in wrong order, there was an "if CPU" to work around these very bugs and worst of all there were minimum values that (apart from crashing on weak devices) negated the whole idea of environment LWS/GWS. The latter made the self test take 12 seconds on my laptop even with --test=0. After my fixes it takes three seconds.
The patch only affects --test=0 or when using LWS/GWS environment variables. The format now works fine with single or multiple devices on Bull as well as on my laptop.
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.
