Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <624f8eb4bcc45acb788e0a6ff982945e@smtp.hushmail.com>
Date: Wed, 05 Aug 2015 15:40:05 +0200
From: magnum <john.magnum@...hmail.com>
To: john-users@...ts.openwall.com
Subject: Re: CUDA support

On 2015-08-05 03:21, Solar Designer wrote:
> On Tue, Aug 04, 2015 at 11:24:48PM +0200, magnum wrote:
>> Most of our CUDA formats are rather useless and we have no active
>> maintainer for any of them nor for the core CUDA stuff. Unless some
>> active maintainer(s) show up, we'll eventually need to drop the CUDA
>> support.
>>
>> We're not there yet, but from now on bleeding-edge Jumbo defaults to not
>> build for CUDA (hint: nvidia can use OpenCL too). The default until now
>> was to enable CUDA if found.
>>
>> To build with CUDA, use "./configure --enable-cuda" from now on.
>
> This sounds just right to me.  Maybe we should have ./configure itself
> print the hint about OpenCL being the right choice even for NVIDIA?

Good idea. Looks like this now. I tried not to break col 80 so it's kind 
of ugly but hopefully will be seen and understood. Maybe lines are 
wrapped in this mail though.

Optional libraries/features found:
(...)
CUDA support (default disabled by OpenCL) .. yes * NOTE: OpenCL is 
superior, *
OpenCL support ............................. no  *       even for NVIDIA 
     *
(...)

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.