Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1bff2b7e646c0c87e8decf87a772f649@smtp.hushmail.com>
Date: Tue, 29 May 2012 02:12:04 +0200
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: GPU options

On 05/29/2012 02:02 AM, Lukas Odzioba wrote:
>> 2012/5/29 magnum <john.magnum@...hmail.com>:
>>> On 05/29/2012 01:26 AM, Lukas Odzioba wrote:
>>> If we will drop platform option as I stated previously, we could use
>>> --device=LIST for both cuda and opencl. For opencl it could print also
>>> platform information.
>>
>> Yes. Do you have (or know of) some trivial CUDA device listing code?
> 
> This one looks good:
> http://gpucoder.livejournal.com/1064.html
> 
> However in my opinion we need:
> -name
> -combined major and minor number

is that like the sm_30?

> -mem sizes in KB or MB
> -clock
> -concurrent copy..
> -number of multiprocessors
> 
> You want me to prepare this code?

Not until you send me a pull request for the broken stuff :)  but after
that, sure if you think it fits your GSoC schedule. I probably wont have
time this week anyway.

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.