[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 24 Sep 2015 20:59:45 +0200
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: Agnieszka's GSoC formats
On 24/09/15 19:59, Agnieszka Bielec wrote:
> 2015-09-22 1:57 GMT+02:00 magnum <john.magnum@...hmail.com>:
>> On 22/09/15 01:14, Agnieszka Bielec wrote:
>>>
>>> 2015-09-13 21:04 GMT+02:00 Agnieszka Bielec <bielecagnieszka8@...il.com>:
>>>>
>>>> I did PR but it's not ideal.
>>>> additionally Lyra2-opencl using new computations of gws and lws
>>>> doesn't work on 960m
>>>
>>>
>>> repaired, wasn't working also on super, but on test on github
>>> lyra2-opencl fails, can I got an access to this opencl-device?
>>
>>
>> That build-bot is using the CPU device of an fglrx driver. Super's device 3
>> should be the same, did you try that?
>
> are you sure that this is xeon?
I meant same OpenCL driver, not same CPU. I think they run some AMD CPU
with XOP (and it's virtual, but I presume the real gear is the same). I
have no shell access to those hosts, it's a free service.
http://www.travis-ci.org
> can I get lws and gws values for test of lyra2-opencl
> lyra2-opencl has the requirement that LWS must be at least 2 for
> default options, there is also lyra2 with option of nparaller=3 in
> fmt_test
You could always add some temporary debug output to your format, in that
branch of yours. Once you push it, the build bot will use it and output
whatever you want to know. Last time I did something similar, it ran on
"AMD Opteron(tm) Processor 6376" and fglrx was Catalyst 14.1 betav1.3
(from some native Ubuntu package).
> btw. I repaired rest of the formats which I broked down. dynamic(md5)
> fails but also fails on my laptop witchout any of my changes.
Great, I'll try it out and see if I can nail that md5 problem.
Thanks,
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.
