[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 4 Jun 2015 16:14:38 +0200
From: Agnieszka Bielec <bielecagnieszka8@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: PHC: Parallel in OpenCL
2015-06-04 13:38 GMT+02:00 magnum <john.magnum@...hmail.com>:
> On 2015-06-04 12:07, magnum wrote:
>>
>> On 2015-06-04 00:45, Lukas Odzioba wrote:
>>> She also implemented splitted kernel and it itself also degradated
>>> performance (from 28k to 27k c/s).
this isn't true. http://www.openwall.com/lists/john-dev/2015/06/04/1
>>
>> Isn't the loop kernel using the "zeros" sha function (only)? And the
>> other kernels use the full version? Then I can't see how code size would
>> be larger for a given kernel.
yes, but for amd gcn I have only demo
> looks to me each call is 3*5*128 rounds of SHA512?
yes, 3*5*128 for each call parallel_kernel_loop()
> Note these lines (after my patch):
>
> opencl_init_auto_setup(SEED, 3*5*128*1, split_events,
> warn, 4, self, create_clobj, release_clobj, BINARY_SIZE*3, 0);
>
> autotune_run(self, 3*5*128*1, 0, 1000);
>
> If you change the loop kernel to do only 128 rounds per call, you should
> change it accordingly for opencl_init_auto_setup() but not for
> autotune_run(). The latter is total, the former is how much you do per call.
> If you change to a test vector with another cost, change the *1 accordingly
> for both.
is this necessary? I don't see any difference in performance and if I
want to change *1, code will be complicated like in pomelo
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.
