Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sat, 2 May 2015 13:16:27 +0200
From: Agnieszka Bielec <bielecagnieszka8@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: [GSoC] John the Ripper support for PHC finalists

2015-05-02 11:44 GMT+02:00 magnum <john.magnum@...hmail.com>:
> On 2015-05-02 06:14, Agnieszka Bielec wrote:
>>
>> well SSE2 interleaving (I forgot to change algorithm_name to "SSE2")
>>
>> a@...l:~/int/run$ ./john --test --format=pomelo --cost=2:2,2:2
>> Will run 8 OpenMP threads
>> Benchmarking: POMELO, Generic pomelo [AVX2]... (8xOMP) DONE
>
>
> Just some non-critical notes: Your algorithm name should change
> automatically

I changed a function call from avx2 to sse2 in the code

> Also, we use to say eg. [SSE2 128/128 4x] for a
> format that calculates 4 hashes in parallel using SIMD, and eg. [SSE2
> 128/128 4x3] when such format is also interleaved by 3x. Please consider
> using this notion since it would make your posted benchmarks much easier to
> tell apart.

it's 4x "internal interleaving" and 2x interleaving for both sse2 and non-simd
4x "internal interleaving" and 4x interleaving for avx2

simd exploits "internal interleaving" in my case

thanks

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ