Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 14 Sep 2015 16:33:55 +0800
From: Lei Zhang <zhanglei.april@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: SHA-1 H()

On Sep 8, 2015, at 3:16 PM, Lei Zhang <zhanglei.april@...il.com> wrote:
> 
> On Sep 2, 2015, at 11:52 PM, Solar Designer <solar@...nwall.com> wrote:
>> 
>> Lei - I think you haven't gotten around to introducing AVX-512 ternary
>> logic intrinsics yet, have you?
> 
> I'll look into that.
> 
>> Unfortunately, we don't yet have
>> hardware to test them on, but you could test on Intel SDE or/and by
>> emulating them with macros.
> 
> I've been following the news of Skylake. It's disappointing the desktop product line won't support AVX-512; only Xeon and Xeon Phi will have it, which are not launched yet. For the moment, Intel SDE is the only option...

Ok, I've now introduced the ternary logic instruction to JtR, and get it to work on SDE.

In case it's helpful, here're some benchmark figures of JtR running on SDE:

Benchmarking: Raw-MD4 [MD4 512/512 AVX512F 16x3]... DONE
Raw:    312712 c/s real, 309647 c/s virtual

Benchmarking: Raw-MD5 [MD5 512/512 AVX512F 16x3]... DONE
Raw:    264240 c/s real, 264240 c/s virtual

Benchmarking: Raw-SHA1 [SHA1 512/512 AVX512F 16x]... DONE
Raw:    130772 c/s real, 129490 c/s virtual

Benchmarking: Raw-SHA256 [SHA256 512/512 AVX512F 16x]... DONE
Raw:    70304 c/s real, 70304 c/s virtual

Benchmarking: Raw-SHA512 [SHA512 512/512 AVX512F 8x]... DONE
Raw:    30875 c/s real, 30875 c/s virtual


BTW, SDE runs much more smoothly than I expected. At least those formats listed above ran quite fast on it.


Lei

Powered by blists - more mailing lists

Your e-mail address:

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