Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 17 Apr 2013 08:54:15 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: [patch] sse/xop implementation of raw-sha512

On Mon, Apr 15, 2013 at 01:53:32AM +0200, magnum wrote:
> I added OMP support. Scales well on Intel, but worse on AMD:

Thanks!  I made some further optimizations, mostly trivial stuff.  With
this, the speed on FX-8120 with OpenMP is 11M c/s.  Patch attached.

In -hotcold.diff, also attached, I tried changing crypt_key from the
current separate 8 regions (unnecessary complexity and pressure on the
caches and TLB) to just 2, hot and cold - but somehow this did not speed
things up on two machines (FX-8120 and 2xE5649).  Hence separate patch.

When optimizing this for real, I think the cold portion should be gone
completely, and instead cmp_exact() should recompute the SHA-512 hash
when needed (rarely).  Also, we should precompute pieces of the first
few steps, and skip the last few (perhaps 3), and revert a few before
those.  For reuse in iterated formats based on SHA-512, we need the full
implementation as well, though.

Alexander

View attachment "john-rawSHA512_ng_fmt-opt.diff" of type "text/plain" (7519 bytes)

View attachment "john-rawSHA512_ng_fmt-hotcold.diff" of type "text/plain" (4358 bytes)

Powered by blists - more mailing lists

Your e-mail address:

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