john-dev - AVX in Intel Sandy Bridge

Follow @Openwall on Twitter for new release announcements and other news

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <20110430102413.GA28893@openwall.com>
Date: Sat, 30 Apr 2011 14:24:13 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: AVX in Intel Sandy Bridge

Hi,

I just found this blog post by Dango-Chu:

http://dango.chu.jp/tripper/20110303.html

Apparently, it might be possible to achieve some speedup by mixing
128-bit and 256-bit AVX ops.  I tried many different combinations, but
not this one.  I will likely give it a try.

Also, yes, the DES code currently in JtR probably doesn't fit in Sandy
Bridge's micro-op cache, but unlike with P4's trace cache this is not a
performance hit (compared to older CPUs) but rather lack of a potential
speedup.  It is unclear how much of a speedup could be possible with
smaller code, if at all - we'd have to pay for that by omitting some
other optimizations.

Alexander

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.