Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 20 Mar 2012 02:00:36 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: OpenSSL and AES-NI (was: RAR format finally proper)

On Mon, Mar 19, 2012 at 11:20:31PM +0200, Milen Rangelov wrote:
> For small files in the archive I get quite an improvement. When files
> in archive are several megabytes in size, the difference is not that huge
> at all. And when files get several tens of megabytes in size, it really
> does not matter where you use AES-NI or not - speed gets close to zero in
> both cases :(

Maybe this has to do with cache vs. memory reads?  If so, you may try to
improve the locality of reference by decrypting the file in small enough
chunks to fit in L1 data cache, but for multiple candidate passwords
being tested - then move on to the next chunk, and so on.  Each chunk
will then be read from memory to cache only once per the current set of
candidate passwords (e.g., per 1000 of them), not once per candidate.

Disclaimer: I am not familiar with the RAR stuff, so the above might not
actually apply.

Alexander

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.