Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 7 May 2013 11:35:30 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: Enhancements to pbkdf2-sha256

On Tue, May 07, 2013 at 09:10:39AM +0200, magnum wrote:
> We are definitely faster with smd5 (now supported in md5crypt format). With OMP we get ~200K

Right.

> and with -fork we might do 50% more (or will it decrease the clock too much for that?

The clock rate will be the same as with OpenMP, but the increase is a
lot less than 50% because our md5crypt efficiency with OpenMP is quite
good.  There's simply not that much room for improvement.  So OpenMP
gives us ~203k, and --fork gives us ~214k.

--fork is much faster than OpenMP for formats that have poor OpenMP
scalability.  That's mostly fast ones like LM.  Not md5crypt.

> It's a pity you cant --test with --fork).

Yes.  I am considering adding an equivalent of MPI gather via a shared
memory window with locks from OpenMP or libpthread, but that's for later.

[ I am also considering keeping the hash database in such shared memory
window (as opposed to the current copy-on-write pages), and using an
rwlock on it.  It's a pity OpenMP does not provide rwlocks, so we'll
have to use libpthread directly or emulate an rwlock with per-process
OpenMP locks.  It'd be a weird hybrid process/thread model, yet it has
its advantages and it supports our existing formats. ]

> We should split aix-ssha into three different formats (in one same source file).

Yes, because they're of different speed, so those hashes are better
cracked separately (more candidate passwords may be tested against the
faster hashes).

> We currently benchmark a mix of test vectors :-/

We only use first two salts for benchmarking, which currently (in
aix_ssha_fmt.c in bleeding) correspond to {ssha1}06.  That's just 128
SHA-1's, right?  I am surprised the speed is not higher.

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.