[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <9f7375a35815c9dd8e168218ca1eac76@smtp.hushmail.com>
Date: Sat, 29 Aug 2015 21:38:05 +0200
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: optimization of constants in first rounds of raw formats
You tend to post very interesting stuff but often no-one answers!
On 2015-08-27 19:57, Aleksey Cherepanov wrote:
> For raw-sha512, raw-md5 and similar formats, it is possible to reduce
> first round to constant+w[0]. Next rounds can be reduced a bit too. I
> played with it in john-devkit.
I think we need to introduce a separate function for first round. I
already see regression from all branching. This would be trivial, I
might try it out.
In OpenCL, everything is inline or macros so many of these things
actually gets done by the optimizer. We have separate macros for PBKDF2
where the input is always eg. 20 bytes and rest are zero, but they do
very little difference so I suspect the optimizer does most of it for us
anyway.
magnum
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.
