Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sat, 27 Jul 2013 03:23:29 +0400
From: Solar Designer <>
Subject: Re: Parallella: bcrypt

Hi Katja,

On Sat, Jul 27, 2013 at 01:00:42AM +0200, Katja Malvoni wrote:
> I implemented interleaving two bcrpyt instances in C (calls to BF_encrypt
> outside of main loop use optimized BF_2ROUND). I'm getting 947 c/s. Plan
> for tomorrow is to optimize interleaving in assembly.

Cool.  This is higher than what you got for plain C code before, right?

BTW, you may implement the preload of two P's into 36 registers at C
level (well, unless the compiler is somehow too shy to use this many
registers for local variables), even before starting with assembly.

> Code:

(I haven't looked yet.)


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.