[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20130713202555.GA25585@openwall.com>
Date: Sun, 14 Jul 2013 00:25:55 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: Parallella: bcrypt
On Sat, Jul 13, 2013 at 10:17:38PM +0200, Katja Malvoni wrote:
> Oh, I made a mistake - execution time of 20 ms doesn't include data
> transfers, it's only computation and waiting for cores to signal done. With
> data transfers included (loading *.srec files on cores, transferring key
> and salt and starting the cores) it's 85.76 ms, 2246.917 ms and 8781.541
> ms.
This means that we should be reducing the communication time. The code
should be loaded just once, not once per hash computed.
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.
