[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20150317051936.GA8939@openwall.com>
Date: Tue, 17 Mar 2015 08:19:36 +0300
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: 256/128 bit integer arithmatic
Hi Sayantan,
I'm sorry I didn't comment sooner.
On Mon, Mar 09, 2015 at 06:40:23PM +0530, Sayantan Datta wrote:
> Also using less than 128 bits precision of hashes would mean various
> hashes would map to same location,
And that's fine.
> which is against the purpose of building such tables.
Not quite. In fact, I think it's usually pointless to send bigger than
64-bit partial hashes to GPU. We certainly can deal with the rare false
positives on the host.
So perfect hash tables based on 64-bit or smaller partial hashes make
sense to me, _if_ perfect hashes are the right thing to pursue at all.
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.
