|
Message-ID: <20130703165157.GA6345@debian> Date: Wed, 3 Jul 2013 20:51:57 +0400 From: Aleksey Cherepanov <aleksey.4erepanov@...il.com> To: john-users@...ts.openwall.com Subject: Re: [dhiru] write up for Hash Runner 2013 On Wed, Jun 26, 2013 at 12:25:11PM +0400, Aleksey Cherepanov wrote: > On Mon, May 27, 2013 at 10:13:20AM +0530, Dhiru Kholia wrote: > After all I got the idea how to improve speed of your format in place > of comparison. If I understand correctly you compare computed hash > with all hashes to crack. You could use hash function and table to > reduce the list of candidates for comparison. Usual hash function for > comparison in john uses first bytes from computed hash. Due to unknown > bits in input hashes you need different hash function - it should use > only bits known in all input hashes. In case there are no such bits or > their amount is not enough, you could create more tricky hash function > to meet following criteria: this hash function should give you the > same results on hashes with all bits and with question marks inside. > In good case you could craft a hash function to pick only one > candidate for comparison (perfect hash function). Though you would > need to generate a hash function for a defined set of hashes (i.e. per > input file) - there could not be one hash function suitable for all > sets. Ugh, forget it. Everything is much easier: you need to get key for hash table as usual before dereferencing but when you fill the hash table you need to unpack ?s (only in part for keys, first bytes) so you need to put input hash into lists in hash table for all possible keys according to ?s. Also you need modified comparison as you have. To improve the speed you could store a binary hash with 0 bits instead of ?s and mask with 1 for known bits and 0 for ?s. Thanks! -- Regards, Aleksey Cherepanov
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.