[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <BACE5E6642DD4177A65E0FE27B134916@D9VGLK61>
Date: Sat, 14 May 2011 12:58:56 -0500
From: "JimF" <jfoug@....net>
To: <john-dev@...ts.openwall.com>
Subject: Re: Changes to john's unique program.
NOTE,
The code presented has been used by me for quite a while (building and
maintaining wordlists I use). I made comment about this a few months back,
when Solar listed improving the speed of unique. However, the addictions
of -cut=num, -cut=LM and -mem are recent changes (only over the last couple
of days). The -imp=, -ex_file= and -ex_file_only have been used a lot. I
KNOW they work well. The -cut=num and -cut=LM should be pretty trivial,
and I think should work just fine.
The -mem= is new. I know on the mem, it can crash a run, if the -mem size
selected is too large. The 'assumptions' I have taken on mem, is that
1<<num is the size of the hash table (exactly the same as in params.h), and
that the buffer size is (1<<num)*64 which is the same as in params.h when
num==20.
Jim.
----- Original Message -----
From: "JimF" Sent: Saturday, May 14, 2011 12:38 PM
> I have placed a patch to the unique.c (unique program) This patch can be
> placed over 1.7.7-jumbo1 (or the jumbo-2-rc1). It adds this to unique:
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.
