Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 16 Sep 2015 19:02:37 +0200
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: larger bitmaps and hash tables

On 2015-09-16 05:30, Solar Designer wrote:
> On Wed, Sep 16, 2015 at 05:57:21AM +0300, Solar Designer wrote:
>> #define PASSWORD_HASH_SIZE_6            0x40000000
>> #define PASSWORD_HASH_SHR               3
>>
>> real    0m50.080s
>> user    3m9.933s
>> sys     0m18.377s
>
> The above combined with magnum's ldr_split_line() fix:
>
> commit fd3c440015ab733d68828f611702ee5ca3aa2daf
> Author: magnum <john.magnum@...hmail.com>
> Date:   Wed Sep 16 03:37:11 2015 +0200
>
>      Move the logic for "Disabled Formats" out of loader.
>
> gives:
>
> real    0m49.570s
> user    3m8.620s
> sys     0m18.560s
>
> With raw-md5's non-OpenMP max_keys_per_crypt increased from 12 to 60
> (which should reduce function call overhead and make bitmap prefetching
> more effective):
>
> real    0m48.515s
> user    3m2.167s
> sys     0m18.777s
>
> Patch attached.  I didn't even try tuning max_keys_per_crypt yet -
> I only tried the 5x increase.  I similarly didn't try tuning
> CRK_PREFETCH yet (it's currently at 64, which is higher than the 60, so
> only the latter takes effect).
>
> This is getting dangerously close to Fred's reported speed for MDXfind.
> Just another 10% and we'll be there.

This patch is not committed yet (most others are). However, I have 
committed a patch where params.h defines PH_MASK_6 as 
(PASSWORD_HASH_SIZE_6 - 1), and changed ALL formats to use PH_MASK_6. So 
we could commit this, or something similar if/when desired.

BTW we now get some warnings from clang and/or with 32-bit builds.

loader.c:211:18: warning: comparison of unsigned expression >= 0 is 
always true
       [-Wtautological-compare]
         } while (--size >= 0);
                  ~~~~~~ ^  ~
loader.c:212:11: warning: comparison of unsigned expression < 0 is 
always false
       [-Wtautological-compare]
         if (size < 0)
             ~~~~ ^ ~

On another OT note we now reverse last three steps for single SHA-2, 
with boost as expected.

magnum

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.