Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Thu, 28 Jun 2012 07:34:27 +0200
From: Frank Dittrich <frank_dittrich@...mail.com>
To: john-dev@...ts.openwall.com
Subject: Re: Can keepass get a second test case prior to jumbo-6?

On 06/28/2012 07:09 AM, Dhiru Kholia wrote:
> My commit didn't even touch this file! Can you do pull, "make clean"
> and try building again? Also make sure that you don't have any local
> changes.

Yes I immediately noticed this. Sorry for the noise.
But I wanted to test the impact on total run time for the --test=0 case.
And I was afraid I wouldn't be able to test it fast enough and find out
this causes a regression after jumbo-6 got released by Solar.

I immediately checked after making sure I have a clean tree.


$ ./john --list=build-info
Version: 1.7.9-jumbo-6-RC0
Build: linux-x86-sse2i
Arch: 32-bit LE
$JOHN is ./
Rec file version: REC3
CHARSET_MIN: 32 (0x20)
CHARSET_MAX: 126 (0x7e)
CHARSET_LENGTH: 8
Compiler version: 4.6.3 20120306 (Red Hat 4.6.3-2)
gcc version: 4.6.3

For the non-OMP case, the run time almost doubled.

From

$ time ./john --test=0 --format=keepass
Warning: doing quick benchmarking - the performance numbers will be
inaccurate
Benchmarking: KeePass SHA-256 AES [32/32]... DONE
Raw:	6.6 c/s real, 6.6 c/s virtual


real	0m0.377s
user	0m0.362s
sys	0m0.013s


to

$ time ./john --test=0 --format=keepass
Warning: doing quick benchmarking - the performance numbers will be
inaccurate
Benchmarking: KeePass SHA-256 AES [32/32]... DONE
Raw:	6.2 c/s real, 6.2 c/s virtual


real	0m0.684s
user	0m0.672s
sys	0m0.011s

But that's acceptable.

The run time for the OMP case is disturbing (May be I'll create a new
thread discussing this):

$ time ./john --test=0 --format=keepass
Warning: doing quick benchmarking - the performance numbers will be
inaccurate
Benchmarking: KeePass SHA-256 AES [32/32]... (2xOMP) DONE
Raw:	9.8 c/s real, 5.0 c/s virtual


real	0m43.795s
user	1m23.647s
sys	0m0.087s

But that didn't change with your patch.
I was afraid it would double this time as well, for no good reason if
this is not really a salted hash.

Frank


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.