Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 22 Jun 2012 03:21:33 +0200
From: Frank Dittrich <frank_dittrich@...mail.com>
To: john-dev@...ts.openwall.com
Subject: Re: relbench and changed format names

On 06/22/2012 12:04 AM, Frank Dittrich wrote:
> On 06/21/2012 04:48 PM, Solar Designer wrote:
>> Maybe we can even
>> include that script along with jumbo-6.  In fact, maybe that script
>> shouldn't be specific to this one occasion, but should be a script to
>> turn all historical format names to what they're currently called in
>> whatever revision of jumbo the script is included with.
> 
> This gets tricky because now some formats that used different format
> names in jumbo-5 now use the same format name.

Another problem is that formats which used to distinguish between "Many
salts" and "Only one salt" or short and long passwords in a previous
release don't in a new release and vice versa.

$ grep -A 2 "HMAC MD5" j?-sse*
j5-sse2i-omp.txt:Benchmarking: HMAC MD5 [SSE2i 12x]... DONE
j5-sse2i-omp.txt-Raw:	1870K c/s real, 1870K c/s virtual
j5-sse2i-omp.txt-
--
j6-sse2i-omp.txt:Benchmarking: HMAC MD5 [128/128 SSE2 intrinsics 12x]...
DONE
j6-sse2i-omp.txt-Many salts:	1750K c/s real, 1733K c/s virtual
j6-sse2i-omp.txt-Only one salt:	1110K c/s real, 1110K c/s virtual


Don't ask me why this format got slower (on 32 bit linux, with sse2i and
OMP).
The point is, I had just one benchmark result in jumbo-5, now I have 2.

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.