Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Wed, 14 Sep 2011 19:49:27 +0200
From: magnum <rawsmooth@...dband.net>
To: john-dev@...ts.openwall.com
Subject: Re: Re: [moved from john-users] Speed of jtr on your machine?

On 2011-09-14 18:35, jfoug wrote:
> I have found several issues here.  First, the output strings to
> benchmarking, or normal running (and logging), where the format string, and
> algorithm was output was happening PRIOR to init() function being called.
> Well, that works fine if using data that is compiled in (in the md5-gen
> case, it is '[gen]').  However, if we want to harvest that from the md5-gen
> underlying format, then we have to make sure we call init() first.

A good thing with this earlier fmt_init() is in case a format caps the 
max OMP num threads (not sure if any format does that currently, but it 
does make sense in some cases), it will be reflected in the benchmark 
output, the (nxOMP). This can avoid some confusion.

By the way, I guess if a format does that, the capped value will remain 
effective for subsequent formats (when running --test of all formats). 
Not a big deal, we can address it if/when needed.

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.