Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 15 Feb 2012 06:56:03 -0800
From: Alain Espinosa <>
Subject: Re: OMP version not use all CPU

On 2/14/12, Alain Espinosa <> wrote:
> Will do and i report back.

You are right, john-omp is working as expected. I see a 25% CPU use in
the Task Manager and think i john was using only one core. More info
by Resource Monitor (happen the same with jumbo and non jumbo

When using 2 threads there is 2 cores (logical cores) with 0% use, 1
with 25% and other with 75%.
When using 4 threads there is 2 cores (logical cores) with 0% use and
the other 2 with 50%.

>> Oh, and the speed you're reporting for LM hashes (25M c/s) is
>> surprisingly low for JtR 1.7.9 on a modern CPU.  As you can see, we have
>> higher speeds for 1.7.9 here:
> I take a look at this numbers. One difference its that i use a laptop
> CPU, other its that i perform a "real" benchmark and not the "-test"
> option. I consistently found differences in "-test" option and "real"
> crack speed. I will test the "-test" option and review the method i
> use to gather performance data.

I review all the process of benchmarking and found my reported
performance accurate. Here is john -test:

>..\john.exe -test --format:nt
Benchmarking: NT MD4 [128/128 SSE2 + 32/32]... DONE
Raw:    18412K c/s real, 18412K c/s virtual

>..\john.exe -test --format:nt2
Benchmarking: NT v2 [SSE2i 12x]... DONE
Raw:    21509K c/s real, 21509K c/s virtual

>..\john.exe -test --format:lm
Benchmarking: LM DES [128/128 BS SSE2]... DONE
Raw:    29076K c/s real, 29459K c/s virtual

>..\john.exe -test --format:mscash
Benchmarking: M$ Cache Hash [Generic 1x]... DONE
Many salts:     14939K c/s real, 14939K c/s virtual
Only one salt:  5845K c/s real, 5845K c/s virtual


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.