Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <3a301d3e-5936-48f1-9c5d-fac4dee8a5ea@jeffunit.com>
Date: Thu, 26 Oct 2023 12:39:10 -0700
From: jeff <jeff@...funit.com>
To: john-users@...ts.openwall.com
Subject: Re: very low cpu usage when cracking one password

On 10/26/2023 11:41, Solar Designer wrote:
> On Thu, Oct 26, 2023 at 09:52:13AM -0700, jeff wrote:
>> I am running a recent version of john, using descript on a single
>> unknown password.
>> I have a 64 core amd epyc processor.
>> John is running, but it seems to be using only 1 thread.
>> My cpu usage is between 1 and 3 percent.
>> Is there any way to use more threads when I have only 1 psssword to crack?
> The issue is that while John is using 128 threads for hash computation,
> it is only using 1 thread for generating candidate passwords (in your
> case, going through the wordlist and applying rules).  descrypt is so
> fast to compute, especially by 128 threads, that almost all real time
> ends up being spent on generating the candidates.
>
> You can remedy this by using the "--fork=128" option, so that separate
> candidate password streams would be generated, one by each process.
> It's also reasonable to try "--fork=64" and "--fork=32" - these will
> likely be slower, but less cumbersome to manage.
>
> I don't know how well that works on Windows, though.  Please try and let
> us all know.
>
> Alexander
It works great under windows. I am using --fork=32 to keep power 
consumption down.
It is using about 42% of cpu load according to the task manager.

jeff

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.