Tables with len <5 possible?

Discussion of the upcoming GPU accelerated rainbow table implementation
  • Ads

Tables with len <5 possible?

Postby frosty » Sat Sep 22, 2012 5:09 pm

First of all I want to say this is an absolutely awesome piece of software and runs super fast on a GTX 580, it's very impressive.

I have a few questions. I'm downloading the NTLM 8,7 and 6 sets of tables, I've generated my own length 5 table but the parameter for length doesn't accept 4 or below, is there a reason for this?

I realise that tables for smaller lengths are less useful because you can probably brute force these pretty fast without rainbow tables, especially on a fast GPU, but I wanted to gather enough tools to deal with plaintexts of length 1-8 inclusive, is there any way I can either generate rainbow tables for length 1-4 or use Cryptohaze to run without rainbow tables?

Does anyone have a good suggestion for a replacement application if this isn't possible?
frosty
 
Posts: 39
Joined: Sat Sep 22, 2012 4:58 pm

Re: Tables with len <5 possible?

Postby Bitweasil » Sat Sep 22, 2012 6:40 pm

frosty wrote:First of all I want to say this is an absolutely awesome piece of software and runs super fast on a GTX 580, it's very impressive.


Thanks.

I have a few questions. I'm downloading the NTLM 8,7 and 6 sets of tables, I've generated my own length 5 table but the parameter for length doesn't accept 4 or below, is there a reason for this?


If you decide that downloading the tables is taking too long, you can also order them on disk. :)

They're pointless to make rainbow tables for. That's about it - it's faster to brute force than it is to bother with tables. You could go about patching the code, but given that you can literally brute force 1-5 in a few seconds, there's no real point.

I realise that tables for smaller lengths are less useful because you can probably brute force these pretty fast without rainbow tables, especially on a fast GPU, but I wanted to gather enough tools to deal with plaintexts of length 1-8 inclusive, is there any way I can either generate rainbow tables for length 1-4 or use Cryptohaze to run without rainbow tables?

Does anyone have a good suggestion for a replacement application if this isn't possible?


The Cryptohaze-Multiforcer or New-Multiforcer will both allow for brute forcing.
Bitweasil
Site Admin
 
Posts: 912
Joined: Tue Jan 20, 2009 4:26 pm

Re: Tables with len <5 possible?

Postby frosty » Sun Sep 23, 2012 2:32 pm

Yeah I figured, really more than anything it's just convenience of being able to run one app with one set of parameters to find the plaintext of an unknown length, I'll tinker with multiforcer and write some batch files to automate it all, thanks for the advice.

Do you accept donations via paypal? I also have a 5970 which is artefacting, when I have time I'm going to lower clocks and stress test and check for errors, if I can get it working error free I might bake the clock speeds in to the BIOS and ship it to you, ought to number crunch fairly fast.
frosty
 
Posts: 39
Joined: Sat Sep 22, 2012 4:58 pm

Re: Tables with len <5 possible?

Postby Bitweasil » Sun Sep 23, 2012 4:04 pm

frosty wrote:Yeah I figured, really more than anything it's just convenience of being able to run one app with one set of parameters to find the plaintext of an unknown length, I'll tinker with multiforcer and write some batch files to automate it all, thanks for the advice.


That's the intended use. The Multiforcer actually started life as a "pre-filter" for the rainbow tables. The intent was to look for the short stuff with the Multiforcer, then use rainbow tables for the larger spaces. This use case remains, though the Multiforcer has grown massively since then and is still in a rework.

Do you accept donations via paypal?


I do. See front page of the site. Alternately, if you donate enough, I may be convinced to ship a hard drive loaded with the tables to you. :) Shoot me an email if you want to discuss.

I also have a 5970 which is artefacting, when I have time I'm going to lower clocks and stress test and check for errors, if I can get it working error free I might bake the clock speeds in to the BIOS and ship it to you, ought to number crunch fairly fast.


Sure, I'll take a 5970. If it's glitching, I'll stress test it heavily before relying on it, but I can make use of a downclocked card. Let me know.
Bitweasil
Site Admin
 
Posts: 912
Joined: Tue Jan 20, 2009 4:26 pm

Re: Tables with len <5 possible?

Postby frosty » Mon Sep 24, 2012 2:24 am

Makes sense, although tables for length 1-5 even at max charset would be tiny, I think length 5 are 200mb so 1-4 would be negligable I guess. It would be nice to just dump all tables in one directory and run a single unknown length hash against them and leave it running without having to come back to it, especially if it can iterate through lengths in order from smallest to largest. All in all I guess that's fairly lazy though :)

I can't believe I didn't see that button, thanks. It won't be a lot I'm afraid enough for a few beers. The as far as I can tell the GPU and Mem are fine on the card, the VRMs were the issue when overclocking, reasonably sure that slightly less power draw and it'll be fine, it stops artefacting with about 10-15Mhz shaved off the core speeds. When I have time to take apart my test rig and test I'll let you know, the card is quite long, 12.5" from memory, not a lot of cases hold it comfortably.
frosty
 
Posts: 39
Joined: Sat Sep 22, 2012 4:58 pm

Re: Tables with len <5 possible?

Postby Bitweasil » Tue Sep 25, 2012 1:47 am

frosty wrote:Makes sense, although tables for length 1-5 even at max charset would be tiny, I think length 5 are 200mb so 1-4 would be negligable I guess. It would be nice to just dump all tables in one directory and run a single unknown length hash against them and leave it running without having to come back to it, especially if it can iterate through lengths in order from smallest to largest. All in all I guess that's fairly lazy though :)


It's a known weakness, and will eventually be addressed. I just don't have nearly as much time as I'd like...

I can't believe I didn't see that button, thanks. It won't be a lot I'm afraid enough for a few beers.


No problems! It's appreciated! It actually doesn't go to beer - it goes to hardware & hosting. Staying up to date on GPUs for development isn't cheap! :)

The as far as I can tell the GPU and Mem are fine on the card, the VRMs were the issue when overclocking, reasonably sure that slightly less power draw and it'll be fine, it stops artefacting with about 10-15Mhz shaved off the core speeds. When I have time to take apart my test rig and test I'll let you know, the card is quite long, 12.5" from memory, not a lot of cases hold it comfortably.


I have cases that can hold it. I'm familiar with the size - I have some 6990s as well, and they fit fine. They live in Chenboro cases with a lot of airflow. So if you're unhappy with it, I'd be happy to give it a home! Or, at least, run it hard until it burns out fully... I'm fine with dropping the clocks on cards. I do that anyway a lot of the time to reduce power consumption and heat.
Bitweasil
Site Admin
 
Posts: 912
Joined: Tue Jan 20, 2009 4:26 pm

Re: Tables with len <5 possible?

Postby frosty » Tue Sep 25, 2012 9:30 am

You're welcome to spend it on whatever you want, beer is just my personal suggestion :D

I actually need to dig out the HSF for this thing I bought a custom cooler for it which worked really well until it started artefacting, it's laying in pieces on my workbench at the moment, I may just fit the original again before testing.

Not enough hours in the day :s
frosty
 
Posts: 39
Joined: Sat Sep 22, 2012 4:58 pm


Return to GPU Rainbow Tables

Who is online

Users browsing this forum: No registered users and 2 guests

cron