Page 1 of 2

So... tables to generate?

PostPosted: Mon Nov 08, 2010 8:46 pm
by Bitweasil
I've got access to a dual 480 box for a while.

Any table requests?

I'm working on NTLM 6 & 7, and will probably regen the MD5 tables with a new set of indexes, to work around the issues Sc00bz found.

Any other requests of useful things?

Re: So... tables to generate?

PostPosted: Wed Nov 10, 2010 4:38 pm
by Bitweasil
For fun and to play with some numbers, I'm working on a NTLM len8 upperlower table set. Hopefully will be done in a week or so.

Re: So... tables to generate?

PostPosted: Fri Nov 12, 2010 12:28 am
by Bitweasil
Bitweasil wrote:For fun and to play with some numbers, I'm working on a NTLM len8 upperlower table set. Hopefully will be done in a week or so.


Looking at 11GB per perfect table for this, and I'll either make 4 or 5 tables. Looks like I can get 99.2% with 4 and 99.8% with 5, which should be good enough for now.

Re: So... tables to generate?

PostPosted: Fri Nov 12, 2010 3:20 am
by blazer
what are the configs you are thinking?

Re: So... tables to generate?

PostPosted: Fri Nov 12, 2010 3:28 am
by Bitweasil
blazer wrote:what are the configs you are thinking?


For the fullalpha tables, I'm generating the following:

Chain length: 200000
Number of chains per table: 10000000
Number of tables per index: 100

Merging these, I end up with... oh, don't have the exact numbers, but roughly a 72% hit rate per table.

4 tables gets me 99.2%-ish total, 5 gives me 99.8% or so.

As for the others, still working on it.

... and I understand someone has used my tools to generate a private NTLM len8 set. :) I'm hoping to get stats out of him.

Re: So... tables to generate?

PostPosted: Fri Nov 12, 2010 1:35 pm
by Sc00bz
Bitweasil wrote:For the fullalpha tables, I'm generating the following:

Chain length: 200000
Number of chains per table: 10000000
Number of tables per index: 100

Merging these, I end up with... oh, don't have the exact numbers, but roughly a 72% hit rate per table.

Do you mean:
Chain length: 200000
Number of chains per file: 10000000
Number of files per table: 100

because that's 200000 * 10000000 * 100 / 52 ^ 8 = 3.7411 which is near a table work factor of 3.5012x which gets you about 72%.

Re: So... tables to generate?

PostPosted: Fri Nov 12, 2010 2:11 pm
by Bitweasil
Yeah, what Sc00bz said. That's what I get for posting while tired. :)

Re: So... tables to generate?

PostPosted: Sun Nov 21, 2010 11:17 pm
by blazer
any chance of lower num 8 or lower num 9

Re: So... tables to generate?

PostPosted: Sun Nov 21, 2010 11:27 pm
by Bitweasil
blazer wrote:any chance of lower num 8 or lower num 9


For what hash?

I'm working on NTLM upperlowernumberspace len8 currently.

Re: So... tables to generate?

PostPosted: Mon Nov 22, 2010 1:45 am
by blazer
i was thinking either NTLM or MD5, is your NTLM upperlowernumberspace len8 the multi TB one?