Multi-GPU, multi-hash GRT BETA! For Linux... (beta1)

[Bitweasil]

I suppose I should stabilize code & make a release instead of continuing to screw with it... I've been working on the tighter table format, per-position charset support, and some other improvements.

I'm out at coffee shops right now, but I'll look at doing a beta code release soon. The main issue right now is that I don't have NTLM support working yet - just haven't gotten around to porting it.

[RIG]

Ok, I got some new Hardware. Now waiting for realese. Do you think 100k Hashes could be cracked in a few hours?

[Bitweasil]

Ok, I got some new Hardware. Now waiting for realese. Do you think 100k Hashes could be cracked in a few hours?

100k hashes is *well* into the territory that it would be faster to brute force the space...

What's the new hardware?

The delays in release are related to my new table format. I'm trying to get it finished before I push out a full release.

Teaser: The new table format is about 40% the size of the current format with the same performance (actually, probably *slightly* faster since there's less data to read/transfer off the disk).

You'll be able to convert all the current tables to the new format, and I will start distributing in the new format as well.

[RIG]

What's the new hardware?

Teaser: The new table format is about 40% the size of the current format with the same performance (actually, probably *slightly* faster since there's less data to read/transfer off the disk).

You'll be able to convert all the current tables to the new format, and I will start distributing in the new format as well.

40% sounds nice

I got 3 Nvidia 580 GTX and access to 2 Tesla S2050 for some weeks

I'll test the performance of your new version on these cards

[Picch]

I'm getting the same segfault as some others are getting:

Code: Select all

./GRTCrack-CPP -h NTLM -f hash2.txt ../grtntlm7/*.grtGot
1 hashes!
Processing GRT file ../grtntlm7/NTLM-idx0-chr95-cl200000-perfect.grt

Table version:   1
Hash:            NTLM
Password length: 7
Table index:     0
Chain length:    200000
Num chains:      444148604
Perfect table:   Yes
Charset length:  95
Charset:          !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQR
                 STUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~

Elements: 1
Elements per unit: 1
Number of workunits: 1
WU init complete
Cuda error in file 'GRTCandidateHashesMD5.cu' in line 39 : invalid device symbol.
Segmentation fault

Here's my system information:
OS: Ubuntu 10.04 x64
Cards: GTX580 and GTX280
Toolkit Version: 3.5
Dev Driver Version: 260.19.26

[Bitweasil]

Hm...

Alright, I think I have a few things to try. I'll see what I can whip up.