Page 1 of 2

Can i accelerate it?

PostPosted: Wed Sep 22, 2010 10:15 pm
by pacifiste95
Hello,

i'm using Cuda-multiforcer on my windows seven, i have the latest driver for my graphic card.

My GPU is Nvidia GTX260M, and i have a Intel Quad core Q9000@2,4ghz.

I'm trying to crack NTLM hash, can i accelerate it?

c:\>CUDA-Multiforcer.exe -c charsets/charsetfull -f hashntlm.txt -o ntlmpw.txt --m
in 9 --max 10 -h NTLM
Cryptohaze.com CUDA Multiforcer (multiple hash brute forcer)
by Bitweasil
Version 0.72, length 0-14
Currently supported hash types: MD5 FASTMD5 MD4 FASTMD4 NTLM FASTNTLM SHA1 FASTS
HA1
Hash type: NTLM
CUDA Device Information:
Device 0: "GeForce GTX 260M"
Number of cores: 112
Clock rate: 1.25 GHz
Performance Number: 17500
Note: Performance number is clock in mhz * core count, for comparing devices.
Single charset loaded.
Loading & sorting hashes. This may take a while.
Hashes loaded (4 hashes)
You may want to consider the FAST[hash] option if it exists.
Launching kernel for password length 9
Done: 0.00% Step rate: 108.0M/s Search rate: 432.2M/sec


What do you think about my step and search rate?

Thanks a lot !

Re: Can i accelerate it?

PostPosted: Wed Sep 22, 2010 11:16 pm
by blazer
try this

CUDA-Multiforcer.exe -c charsets/charsetfull -f hashntlm.txt -o ntlmpw.txt --m
in 9 --max 10 -h NTLM -m 500 -t 512 -b 14 -l

and post the results up

Re: Can i accelerate it?

PostPosted: Thu Sep 23, 2010 11:06 am
by pacifiste95
i have this error : Cuda error: too many resources requested for launch.

what can i do ? :?

c:\>CUDA-Multiforcer.exe -c charsets/charsetfull -f hashntlm.txt -o ntlmpw.txt --min
0 --max 14 -h NTLM -m 500 -t 512 -b 14 -l
Cryptohaze.com CUDA Multiforcer (multiple hash brute forcer)
by Bitweasil
Version 0.72, length 0-14
Currently supported hash types: MD5 FASTMD5 MD4 FASTMD4 NTLM FASTNTLM SHA1 FASTS
HA1
Hash type: NTLM
CUDA Device Information:
Device 0: "GeForce GTX 260M"
Number of cores: 112
Clock rate: 1.25 GHz
Performance Number: 17500
Note: Performance number is clock in mhz * core count, for comparing devices.
Single charset loaded.
Loading & sorting hashes. This may take a while.
Hashes loaded (6 hashes)
You may want to consider the FAST[hash] option if it exists.
Launching CPU kernel for password length 0
Launching kernel for password length 1


------------------------------------------

Compute done: Reference time 0.0 seconds
Stepping rate: 0.0M MD4/s
Search rate: 0.0M NTLM/s

Launching kernel for password length 2


------------------------------------------

Compute done: Reference time 0.0 seconds
Stepping rate: 0.0M MD4/s
Search rate: 0.0M NTLM/s

Launching kernel for password length 3
Done: 83.33% Step rate: 1.$M/s Search rate: 1.$M/sec

------------------------------------------

Compute done: Reference time 0.0 seconds
Stepping rate: 0.0M MD4/s
Search rate: 0.0M NTLM/s

Launching kernel for password length 4
Cuda error: too many resources requested for launch.

Re: Can i accelerate it?

PostPosted: Thu Sep 23, 2010 1:37 pm
by blazer
try this one

CUDA-Multiforcer.exe -c charsets/charsetfull -f hashntlm.txt -o ntlmpw.txt --m
in 9 --max 10 -h NTLM -m 500 -t 512 -b 14

Re: Can i accelerate it?

PostPosted: Thu Sep 23, 2010 3:58 pm
by pacifiste95
This is the same...no?

Re: Can i accelerate it?

PostPosted: Thu Sep 23, 2010 5:41 pm
by pacifiste95
blazer wrote:try this one

CUDA-Multiforcer.exe -c charsets/charsetfull -f hashntlm.txt -o ntlmpw.txt --m
in 9 --max 10 -h NTLM -m 500 -t 512 -b 14



i have the same error... Cuda error: too many resources requested for launch. ;)

Re: Can i accelerate it?

PostPosted: Thu Sep 23, 2010 5:58 pm
by Bitweasil
That error basically means you're throwing too many threads at it.

Try dropping the '-t' parameter back to 256. It should fix that launch error. Then play with the block count to optimize performance.

Re: Can i accelerate it?

PostPosted: Thu Sep 23, 2010 6:05 pm
by pacifiste95
Bitweasil wrote:That error basically means you're throwing too many threads at it.

Try dropping the '-t' parameter back to 256. It should fix that launch error. Then play with the block count to optimize performance.


Thanks, for the block count => lower is better ?

Re: Can i accelerate it?

PostPosted: Thu Sep 23, 2010 9:34 pm
by Bitweasil
Depends on how long you're running it for.

If you've got smaller -m values, a smaller block size is better. It's kind of hard to fully explain... I really should sit down and write up a howto.

Re: Can i accelerate it?

PostPosted: Thu Sep 23, 2010 10:05 pm
by pacifiste95
c:\>CUDA-Multiforcer.exe -c charsets/charsetfull -f mika.txt -o passwdmika252.tx
t --min 5 --max 11 -h NTLM -t 256 -m 200
Cryptohaze.com CUDA Multiforcer (multiple hash brute forcer)
by Bitweasil
Version 0.72, length 0-14
Currently supported hash types: MD5 FASTMD5 MD4 FASTMD4 NTLM FASTNTLM SHA1 FASTS
HA1
Hash type: NTLM
CUDA Device Information:
Device 0: "GeForce GTX 260M"
Number of cores: 112
Clock rate: 1.25 GHz
Performance Number: 17500
Note: Performance number is clock in mhz * core count, for comparing devices.
Single charset loaded.
Loading & sorting hashes. This may take a while.
Hashes loaded (1 hashes)
You may want to consider the FAST[hash] option if it exists.
Launching kernel for password length 5
Done: 73.43% Step rate: 279.4M/s Search rate: 279.4M/sec

------------------------------------------

Compute done: Reference time 3.5 seconds
Stepping rate: 0.0M MD4/s
Search rate: 0.0M NTLM/s

Launching kernel for password length 6
Done: 99.55% Step rate: 278.4M/s Search rate: 278.4M/sec

------------------------------------------

Compute done: Reference time 206.2 seconds
Stepping rate: 0.0M MD4/s
Search rate: 0.0M NTLM/s

Launching kernel for password length 7
Done: 79.92% Step rate: 272.8M/s Search rate: 272.8M/sec
NTLM:E7590FCEE551C84C4176019DE1E31698:Be12mmp:0x426531326d6d70
All hashes cracked.


I think it's a little bit slow... with a custom block count i can optimize it, and how?

thanks