CUDA Multiforcer 0.7 Source

Discussion and support for the CUDA Multiforcers (Windows and Linux)
  • Ads

Re: CUDA Multiforcer 0.7 Source

Postby xdunk » Wed Aug 05, 2009 3:44 am

i have a huge md5 hashlist (about 1,000,000+) to crack with ver 0.7. so i tried it, then give me this Msg : "MAX LEVELS REACHED. EXITING. "
then i searched the source code and find this msg at line 32 in Multiforcer-common.c:
"define MAX_LEVELS 100000"
is this means the max hashlist limits is 100,000?
so how to take this limit away? thx!
xdunk
 
Posts: 7
Joined: Wed Aug 05, 2009 3:29 am

Re: CUDA Multiforcer 0.7 Source

Postby Bitweasil » Wed Aug 05, 2009 8:57 am

Your list is sorted.

Randomize it.

sort -k 1.2 filename

^ will do it - that sorts by the second character, letting Quicksort do it's thing. It performs really badly on already sorted lists - I need to add a check for this.
Bitweasil
Site Admin
 
Posts: 912
Joined: Tue Jan 20, 2009 4:26 pm

Re: CUDA Multiforcer 0.7 Source

Postby xdunk » Thu Aug 06, 2009 12:28 am

:D
ok! it's working now!
thx!

on my GTX 285:

Done: 99.99% Step rate: 69.0M/s Search rate: 73119368.1M/sec

that's a great job!
xdunk
 
Posts: 7
Joined: Wed Aug 05, 2009 3:29 am

Re: CUDA Multiforcer 0.7 Source

Postby Bitweasil » Thu Aug 06, 2009 12:42 pm

Ah... you have a GT200 core.

I'm guessing you can go faster if you don't mind a badly lagged UI.

Try passing:

--threads 512 --blocks 512 -m 500

That should up the speed a bit, if you're not using that already.

You can also *try* passing -l in - that uses a 512MB bitmap in global memory, which is slow, but on large hash lists may be faster than the search of the hash list it's doing. Try it & see.
Bitweasil
Site Admin
 
Posts: 912
Joined: Tue Jan 20, 2009 4:26 pm

Re: CUDA Multiforcer 0.7 Source

Postby xdunk » Mon Aug 10, 2009 1:37 am

:lol:
i'v done this already...
the speed above is the "top" speed now~

and one opinion i want to give you: like JTR' john.pot file, save the result done before in it, Not load cracked password again and again~ that's will be a lot easy to manage Huge hashlist 8-)

thx ur great job again~ :)
xdunk
 
Posts: 7
Joined: Wed Aug 05, 2009 3:29 am

Re: CUDA Multiforcer 0.7 Source

Postby klaida » Tue Aug 11, 2009 9:37 am

Hi Beatweasil! I'm loving the work you're doing with Multiforcer, but I've a question: will you add support to other hash algorithms (oracle, sha-1, etc...) ? Thanks :mrgreen:
klaida
 
Posts: 1
Joined: Tue Aug 11, 2009 9:32 am

Re: CUDA Multiforcer 0.7 Source

Postby xdunk » Thu Aug 13, 2009 1:54 am

ok
i'wrong...

After i update the video card driver to 6.14.11.8608 then the result is so diffirent:

11.1% Step rate: 472.1M/s Search rate: 500254678.7M/sec

that's amazing... :!:
xdunk
 
Posts: 7
Joined: Wed Aug 05, 2009 3:29 am

Re: CUDA Multiforcer 0.7 Source

Postby Bitweasil » Thu Aug 13, 2009 4:50 am

Hrm. That... doesn't seem right, but I've not tried the new drivers. Is it finding stuff at that search rate? I didn't think it would run that fast with 1M+ entries... it shouldn't?
Bitweasil
Site Admin
 
Posts: 912
Joined: Tue Jan 20, 2009 4:26 pm

Re: CUDA Multiforcer 0.7 Source

Postby xdunk » Thu Aug 13, 2009 8:29 am

: ;) ok,here's the detail:
VIDEO CARD GTX285 : LeadTek GTX 285 , 1G DDR3, PUBLIC VERSION.
overclocking to : 720MHZ/1640MHZ/1408MHZ
windows xp 32 bit, 2G DDR2, INTEL E8400. Q45 INTEL MAINBOARD.

cmd:
CUDA-Multiforcer_7 --hashtype=MD5 --hashfile=1M_r.txt --min=13 --max=14 -c charsetlowernumeric --thread 512 --blocks 512 -m 500 -l

Device 0: "GeForce GTX 285"
Number of cores: 240
Clock rate: 1.66 GHz
Single charset loaded.
Loading & sorting hashes. This may take a while.
Returning from sort.
Hashes loaded (1059558 hashes)
Done with copy to constant
Launching kernel for password length 13

MD5 : C162DE19C4C3731CA3428769D0CD593D: 'aaaaaaaaaaaaa' 0x61616161616161616161616161
Done: 0.00% Step rate: 1.$M/s Search rate: 1.$M/sec
Done: 0.00% Step rate: 490.8M/s Search rate: 520034163.1M/sec
Done: 0.00% Step rate: 513.8M/s Search rate: 544371472.2M/sec
Done: 0.00% Step rate: 513.7M/s Search rate: 544331054.2M/sec
Done: 0.00% Step rate: 513.5M/s Search rate: 544042702.8M/sec
Done: 0.00% Step rate: 512.8M/s Search rate: 543301970.4M/sec
....
=======================================================================
when i make a change: min=1 and max = 6 :
...
...
MD5 : FFD5A936840B36CB7A6B282477A51C26: 'as1200' 0x617331323030
MD5 : FFE698587700ECEDABE1FE4671D64349: 'hsy123' 0x687379313233
------------------------------------------
Compute done: Reference time 5.0 seconds
Stepping rate: 437.5M MD4/s
Search rate: 463591387.4M NTLM/s

=======================================================================
and the 1-6 lowernumeric options give me about 60,000 MD5 hash results

so, this speed is really amazing...
xdunk
 
Posts: 7
Joined: Wed Aug 05, 2009 3:29 am

Re: CUDA Multiforcer 0.7 Source

Postby Bitweasil » Thu Aug 13, 2009 12:40 pm

Alright... I'll have to play with the new drivers. I really don't understand how my code should be running that fast on 1M hashes - the new cards aren't *that* much faster.

What's your single hash rate?
Bitweasil
Site Admin
 
Posts: 912
Joined: Tue Jan 20, 2009 4:26 pm

PreviousNext

Return to CUDA Multiforcers

Who is online

Users browsing this forum: No registered users and 1 guest

cron