DES Cracker BETA

[schwarzwaldhacker]

Well, "it's not too bad". "Could be better" or something. Because I guess, it is still not always faster than Xserg's?

But I want to check the algorithm first. Because I had some ideas using textures? And then, maybe for Christmas, you might get multi-GPU . But sincerely, right now I am very busy(Just got a new car) and I don't think so that you will get multi-GPU soon, sorry!

[keri]

What id like to see is support for more languages like german or french, czech maybe, characters, that arent used in a default english password.
also an option to specify the searchrange a bit more would be nice , say "atleast x numbers, max 2 non. english characters " and so on.

[LordMike]

lol. See what it can deliver on the 9 GPU system

[schwarzwaldhacker]

What id like to see is support for more languages like german or french, czech maybe, characters, that arent used in a default english password.
also an option to specify the searchrange a bit more would be nice , say "atleast x numbers, max 2 non. english characters " and so on.

Sorry Keri, but this is still BETA. And I am still trying to optimize the algorithm. But a lot of people are downloading it. And therefore I will have a look!

lol. See what it can deliver on the 9 GPU system

Lord Mike, nice to hear from you again? But development would definitely proceed much faster, if you would join me?

Anyway, I will be extremely busy in the coming days. As a result, please do not expect to much for this week.

[LordMike]

What id like to see is support for more languages like german or french, czech maybe, characters, that arent used in a default english password.
also an option to specify the searchrange a bit more would be nice , say "atleast x numbers, max 2 non. english characters " and so on.

Sorry Keri, but this is still BETA. And I am still trying to optimize the algorithm. But a lot of people are downloading it. And therefore I will have a look!

lol. See what it can deliver on the 9 GPU system

Lord Mike, nice to hear from you again? But development would definitely proceed much faster, if you would join me?

Anyway, I will be extremely busy in the coming days. As a result, please do not expect to much for this week.

Sorry mate.
Been busy last week.. Working on a project I have to hand in on wednesday (10-20)..
On top of that, another project I'm handing in tomorrow.. (12-20 pages)

And then.. When that's done, I'll have to prepare for my annual 'large' project, which kinda defines how far I've gotten in school..
That project gets started on friday at 15:00. And then I have till friday 12:00 next week to finish it.

School just decided to screw with my time the next few weeks

EDIT:
My project beginning friday is about cryptography..
Looking into a little math, then IT as major subject demands me to make my own algorithm

[schwarzwaldhacker]

Sorry mate.
Been busy last week.. Working on a project I have to hand in on wednesday (10-20)..
On top of that, another project I'm handing in tomorrow.. (12-20 pages)

And then.. When that's done, I'll have to prepare for my annual 'large' project, which kinda defines how far I've gotten in school..
That project gets started on friday at 15:00. And then I have till friday 12:00 next week to finish it.

Good Luck! I used to be very bad at school. But within the last ten years, everything got really bad. So I really wish you good luck!

But contact me once you are finished. Because I have a lot of work for you as well.

[LordMike]

Sorry mate.
Been busy last week.. Working on a project I have to hand in on wednesday (10-20)..
On top of that, another project I'm handing in tomorrow.. (12-20 pages)

And then.. When that's done, I'll have to prepare for my annual 'large' project, which kinda defines how far I've gotten in school..
That project gets started on friday at 15:00. And then I have till friday 12:00 next week to finish it.

Good Luck! I used to be very bad at school. But within the last ten years, everything got really bad. So I really wish you good luck!

But contact me once you are finished. Because I have a lot of work for you as well.

lol.

[RudiRatlos]

Hmm i never had problems when running bars, igu, egb and so on but with your cuda des i get a error

G:\cluster\cuda\schwarz>schwarzwaldDES32BETA4.exe l=I_read_and_understood_the_li
cence s=0 e=5 c=loweralpha-numeric f=1.txt

schwarzwaldDES BETA 4 Win:v5.01
Written 2009 by Schwarzwaldhacker/ALGORITHM BY XSERG
CudaV1.1 GeForce 9800 GTX/9800 GTX+, Processors: 128 Clock rate: 1.674 Mg.
*************************************************
Do you agree NOT TO HARM directly or indirectly
other people with this tool?
(Press y for YES and then return/enter)
y

CharSet loweralpha-numeric = abcdefghijklmnopqrstuvwxyz0123456789
Salt DES(Unix) count 1
CUDA starts with 1 password(s)
cudaSafeCall() Runtime API error in file <c:/Users/der sohn/AppData/Local/NVIDIA
Corporation/NVIDIA CUDA SDK/projects/bandwidthTest/bandwidthTest.cu>, line 2323
: feature is not yet implemented.

the newest xzerg version

G:\cluster\cuda\des>MG_DES.exe f=1.txt s=1 e=5 b=128 d=0 c=loweralpha-numeric
DES(Unix) V2.02 Win:v5.01
1:CudaV1.01 GeForce 9800 GTX/9800 GTX+, Processors: 128 Clock rate: 1.674 Mg.
CharSet loweralpha-numeric = abcdefghijklmnopqrstuvwxyz0123456789
Salt DES(Unix) count 1
Start with 2 passwords 1-5
001 001 aehe9 Speed = 6526 k.pas/sec Time remaining: 0.00.00
Stop devices ...
Devices stoped...

[Rolf]

Maybe, just maybe, downloading the latest WHQL drivers(195.62) and installing latest CUDA toolkit(2.3) will help?
Dont forget to remove old ones using safe mode.

[schwarzwaldhacker]

Hmm i never had problems when running bars, igu, egb and so on but with your cuda des i get a error

Sorry RudiRatlos, I guess you don't have the newest drivers? Because I am for instance using Pinned Memory(Though, only a very small amount of 8 bytes). And therefore I think Rolf is right, you probably should update your drivers?

[schwarzwaldhacker]

I again released a new BETA!

http://s-schwarzwaldhacker.rbcmail.ru/s ... ldDES.html

Some of the Improvements:
* MORE SPEED!!!
* You can test up to 2048 passwords at once!

Actually I tried 2001 passwords loweralpha 1-4 in 3:18. While I got 8:06 for 2001 passwords loweralpha-numeric 1-4!

PLEASE NOTE: This update contains two files. The "schwarzwaldDES32BETA5-registers.exe" is recommended for newer devices, for instance for the GTX 295, the GTX 285, the GTX 280, the GTX 275 or the GTX 260. Since these devices can deal with up to 16384 registers(Please test it! Because I don't have one of these "newer devices" and I might be wrong about that )

[RudiRatlos]

yep was the driver

you mean Passwords or SALTS?

Because 1 SALT with 30 passwords xzerg was faster then yours

[schwarzwaldhacker]

RudiRatlos, thanks for testing!

Because 1 SALT with 30 passwords xzerg was faster then yours

RudiRatlos, unfortunately I don't have access to cuda right now. So I will have to check it tomorrow. But did you try the blocks option("b")? For instance "b=256" or even "b=512"?

I would also like to mention, that I found a bug! Because SchwarzwaldDES now needs the lists to be in the following format="Example:HaHsHxxxYYYxY". Therefore, if there are any characters after the hash, even ":", it could confuse everything. Sorry! I will try to correct this bug tomorrow!

[schwarzwaldhacker]

No, there won't be updates everyday . But there was something seriously wrong with the "schwarzwaldDES.rest"-file, where the unfound Hashes get stored. Besides, sometimes wrong logins were shown though the right passwords got found. I corrected these bugs!

http://s-schwarzwaldhacker.rbcmail.ru/s ... ldDES.html

Concerning speed, I had the following results on my GTX+9800:

1-Hash loweralpha 1-6 = 45 seconds ("b=512")
100-Hashes(same salt) loweralpha-numeric 1-5 = 17 seconds ("b=512")
101-Hashes(59 salts) loweralpha 1-5 = 1:40 minutes ("b=512")
2001-Hashes(>1400 salts) loweralpha-numeric 1-4 = 8:06 minutes ("b=32")

Correct me if I am wrong, but this is fast!

Please don't forget to help me with the "schwarzwaldDESBETA51-registers.exe". Which might be even faster on newer systems, like the GTX260, the GTX275 or even the GTX275. But I don't own one of these devices. Therefore you will have to test it for me.

[SecUpwN]

@schwarzwaldhacker: Thanks for developing this neat little tool! I´d love to see a version for Linux hitting the road though.
Und hey, kommst du wirklich aus dem Wald der berühmt-berüchtigten Schwarzwälder Kirschtorte?

-SecUpwN-

[schwarzwaldhacker]

Und hey, kommst du wirklich aus dem Wald der berühmt-berüchtigten Schwarzwälder Kirschtorte?

-$p!c3-

Das ist Irrtum. Ich bin echt Russisch Mann.

[RudiRatlos]

cheers bro
i`ll make a comparing test and post the reuslts here so we have a realy good overview.
But for now i can say very good

STATUS:i46999 Speed = 21845333 pas/sec Time: 2.18.59
All checked

Average Speed = 16915158 Checks per second!(Range/Seconds)
Average Speed per password = 268494 Checks per second!

SchwarzwaldDES found 62 Password(s), 1 Password(s) left

was 63 salts and 63hashes range 1-6 loweralpha-numeric with b=256

[schwarzwaldhacker]

was 63 salts and 63hashes range 1-6 loweralpha-numeric with b=256

Try the newest Release 5.1:

http://s-schwarzwaldhacker.rbcmail.ru/s ... ldDES.html

It should make you lough even more.

[schwarzwaldhacker]

I still don't have BITSLICE, but this Update has many nice features:

* schwarzwaldDES32BETA can now check 8-Character Passwords(Is this really so important? )
* schwarzwaldDES32BETA can now find the Null/Empty password!!!
* You can now test the Hashes against a list, like password.lst(Sorry, forgot to include it!). The password list can contain up to 16384 entries.
* You can now test the Hashes against their logins(But I found a bug, because it doesn't seem to find them all! However found passwords were always correct)
* If you test a list of more than 2048 Hashes, the last tested Hash is shown first(Good if you want to know where you will have to continue, especially if the list contains different types of Hashes)
* I reintroduced the "g"-option, which can help you to limit the number of Hashes you want to check. For instance "g=50", if you want to check 50 Hashes only

http://s-schwarzwaldhacker.rbcmail.ru/s ... ldDES.html

=> The "single-mode" and the "password-list-mode" are very slow. Quite possibly, I shouldn't use Cuda for it! However, if you have huge lists that you want to test for many hours, you should narrow down things first. And if you are too lazy to do it with “John The Ripper”, you still can do it with SchwarzwaldDES(I did it with 2048 Hashes. I found several hundreds with the "password-mode" and with the "login-mode". I then finished 1-5 loweralpha-numeric in like 2:30 hours).

[schwarzwaldhacker]

I am about to release the next BETA! I corrected the Password/Single-modes. And I also removed a number of smaller bugs.

But I am definitely very busy right now! Even though Russian Orthodox Christmas is still weeks away!

[schwarzwaldhacker]

I just released schwarzwaldDES32BETA7:

http://s-schwarzwaldhacker.rbcmail.ru/s ... ldDES.html

* Speed has only slightly been improved.
* But the "single-mode" and the "password-list-mode" are faster now. And I guess, BUGFREE! It is still better to use "John The Ripper" for wordlists. Of course if you are lazy , it is quite OK to try SchwarzwaldDES first. But please notice: The number of passwords checked by the "password-list-mode" == BLOCKS * 128.
* Other minor improvements, etc..

Theoretically, it should be possible to improve the "single-mode" and the "password-list-mode" dramatically. By directly writing into the global memory. But I am sincerely much too busy and I couldn't try it out.

Anyway, have fun!

[bazmauZze]

hi schwarzwaldhacker, thx for this great tool.. just one wish mm can you add attack by mask plz + ETA if you will have some time?
really wana use something like this in "passwords pro" - ?u?l?l?l?l?l?l?l
?u -alpha
?l - loweralpha

[schwarzwaldhacker]

bazmauZze, thanks for testing!

just one wish mm can you add attack by mask plz + ETA if you will have some time?
really wana use something like this in "passwords pro" - ?u?l?l?l?l?l?l?l
?u -alpha
?l - loweralpha

bazmauZze, I will have a look. But sincerely, unless somebody helps me out with Bitslice, "Multi-GPU" will now be the priority.

[schwarzwaldhacker]

I am getting problems, accessing my Website: http://s-schwarzwaldhacker.rbcmail.ru

Therefore, if you want to download the DES Cracker, please use that link:

http://narod.ru/disk/16285485000/schwar ... 7.zip.html

[schwarzwaldhacker]

This is really kind of a "Mini-Mini-Update"! Because I tried a lot of arithmetic tricks. But the improvement was far less than 1%.

However I found out, that my previous register Versions had not been compiled for "Compute Capability 1.3". I don't have one of these modern GPUs, like the GTX295, the GTX285, the GTX280, the GTX275, the GTX260 and so on. But my guess is that they should now be faster, because they might not need any CPU-"Local-Memory"("Compute Capability 1.3" activates more Registers. And "Local_memory", is uncached memory from the CPU, up to 100 times slower than registers).

http://s-schwarzwaldhacker.rbcmail.ru/s ... ldDES.html

Right now, I am really extremely busy. Besides, I often can't even access Cuda. Which makes it kind of difficult to develop new Crackers? But yet, I tried Multi-GPU. Unfortunately, the Linker didn't find an object or something. And I must admit that I don't understand everything I copied from others? Therefore I guess, it will be better if I give Svarychevski Michail Aleksandrovich the Source Code? As he knows best about Multi-GPU?

As I said, I am really very busy right now. And I even expect to become unable to release any further Crackers soon(This was the main reason, I released this corrected "Register" Version). Of course, I didn't forget about the improvements a lot of people requested! But if I have some time, I will now rather concentrate on FreeBSD Md5($1$, at least, an "experimental Version"). In order to create some kind of a "base". And then, everybody else might continue to develop it.