Cryptohaze.com

Bitweasil wrote:Hm.

Add:

#include <stdint.h>

at the top of CHCommon.cpp

I'll take a look and try to fix that in the source.

where can I find this file "CHCommon.cpp" ?

In the directory you're building from... it's one of the source files.

I'll try to fix it in SVN this week.

user@computername:~/NVIDIA_GPU_Computing_SDK/C/src/Cryptohaze-Multiforcer$ make -f Makefile-Linux
g++: /usr/local/lib/libboost_thread.so: Datei oder Verzeichnis nicht gefunden
g++: /usr/local/lib/libboost_system.so: Datei oder Verzeichnis nicht gefunden
make: *** [../../bin/linux/release/CUDA-Multiforcer-CPP] Fehler 1
user@computername:~/NVIDIA_GPU_Computing_SDK/C/src/Cryptohaze-Multiforcer$

I think it's the last problem

ooh. That's a boost version mismatch.

Easiest way to attempt to fix it:

Make sure you have boost development packages on your system.

Remove the "obj" directory.

Rebuild.

The issue here is that the boost header files included with the SVN aren't actually the ones being used on your platform, so it gets linking errors. The easiest way to fix this is to remove the boost header files from the multiforcer directory and let it use the platform files.

I installed the following package: http://packages.ubuntu.com/maverick/amd ... v/download

Delete the obj folder from the Multiforcer but have the same problem:

......
CHNetworkClient.cpp:177: warning: ‘localWorkunit’ may be used uninitialized in this function
g++: /usr/local/lib/libboost_thread.so: Datei oder Verzeichnis nicht gefunden
g++: /usr/local/lib/libboost_system.so: Datei oder Verzeichnis nicht gefunden
make: *** [../../bin/linux/release/CUDA-Multiforcer-CPP] Fehler 1
*:~/NVIDIA_GPU_Computing_SDK/C/src/Cryptohaze-Multiforcer$

Can you make a build or help me with the problem ?

EDIT: solved the problem. I forgot to fix the libboost links (usr/lib -> usr/local/lib)

RIG wrote:EDIT: solved the problem. I forgot to fix the libboost links (usr/lib -> usr/local/lib)

That would do it! I need to fix the Makefiles.

So, does it run without errors and find results?

It runs and find Plaintexts

But always find one error:

"CUDA error 5: setting the device when a process is active is not allowed"

But I'm not sure if this error has an effect on the cracking progress... because the bruteforcer is still running

The server and client mode has one bug. Sometimes when I crack hashes from different password lengths the server start the next higher length but the client is still running the finished length.
For example:
loweralpha-numeric length 6 =100% => server is going to crack loweralpha-numeric length 7
=>client still shows on the pw length number 6 and not 7

But still a very good bruteforcer. Good job Bitweasel

I think the password length thing may be a display bug and not an actual bug (it should be running the longer kernels). I will verify and fix.

RIG wrote:But always find one error:

"CUDA error 5: setting the device when a process is active is not allowed"

The cuda toolkit 3.2 installation fixed the issue

RIG wrote:The cuda toolkit 3.2 installation fixed the issue

Awesome. Hopefully it fixes the RT issue as well.

Good to know I have some possible 4.0 compatibility issues. I've tried to target the most modern official toolkit, not the beta stuff.