What steps will reproduce the problem? 1. Clean install of pyrit with cuda 2. Run pyrit list_cores 3.

What is the expected output? What do you see instead? I expect the cores (mine is GeForce GT 330M)

what I see instead is: Pyrit 0.4.1-dev (svn r308) (C) 2008-2011 Lukas Lueg http://pyrit.googlecode.com This code is distributed under the GNU General Public License v3+

Traceback (most recent call last): File "/usr/local/bin/pyrit", line 6, in <module> pyrit_cli.Pyrit_CLI().initFromArgv() File "/Library/Python/2.7/site-packages/pyrit_cli.py", line 117, in initFromArgv func(self, **options) File "/Library/Python/2.7/site-packages/pyrit_cli.py", line 293, in list_cores with cpyrit.cpyrit.CPyrit() as cp: File "/Library/Python/2.7/site-packages/cpyrit/cpyrit.py", line 440, in init self.cores.append(CUDACore(queue=self, dev_idx=dev_idx)) File "/Library/Python/2.7/site-packages/cpyrit/cpyrit.py", line 245, in init _cpyrit_cuda.CUDADevice.init(self, dev_idx) SystemError: CUDA_ERROR_NO_BINARY_FOR_GPU

What version of the product are you using? On what operating system?

I'm using the latest svn, on OSX 10.9.4. Pyrit is working fine, cpyrit installs without issues after modify setup.py to identify the path for nvcc. But it doesn't run properly as I reported above. CUDA drivers and toolkit are 6.5.14.

Please provide any additional information below.

I've followed the instructions to test CUDA: it is working properly on my system (I've tested all of the samples that come with the toolkit). I checked the previous posts for similar issues. Some of them seem similar but the solutions do not fit my specific case. I've checked google and the various tutorials and guides. Probably I'm missing something stupid. But it is missing. Annoying.

Same issue here.

I also have this problem

Ran into the same problem with no avail. All CUDA samples run fine, but receive this error from pyrit's list_cores and selftest.

I am using Linux Mint 17 (Ubuntu 14.04 base) and complied both pyrit and cpyrit from source, running 0.4.0

Comment deleted

I had the same problem and fixed it by modifying setup.py and specifying the correct gpu architecture for my graphics card. I had a GeForce GT 330 which is compatible with compute 1.2, so in the call to nvcc you have to add an option to specify gpu architecture:

nvcc_cmd = NVCC + bit_flag + ' --host-compilation C'\ ' --gpu-architecture compute_12'\ ' --verbose'\ ' -Xcompiler "-fPIC" --ptx' \ ' ./_cpyrit_cudakernel.cu'

otherwise it specifies compute_20 which is not supported by the card.

unfortunately this old card doesn't provide too much better speed, but at least it works

i had the same problem. Geforce GTX 660 OC, AMD FX 6100 Sixcore and 16GB RAM DDR3.

The Solution by "#5" is not working. My GPU is 3.0

Not working whith: ' --gpu-architecture compute_12'\ ' --gpu-architecture compute_20'\ ' --gpu-architecture compute_30'\

nvida driver and cuda toolkit working. no one has a solution? There are other programs that crack with GPU?