No NVidia compatible platform found

[joshi982]

nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2012 NVIDIA Corporation
Built on Fri_Sep_21_17:28:58_PDT_2012
Cuda compilation tools, release 5.0, V0.2.1221

dmesg | grep -i nvidia
nvidia: module license 'NVIDIA' taints kernel.
NVRM: loading NVIDIA UNIX x86_64 Kernel Module 310.19 Thu Nov 8 00:52:03 PST 2012

./cudaExample.sh
ERROR: No NVidia compatible platform found

I have two Tesla M2070.
what is my problem?
Iam going crazy!!

[atom]

Maybe you are on 32 bit? The example uses 64 bit. If you are on 32 bit you have to edit the .sh File

[joshi982]

its x86_64 and i receive the error from both 32 & 64 bit cudahashcat, Is this necessary to execute cudahashcat with root access?

[atom]

usually not, but you should try to make that clear.

[joshi982]

How the hashcat-lite find the Graphic-Card? I think its use the 'lspci' command in linux to find them.
Is this true?
and the 'lspci' command don't works on my SSH, for ex:

lspci | grep VGA
-bash: lspci: command not found

but its works:

/sbin/lspci
00:00.0 Host bridge: Intel Corporation 440FX - 82441FX PMC [Natoma] (rev 02)
00:01.0 ISA bridge: Intel Corporation 82371SB PIIX3 ISA [Natoma/Triton II]
00:01.1 IDE interface: Intel Corporation 82371SB PIIX3 IDE [Natoma/Triton II]
00:01.3 Bridge: Intel Corporation 82371AB/EB/MB PIIX4 ACPI (rev 01)
00:02.0 VGA compatible controller: Cirrus Logic GD 5446
00:03.0 3D controller: nVidia Corporation GF100 [Tesla S2050] (rev a3)
00:04.0 3D controller: nVidia Corporation GF100 [Tesla S2050] (rev a3)
00:05.0 Unassigned class [ff80]: XenSource, Inc. Xen Platform Device (rev 01)

i think it's my problem

[philsmd]

I don't think that lspci will be involved in any way, but you might check if libcuda.so and libnvidia*.so* are available.
Also checking nvidia-settings on linux systems could be useful.

See also:
modprobe nvidia
cat /proc/driver/nvidia/version
ls /dev/nvidia*

Just guesses, I' m no expert.
BTW: the lspci command fails just because the PATH environment variable is NOT set correctly (to include also the sbin directory, excluding it could be a valid choice sometimes :-) )

I think your nvidia module is not loaded or failed somehow to load... please also re-check the dmesg, maybe there is no "Nvidia" identifier on the particular line.
Best way is to check nvidia-settings !?

EDIT: please also check the nvidia-smi command

[joshi982]

its very very complex!!!

cat /proc/driver/nvidia/version
NVRM version: NVIDIA UNIX x86_64 Kernel Module 310.19 Thu Nov 8 00:52:03 PST 2012
GCC version: gcc version 4.1.2 20080704 (Red Hat 4.1.2-50)

ls /dev/nvidia*
ls: /dev/nvidia*: No such file or directory

and stranger!!

nvidia-smi
NVIDIA: could not open the device file /dev/nvidiactl (No such file or directory).
NVIDIA-SMI has failed because it couldn't communicate with NVIDIA driver. Make sure that latest NVIDIA driver is installed and running.

[philsmd]

The nvidia modules are not loaded correctly. They will create (automatically) the /dev/nvidia* devices, that are needed (by the libcuda and/or libnvidia, both? I am no expert).
For sure the module is NOT loaded correctly, or you use other drivers? are you using the proprietary drivers shipped by nvidia (or default once provided by your distro)?

See:
jockey-text --list
or if you prefer
jockey-gtk
EDIT: KDE uses: jockey-kde, you may need to run it as root (for sure) and if you are not using kde or the konsole version try to tab to find the jockey shipped with your distro! Hopefully there is one.
... maybe you are not using the drivers shipped by nvidia.
AND: as @atom told me recently cuda sdk is not needed so nvcc is optional I think, but it doesn't hurt.

Before hashcat can be using your GPU nvidia-smi, nvidia-settings etc should ALL work...
hashcat interfaces with the libraries provided by nvidia (for your GPU) and does NOT access the /dev/ directly. Please correct me iff that is wrong!

[BlackAlbatross]

I'm uberly noob on this forum(first post), but whenever I want to check what modules are loaded(especially VGA), I use command ''lspci -kk''.
This way I can see what modules are available(kernel modules), what are loaded into kernel(kernel driver in use) and sometimes a hint what is conflicting(for example - nvidia & nouveau).
Edit: nvidiafb module doesn't work(if you have that available). Or at least it didn't for me.

[epixoip]

neither dmesg nor lspci are a common or proper way to check if a module is loaded. normally one would do ''lsmod | grep nvidia'' or ''grep nvidia /proc/modules''