CUDA on Ubuntu 8.10

I posted this on the Ubuntu forums, but have yet to get a reply so thought I’d see if you guys could help.

I am trying to install the NVIDIA CUDA drivers on a fresh install of Intrepid. The version I am using is NVIDIA-Linux-x86_64-177.73-pkg2.run. It detects that I am using a xen kernel, which from some googling seems related to the xen virtual server. This halts the installer as CUDA apparently doesn’t support the xen kernel. I am, however, totally unaware of running such a kernel (I have never used xen anyway). The kernel I am running is 2.6.27-7-generic .

Logs are below. Any bright ideas? Thanks in advance!

/var/log/nvidia-installer.log :

nvidia-installer log file '/var/log/nvidia-installer.log'

creation time: Sat Nov  8 20:11:47 2008

installer version: 1.0.7

option status:

  license pre-accepted	: false

  update				  : false

  force update			: false

  expert				  : false

  uninstall			   : false

  driver info			 : false

  precompiled interfaces  : true

  no ncurses color		: false

  query latest version	: false

  OpenGL header files	 : true

  no questions			: false

  silent				  : false

  no recursion			: false

  no backup			   : false

  kernel module only	  : false

  sanity				  : false

  add this kernel		 : false

  no runlevel check	   : false

  no network			  : false

  no ABI note			 : false

  no RPMs				 : false

  no kernel module		: false

  force SELinux		   : default

  no X server check	   : false

  no cc version check	 : false

  force tls			   : (not specified)

  force compat32 tls	  : (not specified)

  X install prefix		: (not specified)

  X library install path  : (not specified)

  X module install path   : (not specified)

  OpenGL install prefix   : (not specified)

  OpenGL install libdir   : (not specified)

  compat32 install chroot : (not specified)

  compat32 install prefix : (not specified)

  compat32 install libdir : (not specified)

  utility install prefix  : (not specified)

  utility install libdir  : (not specified)

  doc install prefix	  : (not specified)

  kernel name			 : (not specified)

  kernel include path	 : (not specified)

  kernel source path	  : (not specified)

  kernel output path	  : (not specified)

  kernel install path	 : (not specified)

  proc mount point		: /proc

  ui					  : (not specified)

  tmpdir				  : /tmp

  ftp mirror			  : ftp://download.nvidia.com

  RPM file list		   : (not specified)

Using: nvidia-installer ncurses user interface

-> License accepted.

-> Installing NVIDIA driver version 177.73.

-> No precompiled kernel interface was found to match your kernel; would you li

   ke the installer to attempt to download a kernel interface for your kernel f

   rom the NVIDIA ftp site (ftp://download.nvidia.com)? (Answer: Yes)

-> No matching precompiled kernel interface was found on the NVIDIA ftp site;

   this means that the installer will need to compile a kernel interface for

   your kernel.

-> Performing CC sanity check with CC="cc".

-> Performing CC version check with CC="cc".

-> Kernel source path: '/lib/modules/2.6.27-7-generic/build'

-> Kernel output path: '/lib/modules/2.6.27-7-generic/build'

-> Performing rivafb check.

-> Performing nvidiafb check.

-> Performing Xen check.

ERROR: The kernel you are installing for is a Xen kernel!

	   

	   The NVIDIA driver does not currently work on Xen kernels. If 

	   you are using a stock distribution kernel, please install 

	   a variant of this kernel without Xen support; if this is a 

	   custom kernel, please install a standard Linux kernel.  Then 

	   try installing the NVIDIA kernel module again.

ERROR: Installation has failed.  Please see the file

	   '/var/log/nvidia-installer.log' for details.  You may find suggestions

	   on fixing installation problems in the README available on the Linux

	   driver download page at www.nvidia.com.

CUDA is not supported on Ubuntu-8.xx at this time.

CUDA isn’t officially supported, but that doesn’t mean it doesn’t work, as the thread I opened revealed.
Alas, this doesn’t seem to be an issue with CUDA, but rather an issue with trying to install the drivers, from
NVidia’s site, on Ubuntu, where the running kernel is said to have XEN support enabled. This doesn’t seem
to be a new issue with ubuntu either.

It seems that there was a known issue with server kernels in ubuntu which for some reason were compiled
with XEN support. Now you’re saying the kernel you’re running is the generic 64 bit kernel, which would
suggest that the problem had propagated into the generic kernels. I suggest you open a bug with Ubuntu
on this issue.

I myself am running a generic kernel on a 32-bit 8.10. I didn’t try to install the driver from NVidia, instead
I just installed 177.80 using synaptic, (I think it’s on the partners repository). CUDA’s SDK was compiled
once I installed gcc 4.1, and the GLUT development libraries. All SDK applications work fine for me. You
can try installing the driver like I did, and post your conclusions. However, I strongly recommend reporting
the bug on the generic kernel with Ubuntu.

I stuck with the NVIDIA drivers installed via the apt, changed to using gcc 4.1 and it now works. Thanks for the pointers!

Someone else just reported the bug to ubuntu regarding xen being enabled by default (https://bugs.launchpad.net/ubuntu/+source/linux-meta/+bug/264763).

CUDA is not supported on Ubuntu-8.xx at this time?¿¿?¿?¿?
I’ve ubuntu 8-04 and CUDA without any problem! so where is the problem??

9800GT. ubuntu 8.04 32. the driver installed is 177.80.

“Not supported” isn’t the same as “doesn’t work.” NVIDIA hasn’t tested a the current CUDA release on Ubuntu 8.x, but if it works for you, that’s great. The next CUDA release will have been tested on Ubuntu 8.04, so there will be some guarantee it will work.

OK! thank you seibert!

Works fine if you run the .run file with --update or nvidia-installer -f it will download the 177.80 driver and work with the default kernel on ibex

I’m on intrepid 8.10, and the 2.1 driver compiled cleanly. The problem comes when I build my code, it says:

nvcc   -m64 -shared --compiler-options -fPIC --compiler-options -Wall --compiler-options -fno-strict-aliasing  -I. -I/home/mnitsche/cuda_libs/include -I/home/mnitsche/cuda_libs/../sdk/common/inc -DUNIX -O3 --linker-options -fPIC  -o cuda/exchnum.cu_o -c cuda/exchnum.cu

/usr/include/bits/wchar2.h(290): error: identifier "__builtin_va_arg_pack" is undefined

/usr/include/bits/wchar2.h(291): error: identifier "__builtin_va_arg_pack" is undefined

The problem is that /usr/include/bits comes with the kernel, so I think this means I have no other choice than to go back to 8.04. Is that so?