FATAL ERROR: CUDA error in cudaGetDeviceCount on Pe 0 (thomasASUS): CUDA driver version is insufficient for CUDA runtime version

From: Thomas Evangelidis (tevang3_at_gmail.com)
Date: Sat Sep 22 2012 - 07:41:25 CDT

Greetings,

I have an ASUS N56V laptop with a GeForce GT 650 M, running Fedora 17
64-bit. The proprietary NVIDIA drivers failed to install so I followed the
instructions from the following page to set up nouveau and bumblebee in
order to install CUDA 4.2.

http://duxyng.wordpress.com/2012/01/26/finally-working-nvidia-optimus-on-fedora-16/

The CUDA tests run successfully. However, when I try to run the ApoA1
benchmark within the latest NAMD nightly build directory, I get the
following error:

$ export LD_LIBRARY_PATH=`pwd` && ./namd2 +idlepoll apoa1.namd
Charm++: standalone mode (not using charmrun)
Converse/Charm++ Commit ID: v6.4.0-beta1-0-g5776d21
CharmLB> Load balancer assumes all CPUs are same.
Charm++> Running on 1 unique compute nodes (8-way SMP).
Charm++> cpu topology info is gathered in 0.000 seconds.
Info: NAMD CVS-2012-09-22 for Linux-x86_64-multicore-CUDA
Info:
Info: Please visit http://www.ks.uiuc.edu/Research/namd/
Info: for updates, documentation, and support information.
Info:
Info: Please cite Phillips et al., J. Comp. Chem. 26:1781-1802 (2005)
Info: in all publications reporting results obtained with NAMD.
Info:
Info: Based on Charm++/Converse 60400 for multicore-linux64-iccstatic
Info: Built Sat Sep 22 02:25:04 CDT 2012 by jim on lisboa.ks.uiuc.edu
Info: 1 NAMD CVS-2012-09-22 Linux-x86_64-multicore-CUDA 1 thomasASUS
thomas
Info: Running on 1 processors, 1 nodes, 1 physical nodes.
Info: CPU topology information available.
Info: Charm++/Converse parallel runtime startup completed at 0.00364304 s
FATAL ERROR: CUDA error in cudaGetDeviceCount on Pe 0 (thomasASUS): CUDA
driver version is insufficient for CUDA runtime version
------------- Processor 0 Exiting: Called CmiAbort ------------
Reason: FATAL ERROR: CUDA error in cudaGetDeviceCount on Pe 0 (thomasASUS):
CUDA driver version is insufficient for CUDA runtime version

[0] Stack Traceback:
  [0:0] CmiAbort+0x95 [0xbfcb85]
  [0:1] _Z8NAMD_diePKc+0x62 [0x5aed32]
  [0:2] _Z13cuda_errcheckPKc+0x14a [0x797538]
  [0:3] _Z15cuda_initializev+0x194 [0x796d3a]
  [0:4] _Z8all_initiPPc+0x55 [0x5b80a5]
  [0:5] _Z11master_initiPPc+0x67 [0x5b7cd7]
  [0:6] _ZN7BackEnd4initEiPPc+0xe8 [0x5b7c50]
  [0:7] main+0x2f [0x5b2e57]
  [0:8] __libc_start_main+0xf5 [0x3f20221735]
  [0:9] _ZNSt8ios_base4InitD1Ev+0x6a [0x548a8a]
Charm++ fatal error:
FATAL ERROR: CUDA error in cudaGetDeviceCount on Pe 0 (thomasASUS): CUDA
driver version is insufficient for CUDA runtime version

[0] Stack Traceback:
  [0:0] [0xbfdab5]
  [0:1] CmiAbort+0xd3 [0xbfcbc3]
  [0:2] _Z8NAMD_diePKc+0x62 [0x5aed32]
  [0:3] _Z13cuda_errcheckPKc+0x14a [0x797538]
  [0:4] _Z15cuda_initializev+0x194 [0x796d3a]
  [0:5] _Z8all_initiPPc+0x55 [0x5b80a5]
  [0:6] _Z11master_initiPPc+0x67 [0x5b7cd7]
  [0:7] _ZN7BackEnd4initEiPPc+0xe8 [0x5b7c50]
  [0:8] main+0x2f [0x5b2e57]
  [0:9] __libc_start_main+0xf5 [0x3f20221735]
  [0:10] _ZNSt8ios_base4InitD1Ev+0x6a [0x548a8a]
Aborted (core dumped)

Is it because I don't have the proprietary NVIDIA drivers installed, or
because I have Cuda 4.2 instead of 4.0? Then why can I run successfully
other programs with the GPU, including GROMACS?

Thanks in advance.
Thomas

-- 
======================================================================
Thomas Evangelidis
PhD student
University of Athens
Faculty of Pharmacy
Department of Pharmaceutical Chemistry
Panepistimioupoli-Zografou
157 71 Athens
GREECE
email: tevang_at_pharm.uoa.gr
          tevang3_at_gmail.com
website: https://sites.google.com/site/thomasevangelidishomepage/

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:22:06 CST