From: John Stone (johns_at_ks.uiuc.edu)
Date: Thu Mar 24 2011 - 21:02:06 CDT

Jerome,
  Are you using the GPU for anything besides VMD, like Google Chrome?
I've had a few people telling me lately that some other apps have started
interfering with VMD if the GPU doesn't have much RAM. I've reported
a driver bug to NVIDIA about this already, and they are working on it so
that at least the error codes returned are more indicative of the real
problem. How long is the first VMD running before you start the second
one?

Cheers,
  John

On Wed, Mar 23, 2011 at 03:40:20PM +0100, Jérôme Hénin wrote:
> Hi all,
>
> I am running VMD 1.9 on a Linux x86_64 machine with an nVidia GPU and
> the assorted proprietary driver. The first instance says:
>
> Info) VMD for LINUXAMD64, version 1.9 (March 14, 2011)
> (...)
> Info) Detected 1 available CUDA accelerator:
> Info) [0] Quadro FX 3700 14 SM_1.1 @ 1.25 GHz, 511MB RAM, KTO, OIO, ZCP
> Info) OpenGL renderer: Quadro FX 3700/PCI/SSE2
> Info) Features: STENCIL MSAA(16) MDE CVA MTX NPOT PP PS GLSL(OVF)
> Info) Full GLSL rendering mode is available.
> Info) Textures: 2-D (8192x8192), 3-D (2048x2048x2048), Multitexture (4)
>
> If I try to start a second instance of VMD though, it says:
>
> Info) Creating CUDA device pool and initializing hardware...
> Warning) thread[0] can't init GPU[0] found by device query
> Warning) CUDA error: no CUDA-capable device is available
> Segmentation fault
>
> I don't remember this happening with previous versions. Any ideas?
>
> Cheers,
> Jerome

-- 
NIH Resource for Macromolecular Modeling and Bioinformatics
Beckman Institute for Advanced Science and Technology
University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
Email: johns_at_ks.uiuc.edu                 Phone: 217-244-3349
  WWW: http://www.ks.uiuc.edu/~johns/      Fax: 217-244-6078