VMD-L Mailing List

From: John Stone (johns_at_ks.uiuc.edu)
Date: Tue Mar 29 2005 - 12:52:55 CST

Sabuj,
  Yeah, it could be that Chimera and O don't use as large of a variety of
OpenGL primitives and VMD and Pymol do. Last I checked, Rasmol didn't use
OpenGL at all, it renders using its own algorithm and displays using Xlib
calls. Do you have any other OpenGL programs to try, I suspect that VMD
and Pymol are not the only programs that'll stress the drivers in this way.

  John Stone
  vmd_at_ks.uiuc.edu

On Tue, Mar 29, 2005 at 11:55:03AM -0600, Sabuj Pattanayek wrote:
> Hi,
>
> Unfortunately I can't get pymol-0.97 & 0.98beta29/27 to open a pdb
> without crashing instantly. I also tested Chimera (1.2105), O v9.0.7,
> and rasmol and these seem to work ok but I guess they aren't using
> quads/quad strips. Thanks for your insight, I'll email nvidia with the
> contents of your email.
>
> ..Sabuj Pattanayek
>
> John Stone wrote:
> >Sabuj,
> > That's really interesting. It sounds to me like NVidia's new driver
> >has some sort of problem with drawing quads or quad strips in immediate
> >mode.
> >If that's what's going on, then a lot of other programs are going to crash
> >too, not just VMD and Pymol, so I suspect they'll have to fix this very
> >soon :-) All of the VMD reps you listed as having problems draw
> >quadrilateral strips at some point, so that's my personal guess as
> >to what is happening with their new driver. The Isosurface rep is
> >one exception, it is drawing triangle strips, so I'm not sure why you'd
> >have trouble with it, given that several of the others you listed do
> >as well. Please email these results to linux-bugs_at_nvidia.com so they
> >are aware of the problem. If you can do similar tests with Pymol,
> >that will help convince them it's a new problem with the driver and not
> >just one bad software package, which is likely their default assumption.
> >
> > John
> >
> >On Tue, Mar 29, 2005 at 11:11:54AM -0600, Sabuj Pattanayek wrote:
> >
> >>Hi John,
> >>
> >>This is what I was able to test on vmd 1.8.3, I think 1.8.2 will do the
> >>same (note this only affects the quadro cards and not geforce):
> >>
> >>Representation Result
> >>-------------- ------
> >>Lines (Default) ok
> >>Bonds crash
> >>DynamicBonds crash
> >>HBonds ok
> >>Points ok
> >>VDW ok
> >>CPK crash
> >>Licorice crash
> >>Trace crash
> >>Tube crash
> >>Ribbons ok
> >>NewRibbons ok
> >>Cartoon crash
> >>NewCartoon ok!!!
> >>MSMS ok
> >>Surf ok
> >>Dotted ok
> >>Solvent ok
> >>Isosurface(points) ok
> >>Isosurface(solid) crash
> >>Isosurface(wireframe) ok
> >>VolumeSlice ok? ..not sure what datatype this wants (never used
> >>this before), i was using a brix map but couldn't see anything
> >>
> >>Thanks,
> >>Sabuj Pattanayek
> >>
> >>David A. Horita wrote:
> >>
> >>>Hi,
> >>>I Haven't tried VMD yet, but Windows XP-pro pymol crashes upon file open
> >>>under the 71.84 drivers, but not the 70.78 drivers. (Quadro 1300,
> >>>XP-Pro servicepack 2)
> >>>David Horita
> >>>
> >>>-----Original Message-----
> >>>From: pymol-users-admin_at_lists.sourceforge.net
> >>>[mailto:pymol-users-admin_at_lists.sourceforge.net] On Behalf Of John Stone
> >>>Sent: Monday, March 28, 2005 10:10 AM
> >>>To: Sabuj Pattanayek
> >>>Cc: pymol-users_at_lists.sourceforge.net; vmd-l_at_ks.uiuc.edu
> >>>Subject: Re: vmd-l: vmd and Re: [PyMOL] pymol crashes with new [7167]
> >>>nvidia driver
> >>>
> >>>
> >>>Hi,
> >>>Have you reported the crashes to NVidia? I'd suggest sending an email
> >>>to their linux-bugs_at_nvidia.com mail address. The fact that multiple
> >>>packages crash with the new driver indicates that they may have a new
> >>>bug in their drivers. You mentioned that making a licorice rep crashes
> >>>your VMD on that driver, what about VDW or bonds? Do they have a
> >>>similar effect?
> >>>
> >>>John Stone
> >>>vmd_at_ks.uiuc.edu
> >>>
> >>>On Sat, Mar 26, 2005 at 04:42:42PM -0600, Sabuj Pattanayek wrote:
> >>>
> >>>
> >>>>Yeah I have the same problem in the beta pymol, haven't tried the
> >>>>stable
> >>>>version:
> >>>>
> >>>>OpenGL-based graphics engine:
> >>>>GL_VENDOR: NVIDIA Corporation
> >>>>GL_RENDERER: Quadro4 750 XGL/AGP/SSE/3DNOW!
> >>>>GL_VERSION: 1.5.3 NVIDIA 71.67
> >>>>Hardware stereo capability detected.
> >>>>Adapting to Quadro hardware...
> >>>>Detected 2 CPUs. Enabled multithreaded rendering.
> >>>>/usr/bin/pymol: line 14: 8852 Segmentation fault
> >>>>$PYMOL_PATH/pymol.exe $*
> >>>>
> >>>>and 7167 won't compile on 2.6.10-11? kernels either, need 2.6.12-rc1,
> >>>>haven't tested others. I get similar crashes in VMD 1.8.2+ using 7167
> >>>>also. For instance, if you load a pdb and try to make a "licorice"
> >>>>representation, vmd just closes/crashes.
> >>>>
> >>>>I also think this maybe a Quadro problem and not a a problem with
> >>>>GeForce cards. I have a GeForce 5500 FX on another machine and it
> >>>>doesn't have these programs, and realtime GLSL rendering works very
> >>>>nicely in vmd 1.8.3.
> >>>>
> >>>>Sorry for posting this on both mailing lists, just trying to avoid
> >>>>reposting.
> >>>>
> >>>>..Sabuj Pattanayek
> >>>>
> >>>>HuiZhe Li wrote:
> >>>>
> >>>>
> >>>>>just upgraded nvidia driver from 6629 to 7167 (Linux, RHEL 4) and
> >>>>>pymol crashes when a molecule is load. is there a fix? (pymol worked
> >>>>>fine with 6629 driver).
> >>>>>
> >>>>>OpenGL-based graphics engine:
> >>>>>GL_VENDOR: NVIDIA Corporation
> >>>>>GL_RENDERER: Quadro FX 1400/PCI/SSE2
> >>>>>GL_VERSION: 1.5.3 NVIDIA 71.67
> >>>>>Adapting to Quadro hardware...
> >>>>>/usr/local/bin/pymol: line 14: 4575 Segmentation
> >>>>>fault (core dumped) $PYMOL_PATH/pymol.exe $*
> >>>>>
> >
> > 

-- 
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