VMD-L Mailing List

From: John Stone (johns_at_ks.uiuc.edu)
Date: Mon Apr 21 2003 - 11:31:59 CDT

Hi David,
  As far as I know there are no memory leaks in VMD 1.8 that would
cause the behavior you are observing. Its quite possible that this
is a bug in the Radeon 9700 driver on Windows XP. A couple of questions
for you that might help narrow this down, so that we can find the source
of the problem:
  1) What molecular representations are you drawing when you observe
     the memory leak occuring?

  2) If you load VMD with no molecules drawn, just displaying the title
     screen, do you observe the same behavior?

  3) Can you see a difference between the rate of leaking if you draw a
     molecule in "VDW", "Surf", or "Bonds". Those three reps consist of
     very different OpenGL commands. If one leaks and the others don't,
     then we can send a bug report to ATI and tell them what is causing
     the problem.

If you can provide more info we can find the bug in VMD (if any) or report
the problem to ATI if its a driver problem. So far you are the first to
experience this problem so that's why I suspect its an issue with the
Radeon driver.

Thanks,
  John Stone
  vmd_at_ks.uiuc.edu

> Date: Sat, 19 Apr 2003 15:21:31 -0500
> From: David Henthorn <henthorn_at_mail.utexas.edu>
> To: vmd-l_at_ks.uiuc.edu
> Subject: Memory problems with WinXP
>
> Memory problems with VMD, Win XP, and Radeon 9700?
>
> I apologize if this has already been discussed here and I missed the
> relevant info....
>
> I'm running VMD 1.8 on a P4 with WinXP SP1 and a Radeon 9700. If I let
> VMD run for some time, then memory usage steadily increases until XP
> kills VMD because it's referenced memory out of bounds.
>
>
> "Unhandled exception at 0x69244caa in winvmd.exe: 0xC0000005: Access
> violation writing location 0x00000000."
>
> When I started watching this with the Task Manager, VMD increases in
> physical memory up to about 300MB and 1.5 GB of virtual memory before XP
> kills it.
>
> This happens for pretty simple PDB files... one that seems repeatable is
> to load 1A06 from the PDB and then change the graphics to VDW. The
> memory increase only happens when the molecule is rotating, and then it
> seems that the faster it rotates, the faster it eats memory.
>
> When it crashes MS Visual Studio Debugger fires up and shows the
> following info:
>
> Program: winvmd.exe
> Thread: Win32 Thread
> Stack Frame: atioglxx.dll!69244caa()
>
> So, am I correct to assume that it's not really VMD that's errant here,
> rather the ATI OpenGL driver that is being linked to by VMD?
>
> ATI Catalyst Drivers 3.2 Version 6.14.01.6307.
>
> Any help would be appreciated.
>
> Thanks,
>
> Dave
>
>
>
>
>
>
> 

-- 
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