VMD-L Mailing List

From: Oliver Beckstein (orbeckst_at_jhmi.edu)
Date: Mon Jan 21 2008 - 10:42:05 CST

John,

> I was checking on the crash symptoms you describe, and these
> sound like the known problem with older Linux kernels and
> multithreading
> on Linux kernels earlier than 2.6.9-42. As you are apparently running
> 2.6.9-22, that would explain the crash problems you've described, as
> each of the crashes you mention are in pieces of code that are
> heavily multithreaded.

Ok, makes sense. Unfortunately I can't really update the kernel
easily on this machine.
(Btw, with 'VMD for LINUX, version 1.8.5 (August 25, 2006)' these
bugs do not occur.)

> The MacOS X oddity you mention sounds like a [bug] but in their X
> server.

I upgraded to Mac OS X 10.4.11 and their latest X11 and at least the
resizing crash did not happen again.

Could you possibly make an OSX/ppc build of
vmd-1.8.7a14extrabonds.bin available as well? This would allow me to
do all my work through the local display.

Many thanks,
Oliver

> Let me know if you have more questions about these
> issues.
>
> Cheers,
> John Stone
> vmd_at_ks.uiuc.edu
>
> On Thu, Dec 20, 2007 at 04:50:10PM -0500, Oliver Beckstein wrote:
>> John,
>>
>>> I've posted two Linux builds with 64 bonds per atom in the alpha
>>> directory in the VMD Public project on BioCoRE:
>>> http://www.ks.uiuc.edu/Research/vmd/alpha/
>>
>> thanks for the Linux builds. I only tested the 32bit one
>> (vmd-1.8.7a14extrabonds.bin.LINUX.opengl). It works in principle. In
>> particular, I get all my 'bonds' displayed.
>>
>> I'm having a few stability issues. Although they won't prevent me
>> from
>> using the build I'll mention them nevertheless.
>>
>> For testing, I ran Linux-VMD remotely (bad, I know) through X11 on my
>> PowerBook laptop:
>>
>> VMD: 1.8.7a14extrabonds (December 19, 2007)
>>
>> remote:
>> Linux: CentOs 4.2 (Rocks 4.1)
>> Linux 2.6.9-22.ELsmp #1 SMP Sat Oct 8 19:11:43 CDT 2005 i686 i686
>> i386
>> GNU/Linux
>>
>> local:
>> Mac OS X 10.3.9, Apple's X11
>>
>> VMD's startup output:
>> Info) ATI Linux driver detected, limiting features to avoid driver
>> bugs.
>> Info) Set the environment variable VMDDISABLEATILINUXWORKAROUND
>> Info) to enable full functionality on a known-safe driver version.
>> Info) OpenGL renderer: ATI Radeon 9600 XT OpenGL Engine
>> Info) Features: STENCIL MDE MTX
>> Info) GLSL rendering mode is NOT available.
>> Info) Textures: 2-D (2048x2048), Multitexture (8)
>>
>>
>> The following operations crashed VMD:
>> * Render -> Tachyon Internal
>> (Tachyon external works)
>> * Trying to load any pdb file on its own (psf + pdb works)
>> * Trying to load a pdb from the web
>>
>> (Also, when I resize the VMD OpenGL window with a molecule
>> present my
>> local MacOS X freezes (either hangs or the display becomes
>> brighter and
>> brighter while hanging until the user panics and hits the power-off
>> button...). This happens with both 1.8.5 and 1.8.7a14extrabonds
>> running
>> remotely on Linux (and it doesn't matter if I do 'export
>> VMDDISABLEATILINUXWORKAROUND=1' ). --- I suspect that's more a
>> problem
>> with my ancient version of MacOS X than with VMD per se, though;
>> native
>> MacOS X ppc VMD 1.8.6 has no problems.)
>>>
>>> Give those a spin and let me know if all is well. I can do a MacOS
>>> build
>>> for you tomorrow if the Linux builds work for you.
>>
>> I'd like to test the OS X build as well, please (hoping that it works
>> around the OpenGL problems and so that I can run VMD locally).
>>
>> Many thanks,
>> Oliver
>>
>>>
>>> Cheers,
>>> John Stone
>>> vmd_at_ks.uiuc.edu
>>>
>>> On Tue, Dec 18, 2007 at 05:05:02PM -0500, Oliver Beckstein wrote:
>>>> Hello John,
>>>>
>>>>> I can make you special builds with larger max bond counts.
>>>>
>>>> I appreciate this very much.
>>>>
>>>>> Will 30 be adequate? I can make it any number you like
>>>>> up to 255, but the larger the number the more memory VMD will use
>>>>> on a per-atom basis.
>>>>
>>>> To be on the safe side let's choose 64 – more are really not
>>>> going to
>>>> come up in my problems.
>>>>
>>>> Thank you,
>>>> Oliver
>>>>
>>>>>
>>>>> Cheers,
>>>>> John Stone
>>>>> vmd_at_ks.uiuc.edu
>>>>>
>>>>>
>>>>> On Mon, Dec 17, 2007 at 02:37:42PM -0500, Oliver Beckstein wrote:
>>>>>> Hello,
>>>>>>
>>>>>> is there a way to obtain binaries for Mac OS X and Linux i686
>>>>>> that
>>>>>> have
>>>>>> MAXATOMBONDS increased to, say, 30? I am (ab)using VMD for graph
>>>>>> plotting and some of my nodes simply don't behave like atoms
>>>>>> although
>>>>>> I'd like to keep using atom-like representations instead of
>>>>>> drawing
>>>>>> bonds as graphics objects because it makes it so much easier to
>>>>>> change
>>>>>> colors, materials etc.
>>>>>>
>>>>>> As I understand it, MAXATOMBONDS is defined to be 12 in Atom.h:
>>>>>> http://www.ks.uiuc.edu/Research/vmd/doxygen/Atom_8h.html
>>>>>>
>>>>>> I don't have the VMD source and as I understand it, it can be
>>>>>> a bit
>>>>>> complicated compiling it, though I'd give it a shot if this is
>>>>>> too
>>>>>> much
>>>>>> work for the developers (which I would certainly understand).
>>>>>>
>>>>>> Thanks,
>>>>>> Oliver
>>>>>>
>>>
>>
> --
> 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 

--
Oliver Beckstein * orbeckst_at_jhmi.edu
Johns Hopkins University, School of Medicine
Dept. of Physiology, Biophysics 206
725 N. Wolfe St
Baltimore, MD 21205, USA
Tel.: +1 (410) 614-4435