VMD-L Mailing List

From: Norman Geist (norman.geist_at_uni-greifswald.de)
Date: Thu Oct 30 2014 - 09:25:36 CDT

Hey,

> -----Ursprüngliche Nachricht-----
> Von: John Stone [mailto:johns_at_ks.uiuc.edu]
> Gesendet: Donnerstag, 30. Oktober 2014 14:50
> An: Norman Geist
> Cc: VMD Mailing List
> Betreff: Re: vmd-l: BUG(s): vmd-1.9.2b
>
> Hi Norman,
> Items 1) and 2) should already be fixed in the next VMD builds.
> I believe 1) may have been a bug in OptiX itself, as I was unable to
> reproduce that problem with the current OptiX 3.6.3, and I recall
> having written code to handle this case almost a year ago.
> Item 2) I have improved in the current test versions, although it
> could likely still be further improved.

Ok, I'll get the latest build and try it next week.

>
> I don't fully understand the issue 3) you describe below, can you
> explain what you mean by "background"?

VMD Main->Save Coordinates->"[x] Save in background" doesn't work, only an
empty file is created. ;)

>
> Item 4) is unavoidable. For good performance ray tracing requires
> that almost all of the geometry in the scene be directly accessible
> in GPU RAM. Representations that are disabled do not take up GPU RAM
> in general, but there is one exception to this:
>
> The QuickSurf representation keeps a cache in GPU RAM which can become
> quite large in some cases. If the size of the QuickSurf memory cache
> gets too large, it doesn't leave much memory for use by OptiX.
> You can overcome this issue by setting this environment variable
> in your .vmdrc (or in the VMD startup scripts):
> set env(VMDQUICKSURFMINMEM) 1

I see, that's a hint, I'll see if this is the problem for us.

Thanks =D

Norman Geist

>
> This will force QuickSurf to release its memory upon completion of
> surface calculations.
>
> I'm working on adding a code path to allow VMD/OptiX to force QuickSurf
> to dump its GPU memory cache upon entry into ray tracing.
>
> Cheers,
> John
>
>
> On Wed, Oct 29, 2014 at 10:10:16AM +0100, Norman Geist wrote:
> > Hey,
> >
> > seems I encountered two bugs:
> > 1. The Tachyon-Optix renderer fails unrecoverable when all
> > representations are disabled and also the axis, so in fact there's
> nothing
> > to render.
> >
> > It might sound a stupid idea to render _nothing_ in the 1^st
> place, but
> > can IMHO happen when doing user defined animated movies, at least
> >
> > it shouldn't crash.
> >
> > 2. Tachyon Optix doesn't correctly render msms wireframes
> compared
> > to the traditional Tachyon.
> >
> > 3. Saving coordinates in background seems to not be
> functional, only
> > an empty file is created. (Tested from the the GUI only)
> >
> > 4. This one might be unavoidable, but VMD seems to have all
> > representations in VRAM, also if they are disabled, this makes
> >
> > it hard to use Tachyon-Optix for complex video animations as
> there's not
> > enough VRAM left.
> >
> >
> >
> > Norman Geist
> >
> >
> >
> > ------------------------------------------------------------------
> ---------
> >
> > [1][IMG] Diese E-Mail ist frei von Viren und Malware, denn der
> [2]avast!
> > Antivirus Schutz ist aktiv.
> >
> > References
> >
> > Visible links
> > 1. http://www.avast.com/
> > 2. http://www.avast.com/
>
> --
> NIH Center for Macromolecular Modeling and Bioinformatics
> Beckman Institute for Advanced Science and Technology
> University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
> http://www.ks.uiuc.edu/~johns/ Phone: 217-244-3349
> http://www.ks.uiuc.edu/Research/vmd/ 

---
Diese E-Mail ist frei von Viren und Malware, denn der avast! Antivirus Schutz ist aktiv.
http://www.avast.com