From: Jason Hlozek (jason.hlozek_at_live.co.za)
Date: Mon Sep 20 2021 - 06:19:57 CDT

Hi John

I got around the issue - that specific scene was smaller and rendered correctly in VMD 1.9.3 with the 32-bit Tachyon.
Strangely, loading the same visualisation state in the same way in the newer 64bit VMD versions does produce the speckling on the VDW spheres of the water.
The water selection is done with "water and same resid as within 5 of resname AY6" - I'm not sure if the processing has changed for this then to form duplicate atom selection.

My main follow-up question though was about the 64bit Tachyon: VMD couldn't call "C:\Program Files\VMD\tachyon_WIN64.exe" and I don't see it in the installed directory.

Kind regards
Jason

________________________________
From: John Stone <johns_at_ks.uiuc.edu>
Sent: Saturday, 18 September 2021 06:54
To: Carlos Simmerling <carlos.simmerling_at_gmail.com>
Cc: Jason Hlozek <jason.hlozek_at_live.co.za>; Axel Kohlmeyer <akohlmey_at_gmail.com>; VMD Mailing List <vmd-l_at_ks.uiuc.edu>
Subject: Re: vmd-l: Tachyon not rendering large scenes

Hi,
  I'm circling back on a few threads, and I didn't see any new post
from Jason after Carlos' suggestion to watch out for replicated
geometry causing speckles. In general you want to avoid drawing any
geometry that precisely overlaps as this creates problems for ray tracing
renderers that use ray-surface intersection testing. If you have two
objects (triangles, spheres, etc) that are either duplicated or that
precisely overlap, this is a scenario that is bound to cause problems
in most ray tracing engines. Since there hasn't been any further discussion
since Carlos' post, I'm assuming that this was the problem,
However, if this was not the case and if you have ruled out this scenario
as a source of your speckling problem, I would want to have a copy of
your VMD scene to check more carefully myself.

Best,
  John Stone
  vmd_at_ks.uiuc.edu

On Fri, Sep 03, 2021 at 06:53:07AM -0400, Carlos Simmerling wrote:
> Tachyon sometimes results in speckled surfaces if you have duplicate
> representations for that region. It may not be seen in the normal VMD
> view, but will show with the rendered version.¬
> On Thu, Sep 2, 2021, 6:18 PM Jason Hlozek <[1]jason.hlozek_at_live.co.za>
> wrote:
>
> Ah I think Axel did get it right. I can render using the internal
> Tachyon (as well as TachyonL OptiX/OSPRay) on the 64 bit VMD and not 32
> bit.
> However, is the non-internal 64-bit Tachyon built into the VMD 1.9.4
> install location? Both the a54 and a51 versions attempt and fail to call
> 'tachyon_WIN64.exe' and I don't see it in the installed directory.
> The internal Tachyon seems to render VDW waters as very speckled instead
> of smooth.
> Thanks for the help and insight.
> Kind regards
> Jason
>
> --------------------------------------------------------------------------
>
> From: John Stone <[2]johns_at_ks.uiuc.edu>
> Sent: Thursday, 02 September 2021 18:33
> To: Axel Kohlmeyer <[3]akohlmey_at_gmail.com>
> Cc: Jason Hlozek <[4]jason.hlozek_at_live.co.za>; [5]vmd-l_at_ks.uiuc.edu
> <[6]vmd-l_at_ks.uiuc.edu>
> Subject: Re: vmd-l: Tachyon not rendering large scenes
> ¬
> Hi,
> ¬ Given that you're running the older 32-bit VMD, I suspect Axel is
> exactly right.¬ I would suggest that you consider installing the new
> 64-bit VMD 1.9.4 test build and see if that cures your problem.
>
> Best,
> ¬ John
>
> On Thu, Sep 02, 2021 at 12:06:30PM -0400, Axel Kohlmeyer wrote:
> >¬ ¬ ¬ The free 7GB do not matter since you are using 32-bit
> executables and they
> >¬ ¬ ¬ are limited in how much memory they can address.
> >¬ ¬ ¬ On Thu, Sep 2, 2021 at 12:03 PM Jason Hlozek
> <[1][7]jason.hlozek_at_live.co.za>
> >¬ ¬ ¬ wrote:
> >
> >¬ ¬ ¬ ¬ ¬ Hi John
> >¬ ¬ ¬ ¬ ¬ The conventional Snapshot OpenGL render does produce the
> correct
> >¬ ¬ ¬ ¬ ¬ '.bmp'.√*
> >¬ ¬ ¬ ¬ ¬ Re system resources for Tachyon, the GPU is not engaged at
> all and RAM
> >¬ ¬ ¬ ¬ ¬ usage increases by 2GBs but there's still 7GBs of free head
> room by the
> >¬ ¬ ¬ ¬ ¬ time Tachyon aborts.√*
> >¬ ¬ ¬ ¬ ¬ The external '.dat' scene file is 422MBs and seems to
> consist
> >¬ ¬ ¬ ¬ ¬ predominantly of 3D coordinates so it seems like the grid
> you reference
> >¬ ¬ ¬ ¬ ¬ is at least partially built?
> >¬ ¬ ¬ ¬ ¬ Kind regards
> >¬ ¬ ¬ ¬ ¬ Jason
> >
> >¬ ¬ ¬
> --------------------------------------------------------------------------
> >
> >¬ ¬ ¬ ¬ ¬ From: John Stone <[2][8]johns_at_ks.uiuc.edu>
> >¬ ¬ ¬ ¬ ¬ Sent: Thursday, 02 September 2021 15:43
> >¬ ¬ ¬ ¬ ¬ To: Jason Hlozek <[3][9]jason.hlozek_at_live.co.za>
> >¬ ¬ ¬ ¬ ¬ Cc: [4][10]vmd-l_at_ks.uiuc.edu <[5][11]vmd-l_at_ks.uiuc.edu>
> >¬ ¬ ¬ ¬ ¬ Subject: Re: vmd-l: Tachyon not rendering large scenes
> >¬ ¬ ¬ ¬ ¬ √*
> >¬ ¬ ¬ ¬ ¬ Have you checked to ensure that you're not running your
> system out of
> >¬ ¬ ¬ ¬ ¬ memory?√*¬ Do you have a conventional OpenGL rendering of
> the scene in
> >¬ ¬ ¬ ¬ ¬ question for comparison?√*¬ When you render the scene to an
> external
> >¬ ¬ ¬ ¬ ¬ file,
> >¬ ¬ ¬ ¬ ¬ how large is the scene file?√*¬ I notice that in the output
> you included
> >¬ ¬ ¬ ¬ ¬ below, there's no console output indicating that spatial
> acceleration
> >¬ ¬ ¬ ¬ ¬ grid
> >¬ ¬ ¬ ¬ ¬ structures were built, despite the fact that you have the +V
> flag
> >¬ ¬ ¬ ¬ ¬ enabled.
> >¬ ¬ ¬ ¬ ¬ That is one of the reasons I'm wondering about the memory
> exhaustion
> >¬ ¬ ¬ ¬ ¬ issue.√*¬ Another possibility is a runaway grid build of
> some sort, but
> >¬ ¬ ¬ ¬ ¬ I haven't seen a problem like that for a very long time, so
> the odds
> >¬ ¬ ¬ ¬ ¬ are somewhat against it.
> >
> >¬ ¬ ¬ ¬ ¬ Best,
> >¬ ¬ ¬ ¬ ¬ √*¬ John Stone
> >¬ ¬ ¬ ¬ ¬ √*¬ [6][12]vmd_at_ks.uiuc.edu
> >
> >¬ ¬ ¬ ¬ ¬ On Thu, Sep 02, 2021 at 12:29:11PM +0000, Jason Hlozek
> wrote:
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ Hi
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ I'm trying to render an overlaid view of
> multiple ligands in the
> >¬ ¬ ¬ ¬ ¬ same
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ protein pocket (i.e. a very busy scene).
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ If I go above 2-3 protein/ligand pairs in the
> scene, then VMD
> >¬ ¬ ¬ ¬ ¬ still writes
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ the scene file and calls Tachyon as usual but
> after a few
> >¬ ¬ ¬ ¬ ¬ seconds of
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ processing, instead of proceeding to the
> completion % bar,
> >¬ ¬ ¬ ¬ ¬ Tachyon instead
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ just ends with VMD saying the render is complete
> even though it
> >¬ ¬ ¬ ¬ ¬ didn't
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ run. The same thing happens if I try to call
> tachyon from the
> >¬ ¬ ¬ ¬ ¬ command
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ line:
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ "C:\Program Files (x86)\University of
> Illinois\VMD>
> >¬ ¬ ¬ ¬ ¬ .\tachyon_WIN32.exe +V
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬
> "C:\Users\user-pc\AppData\Local\VirtualStore\Program Files
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ (x86)\University of
> Illinois\VMD\mmpl3_overlay_test"
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ Tachyon Parallel/Multiprocessor Ray Tracer√*
> √*¬ Version 0.99.1
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ Copyright 1994-2014,√* √* √*¬ John E. Stone
> >¬ ¬ ¬ ¬ ¬ <[7][13]john.stone_at_gmail.com>
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬
> ------------------------------------------------------------
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ C:\Program Files (x86)\University of
> Illinois\VMD>"
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ I thought it might be timing out and tried
> increasing the
> >¬ ¬ ¬ ¬ ¬ windows Timeout
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ Detection and Recovery (TDR) registry variables
> as well as
> >¬ ¬ ¬ ¬ ¬ disabling them
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ but it made no difference.
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ Any suggestions please?
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ Kind regards
> >¬ ¬ ¬ ¬ ¬ >√* √* √*¬ Jason
> >
> >¬ ¬ ¬ ¬ ¬ --
> >¬ ¬ ¬ ¬ ¬ NIH Center for Macromolecular Modeling and Bioinformatics
> >¬ ¬ ¬ ¬ ¬ Beckman Institute for Advanced Science and Technology
> >¬ ¬ ¬ ¬ ¬ University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
> >¬ ¬ ¬ ¬ ¬ [8][14]http://www.ks.uiuc.edu/~johns/√* √* √* √* √* √* √* √*
> √* √*¬ Phone:
> >¬ ¬ ¬ ¬ ¬ 217-244-3349
> >¬ ¬ ¬ ¬ ¬ [9][15]http://www.ks.uiuc.edu/Research/vmd/√* √* √* √* √*
> >
> >¬ ¬ ¬ --
> >¬ ¬ ¬ Dr. Axel Kohlmeyer √* [10][16]akohlmey_at_gmail.com √*
> [11][17]https://urldefense.com/v3/__http://goo.gl/1wk0__;!!DZ3fjg!q4u7ujN0GlH2Sq1TU6q3Y4JWOv8vt-T-GNeaFoJhPpa6WJO9dYyfTW6pZsLJT09rZw$
> >¬ ¬ ¬ College of Science & Technology, Temple University, Philadelphia
> PA, USA
> >¬ ¬ ¬ International Centre for Theoretical Physics, Trieste. Italy.
> >
> > References
> >
> >¬ ¬ ¬ Visible links
> >¬ ¬ ¬ 1. [18]mailto:jason.hlozek_at_live.co.za
> >¬ ¬ ¬ 2. [19]mailto:johns_at_ks.uiuc.edu
> >¬ ¬ ¬ 3. [20]mailto:jason.hlozek_at_live.co.za
> >¬ ¬ ¬ 4. [21]mailto:vmd-l_at_ks.uiuc.edu
> >¬ ¬ ¬ 5. [22]mailto:vmd-l_at_ks.uiuc.edu
> >¬ ¬ ¬ 6. [23]mailto:vmd_at_ks.uiuc.edu
> >¬ ¬ ¬ 7. [24]mailto:john.stone_at_gmail.com
> >¬ ¬ ¬ 8. [25]http://www.ks.uiuc.edu/~johns/
> >¬ ¬ ¬ 9. [26]http://www.ks.uiuc.edu/Research/vmd/
> >¬ ¬ 10. [27]mailto:akohlmey_at_gmail.com
> >¬ ¬ 11.
> [28]https://urldefense.com/v3/__http://goo.gl/1wk0__;!!DZ3fjg!rj4UDi-KzG2GSD5hO09MyAxrj5BSDVzxkpJZHIwlgXP4XNJHGakyU-WSF9s_cVoelQ$
>
> --
> NIH Center for Macromolecular Modeling and Bioinformatics
> Beckman Institute for Advanced Science and Technology
> University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
> [29]http://www.ks.uiuc.edu/~johns/¬ ¬ ¬ ¬ ¬ ¬ ¬ ¬ ¬ ¬ Phone:
> 217-244-3349
> [30]http://www.ks.uiuc.edu/Research/vmd/¬ ¬ ¬ ¬ ¬
>
> References
>
> Visible links
> 1. mailto:jason.hlozek_at_live.co.za
> 2. mailto:johns_at_ks.uiuc.edu
> 3. mailto:akohlmey_at_gmail.com
> 4. mailto:jason.hlozek_at_live.co.za
> 5. mailto:vmd-l_at_ks.uiuc.edu
> 6. mailto:vmd-l_at_ks.uiuc.edu
> 7. mailto:jason.hlozek_at_live.co.za
> 8. mailto:johns_at_ks.uiuc.edu
> 9. mailto:jason.hlozek_at_live.co.za
> 10. mailto:vmd-l_at_ks.uiuc.edu
> 11. mailto:vmd-l_at_ks.uiuc.edu
> 12. mailto:vmd_at_ks.uiuc.edu
> 13. mailto:john.stone_at_gmail.com
> 14. http://www.ks.uiuc.edu/~johns/¬
> 15. http://www.ks.uiuc.edu/Research/vmd/¬
> 16. mailto:akohlmey_at_gmail.com
> 17. https://urldefense.com/v3/__http://goo.gl/1wk0__;!!DZ3fjg!pHrdmIBjl-44ZUSX62jQy1svGMT-P7vhnlpSYAlcI1rOOJCLhI76AgbL1krljKKIXg$
> 18. mailto:jason.hlozek_at_live.co.za
> 19. mailto:johns_at_ks.uiuc.edu
> 20. mailto:jason.hlozek_at_live.co.za
> 21. mailto:vmd-l_at_ks.uiuc.edu
> 22. mailto:vmd-l_at_ks.uiuc.edu
> 23. mailto:vmd_at_ks.uiuc.edu
> 24. mailto:john.stone_at_gmail.com
> 25. http://www.ks.uiuc.edu/~johns/
> 26. http://www.ks.uiuc.edu/Research/vmd/
> 27. mailto:akohlmey_at_gmail.com
> 28. https://urldefense.com/v3/__http://goo.gl/1wk0__;!!DZ3fjg!rj4UDi-KzG2GSD5hO09MyAxrj5BSDVzxkpJZHIwlgXP4XNJHGakyU-WSF9s_cVoelQ$
> 29. http://www.ks.uiuc.edu/~johns/
> 30. http://www.ks.uiuc.edu/Research/vmd/

--
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/