VMD-L Mailing List

From: Bart Bruininks (b.m.h.bruininks_at_rug.nl)
Date: Mon Apr 29 2019 - 03:57:59 CDT

Hello all,

I was currently testing our visualization server and I ran into some bugs.
On top of that I had some questions.

SPECS:
1x Intel(R) Xeon(R) Gold 6150 CPU @ 2.70GHz
1x TESLA V100 (32 GB VRAM)
700 GB RAM DDR4

BUGS(?):
1) When create a .dat file of my render so i can change resolution later on
(using visual VMD 1.9.4a29), I run into a syntax error when I try to use
the latest build of the terminal tachyon (0.99 installed through apt-get).
I could change some values and see if it starts to work, but for me this is
an indication that these version have not been tested together at all and I
can expect more things to go wrong.

*$ tachyon angle_bonds_3a.dat +V -res 3840 4064 -format TARGA -o
./renders/higher_res_angle_bonds_3a.tga*
*Tachyon Parallel/Multiprocessor Ray Tracer Version 0.99 *
*Copyright 1994-2013, John E. Stone <john.stone_at_gmail.com
<john.stone_at_gmail.com>> *
*------------------------------------------------------------ *
*Parse Error:*
* Encountered a syntax error in file angle_bonds_3a.dat*
* Expected to find FOCALLENGTH*
* Actually found: FocalDist*
* Error occurred at or prior to file offset 468, line 21*
* Error position is only approximate, but should be close*

*Parser failed due to an input file syntax error.*
*Aborting render.*

2) I have two very similar files which I can send for testing, but would
rather not yet send to the whole mailing list, due to the fact that they
are part of the work of a colleague of mine. They are two gro files of the
same system with equal amount of particles and rougly the same amount of
surface (clustering). If I try to generate a quicksurf in one of them it
works just fine, if I try the same in the other, I get a CUDA domain
decomposition error (which is pretty vague...). I would not mind sending
the files and the exact instructions to the VMD team to see if you can
recreate the error. Though it might even be hardware specific.

QUESTIONS:
1) I created a .dat file (over 50 gigs) and I would like to post render it
at a chosen resolution and -AA level, bounces etc. Lets say the real deal,
and make it as pretty as possible. What would be the intended way of
achieving that? The 0.99 terminal tachyon doesn't seem to work with the
.dat's created by this VMD version and I couldn't figure out how to load a
.dat file for rendering into VMD (I do understand it has lost all its
molecular selection etc, but just rendering using the GPU accelerated OPTIX
code would be nice).

2) I reside in the Netherlands and work in the Martini CG lab of SJ
Marrink. I will be spending part of my autumn in Los Alamos and I/we were
wondering if it would be possible to visit your lab in the same period for
something like 2 weeks to become more familiar with the new ins and outs of
VMD. I would like to discuss the issues we have with CG and learn from you
guys how to get the most out of VMD. I have affinity for computers and
software development, but I think this visit would be mainly focused on
practical application of VMD on high performance machines and clusters.

Cheers and I look forward to your reaction,

Bart