[Molvis-list] hardware upgrades to increase rasmol performance on very large systems.

[Keith McLaughlin]

Herbert,

1. I believe the coloring problems are a known issue which was mentioned
during the 2.7.2 release, though I am unable to find this documented right
now. I'm running linux with 32bit rasmol. In 2.7.2 if I load an image, and
use "color temperature" everything is fine, but if I then zap the image and
load a second one, color temperature no longer works.

2. We are using a sequence of commands as follows:
load xyz somefile.xyz
rotate y -90
color temperature
zoom 440
slab 46
set ambient 100
write someoutputfile.ppm
zap

the .xyz files are like the following:

  1310720     1    32    32   160
  0.114102400000E+03  0.114102400000E+03  0.570512000000E+03  90.0000000
90.0000000  90.0000000 Time(ps)=   4.0000
 C -5.6590E+01 -5.6606E+01 -2.9792E+02 -3.248E+00  0.000E+00 1 0
.... more atoms....
.... etc...


I usually only run 2-3 copies of rasmol. Running fewer or more makes no real
difference, and I begin to see failures at around 500-700 loads.

The workstation I use is an AMD sempron 1800+ with 1.5gb of RAM (Centos 4.4
).

Thanks for your help.
Keith McLaughlin


On 4/13/07, Herbert J. Bernstein <yaya at bernstein-plus-sons.com> wrote:
>
> Dear Keith,
>
>    Thank you for bringing this to my attention.  We would like to fix
> it.  You have raised two problems that clearly need to be fixed in
> the latest version:
>
>    1.  You have described "coloring problems when loading images one
> after the other."  Could you clarify and provide an example?
> It sounds like the color maps are not getting reset properly.
> Are you using a 32-bit, 16-bit or 8-bit version of RasMol?
>
>    2.  You say that after about "500 images rasmol gives an error and
> is unable to load any new images, saying that the file is not found".
> Could you be more specific about the sequence of commands and data files
> used, so we can reproduce this error and fix it.  it sounds like a
> memory leak.
>
> Are you running one copy of RasMol and loading and zapping multiple images
> or are you running multiple copies of RasMol.  In the former case the
> problem need to be addressed primarily directly in RasMol.  In the
> later case, the problem needs to be adressed more at the operating
> system level.  In either case, we would like to solve it.
>
> Please specify the hardware platform and operating system you use, so
> we can try to work in a similar environment.
>
> In the meantime, I would agree with your guess that adding more memory
> would help, but if you are keeping 500 copies of RasMol open and are
> in a unix environment, you probably need to increase process limits
> as well.  If you are in an MS windows environment, there may not be
> any easy answer.
>
>    Regards,
>      Herbert
>
>
> At 7:28 PM -0400 4/12/07, Keith McLaughlin wrote:
> >Hello all;
> >
> >My research involves 4 million atom MD simulations and we've decided to
> use
> >rasmol for visualization. From my experiences, I have two questions and
> am
> >hoping that someone here may be able to help.
> >
> >1. Currently I'm using version 2.7.1 because it doesn't have the coloring
> >problems when loading images one after another. Unfortunately, after
> about
> >500 images rasmol gives an error and is unable to load any new images
> saying
> >that the file is not found, or something to that effect. Does anyone have
> >experience with handling this?
> >
> >2. We are planning on some upgrades, and are kind of curious as to which
> >components will help the most with increasing the speed in which these
> files
> >are loaded: memory or video memory, or perhaps processor. We mostly
> suspect
> >that adding more memory would be the best choice, but I'd like to hear
> other
> >opinions.
> >
> >Thanks!
> >--
> >Keith McLaughlin
> >University of South Florida
> >Physics Dept. PHY225
> >4202 E. Fowler Ave
> >Tampa, FL 33620
> >
> >--
> >Keith McLaughlin
> >University of South Florida
> >Physics Dept. PHY225
> >4202 E. Fowler Ave
> >Tampa, FL 33620
> >_______________________________________________
> >Molvis-list mailing list
> >Molvis-list at bioinformatics.org
> >https://bioinformatics.org/mailman/listinfo/molvis-list
>
>
> --
> =====================================================
>   Herbert J. Bernstein, Professor of Computer Science
>     Dowling College, Kramer Science Center, KSC 121
>          Idle Hour Blvd, Oakdale, NY, 11769
>
>                   +1-631-244-3035
>                   yaya at dowling.edu
> =====================================================
> _______________________________________________
> Molvis-list mailing list
> Molvis-list at bioinformatics.org
> https://bioinformatics.org/mailman/listinfo/molvis-list
>



-- 
Keith McLaughlin
University of South Florida
Physics Dept. PHY225
4202 E. Fowler Ave
Tampa, FL 33620