3D Simulation / Visualization

MitchellJames · Post by **MitchellJames** » Sat May 31, 2008 3:46 am

I found an interesting visualization tool at
http://software.sci.utah.edu/scirun.html

To try it out download version 3.0.2. What is interesting is that it is a generalization of what Indrek Mandre was doing with ephi. The concept would be to create a data set within a framework like www.cactuscode.org dumping the data into an HDF5 output file which could be read into scirun.

This way lots of interesting simulation could be done without much code investment. I would like to see a complete 3D WB-6 simulation done within cactus and scirun used for visualization. Only a few plugin modules may need to be added and I expect some work on the I/O routines because the HDF5 people just made a major new release which will probably have to be incorporated.

I have been working through the tutorials and took a picture of a current file that shows a vector field (in this case electrical) a mesh, and surface boundaries at:
http://www.nuclearfusionpower.us/misc/t ... _field.png

There is a fusion app based on scirun at http://software.sci.utah.edu/fusion.html#fusionviewer
which is where I started looking at this tool. But I wouldn't recommend wasting time with this version. It is based on a very old scirun and has a lot of compilation problems that took me several days to resolve by which time I discovered that everything that I wanted was in the new version of scirun.

Some other interesting simulation sites are:

http://www-unix.mcs.anl.gov/petsc/petsc-2/
http://www.extreme.indiana.edu/xcat/#Downloads
http://www.grycap.upv.es/slepc/description.htm
http://www.fusiongrid.org/

It would be extremely helpful if somebody had working knowledge of any of these frameworks.

Mitchell James

drmike · Post by **drmike** » Sat May 31, 2008 1:21 pm

I downloaded the linux version last night, but it looks like it is 1.24.2. Probably not worth dinking with. Do you know what the data format is supposed to look like? Maybe I can generate files in the correct format and other people can get it to work with the viewer.

MitchellJames · Post by **MitchellJames** » Sat May 31, 2008 3:51 pm

drmike wrote:I downloaded the linux version last night, but it looks like it is 1.24.2. Probably not worth dinking with. Do you know what the data format is supposed to look like? Maybe I can generate files in the correct format and other people can get it to work with the viewer.

The first link in my post should have taken you to the right download area. If you followed the fusion link then you would have ended up in the wrong download area. Or you can get to the latest version from the ad on the home page.

I use a Gentoo Linux build with all the latest revisions which is working fine for the scirun distributions. I tried doing the fusion viewer download and build on the latest Ubuntu distribution which was a complete disaster. However the Ubuntu distribution is working great for my wife (except for a minor screen resolution problem) because she just wants it to work like a windows machine, no development capability. Your ability to get it working may depend on how much toward the software development side of things your Linux is set-up.

HDF5 is described at http://hdf.ncsa.uiuc.edu/index.html.

MitchellJames · Post by **MitchellJames** » Sat May 31, 2008 3:57 pm

My main goal for simulation is to be able to watch individual electrons move through various non uniform magnetic and electric fields.

drmike · Post by **drmike** » Sat May 31, 2008 8:55 pm

OK, thanks, I'll go check that out.