Development of data storage and visualization tools for ORBIT

GEOFFREY SHELBURNE (Augsburg College, Minneapolis, MN 55454)

Abstract:

The ORBIT code is a numerical code (developer: Roscoe White), written in Fortran, which integrates the Hamiltonian equations of the guiding center motion for particles. Previously, ORBIT used the National Center for Atmospheric Research (NCAR) graphics package and did not produce data files that were inaccessible by other graphics packages. This did not allow comparison of ORBIT configurations or adaptation for more flexible visualization software. Different options in ORBIT access different subroutines and variables. Therefore a dynamic subroutine capable of differentiating between the various options was necessary. We replaced NCAR graphics calls with a subroutine that stores all resulting data from each run in a binary network common data format (netCDF) file using the easy CDF (ezCDF) libraries. During each run, the data is sent to a single subroutine which determines characteristics of each data array. According to the ORBIT parameters of each run, data are labeled and stored as scalar variables, time indexed profiles, or parameter specific arrays. Data is then accessed directly by the ElVis graphics software and can be viewed in all operating systems. Alternatively, data can be stored in ASCII format and accessed by a package such as SuperMongo. We examine the limitations of the previous data generating configuration as well as the updated version. We find that the newly developed data generation and exploration methods allow greater portability across platforms as well increased ability to explore data generated from various ORBIT configurations. The netCDF format allows greater accessibility, flexibility, and smaller file sizes than ASCII format data. Future work will focus on simplifying the code and moving away from the use of the ezCDF libraries, which are only capable of storing arrays of 3 dimensions or less.


Last modified: Mon Aug 16 17:30:00 EDT 2004