next up previous contents
Next: Frequently Asked Questions Up: Illuminator Distributed Visualization Library Previous: Contents   Contents

Introduction

PETSc is a great distributed object manager and Newton-Krylov nonlinear system solver. Geomview is a wonderful 3-D surface visualization tool. It seemed logical marry the two, and produce isoquant contour surfaces from PETSc distributed objects in parallel, then display them using Geomview.

This is a first cut attempt at such a project. Each node generates its own isoquant contour surface(s), then all of those generated triangles are gathered to the first node for viewing in Geomview. Right now, it does not do a very efficient job of this: it sends three vertices for each triangle to the first node, which sends those three vertices to Geomview, in plain text format. In the future, it may do a number of things to reduce the number of triangles involved, reduce the number of vertices generated per triangle, and more efficiently transmit the vertices to Geomview.

Long-term, there are other nice 3-D viewers out there such as Data Explorer from IBM, some kind of unified front end to them would be nice. Furthermore, with parallel array storage, much more powerful parallel visualization concepts are possible. Like if we can assign to each point a certain luminosity and transparency, it's easy to integrate those values along a line through the locally-stored part of the array to produce a total transparency and luminosity for that line in the local array, then just combine that with other line integrals through the other local sections (maybe in e-vas?), and we have generated a total image in parallel. However, to do this easily each section must be convex, or there will likely be regions with points both in front and behind a given other region; so this will work with PETSc distributed arrays, but not with general partitioned finite element meshes. But as you can see, this opens many interesting avenues for exploration...

For now, this humble beginning. Share and enjoy.


next up previous contents
Next: Frequently Asked Questions Up: Illuminator Distributed Visualization Library Previous: Contents   Contents
root 2002-06-18