|
37:
Commodity-Based Scalable Visualization
Monday, Full Day, 8:30 am - 5 pm
Room 511
A
new approach to scalable graphics systems proposes to
assemble clusters of PC-like nodes that include 3D hardware
acceleration. This clustering approach can also be used
to drive multiple displays arranged in a tiled configuration
that generates a single high-resolution display. The challenge
is to demonstrate that the aggregate performance of such
systems can be harnessed effectively, especially for large
datasets. This course surveys efforts to develop such
systems.
Prerequisites
Interest in parallel graphics and visualization for large
graphics and visualization applications. Some familiarity
with 3D graphics systems and applications.
Topics
Scalable graphics architectures, scientific visualization,
parallel visualization and rendering (polygonal and volume),
graphics sorting, PC graphics, display technology and
tiled-display systems, and hardware integration approaches
for building cluster-based graphics systems. How parallel
rendering can be achieved on clustered architectures.
Overview of recent results.
Organizer
Constantine Pavlakos
Sandia National Laboratories
Lecturers
Matthew Eldridge
Greg Humphreys
Stanford University
Adam Finkelstein
Princeton University
Randall Frank
Lawrence Livermore National Laboratory
Alan Heirich
Compaq Computer Corporation
Allen McPherson
Los Alamos National Laboratory
Constantine Pavlakos
Sandia National Laboratories |
|
