Research Funding:
Algorithmic Issues in High Performance Computing
Support provided by National Science Foundation
Description
Research in science is shifting from its traditional focus on the discovery of new information toward the computationally intensive task of processing it. This proposal aims at developing new algorithmic techniques to support the large-scale computing requirements of high- performance applications. This project will focus on incremental, parallel, and input/output (I/O) algorithms. For each of these three computational models, several problems will be studied, with particular emphasis on data structures, graphs, and computational geometry. These problems include such key high-performance applications as sorting, FFT, matrix operations, neural nets, caching, VLSI layout, robotics, network optimization, computer-aided design, and graphics display. All three computational models are motivated by the extreme demands placed by large-scale applications, and thus it makes sense to consider hybrid computational models in order to get even better performance. Examples include solving dynamic problems in parallel and analyzing the I/O efficiency of parallel and incremental algorithms. Finally, a new framework for incremental complexity and its relation to sequential and parallel complexity will be studied.
Principal Investigator
Co-PIs
Other Investigators
Projects Supported
Details
| Amount: | $351,803 |
| Dates: | 1991-1994 |
| Status: | Complete |
|
Page Owner: saas
|
Last Modified: Fri Nov 3 11:06:22 2006
|
