On the design of a concurrent image analysis system
Because programming parallel computers is challenging, their use has been limited. To make the technology more accessible it is necessary to develop programming environments which hide the details. An environment for image analysis on concurrent computers is described which provides fundamental services for parallelism. The system is organized in a tripartite abstraction hierarchy. The topmost level can be programmed as though it were an ordinary sequential computer. Distributed memory and message passing appear only at lower levels. An operational taxonomy consisting of mathematical transformations, memory management, domain decomposition, and input/output facilities for relevant data structures form the nucleus of the environment. A collection of applications are enumerated; advantages and drawbacks are discussed. 18 refs.
- Research Organization:
- Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
- Sponsoring Organization:
- DOE/ER; DOE/NE
- DOE Contract Number:
- AC05-84OR21400
- OSTI ID:
- 6896112
- Report Number(s):
- CONF-900780-4; ON: DE90011388
- Resource Relation:
- Conference: ISRAM '90: international symposium on robotics and automation, Vancouver (Canada), 18-20 Jul 1990
- Country of Publication:
- United States
- Language:
- English
Similar Records
The multicomputer toolbox: Scalable parallel libraries for large-scale concurrent applications
PREPARING FOR EXASCALE: ORNL Leadership Computing Application Requirements and Strategy
Related Subjects
IMAGE PROCESSING
PARALLEL PROCESSING
COMPUTER ARCHITECTURE
ROBOTS
COMPUTERIZED CONTROL SYSTEMS
ALGORITHMS
DATA PROCESSING
DESIGN
CONTROL SYSTEMS
MATHEMATICAL LOGIC
PROCESSING
PROGRAMMING
990200* - Mathematics & Computers