On the design of parallel numerical methods in message passing and shared memory environments
Conference
·
OSTI ID:5542449
This paper presents a comparative view of the methodologies used in parallel methods for scientific computing. Our goal is to put in contrast the approaches taken when developing scientific software for two broad classes of parallel machines namely shared memory machines and distributed memory machines. We will illustrate our discussion with two specific algorithms namely the Alternating Directions Implicit method and the Preconditioned Conjugate Gradient method.
- Research Organization:
- Illinois Univ., Urbana (USA). Center for Supercomputing Research and Development
- DOE Contract Number:
- FG02-85ER25001
- OSTI ID:
- 5542449
- Report Number(s):
- DOE/ER/25001-58; CONF-870294-2; CSRD-614; ON: DE88003566
- Resource Relation:
- Conference: International seminar on scientific supercomputers, Paris, France, 2 Feb 1987; Other Information: Portions of this document are illegible in microfiche products
- Country of Publication:
- United States
- Language:
- English
Similar Records
Parallel implementation of the hp-version of the finite-element method on a shared-memory architecture. Final report
Parallelization of the NAS Conjugate Gradient Benchmark Using the Global Arrays Shared Memory Programming Model
Message passing and shared address space parallelism on an SMP cluster
Technical Report
·
Thu Nov 01 00:00:00 EST 1990
·
OSTI ID:5542449
Parallelization of the NAS Conjugate Gradient Benchmark Using the Global Arrays Shared Memory Programming Model
Conference
·
Fri Apr 08 00:00:00 EDT 2005
·
OSTI ID:5542449
+1 more
Message passing and shared address space parallelism on an SMP cluster
Journal Article
·
Wed Sep 25 00:00:00 EDT 2002
· Parallel Computing
·
OSTI ID:5542449
+1 more