Graduate Project
 

Integrating a parallel computer and a heterogeneous workstation cluster into a metacomputer system

Öffentlich Deposited

Herunterladbarer Inhalt

PDF Herunterladen
https://ir.library.oregonstate.edu/concern/graduate_projects/5425kk119

Descriptions

Attribute NameValues
Creator
Abstract
  • Two types of parallel computers commonly used or solving large scientific problems are clusters of workstations and distributed-memory multicomputers. Each system has strengths and weaknesses for this task. Workstation clusters have a high performance to cost ratio and the advantage of the latest processors. Workstations are commonly under-utilized and can provide an inexpensive source of CPU cycles. However, clusters of workstations cannot compete with the performance of a dedicated supercomputer. This research proposes that creating a metacomputer combining different types of parallel computers can provide some of the advantages of each separate system. Specifically, I have inte­grated a distributed-memory parallel computer (the MEIKO CS-2) with a heterogeneous cluster of workstations. The integrated system uses the CHARM parallel-programming environment to provide for machine-independence and ease of programming in this heterogeneous environment. The availability of processing capacity limits the size and complexity of the types o[ problems that can be efficiently solved. By creating a meta.computer the amount of processing capacity can be increased at relatively low costs. The low cost of the system and the fact that it is easily reconfigurable make it a good choice for solving large-scale Grand Challenge type scientific problems.
Resource Type
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Academic Affiliation
Urheberrechts-Erklärung
Publisher
Peer Reviewed
Language

Beziehungen

Parents:

This work has no parents.

Artikel