Designing and Building Parallel Programs

1/17/98


Click here to start


Table of Contents

Designing and Building Parallel Programs

Outline

Outline

The Need for Higher Performance

Performance Requirements

Why Parallel Computing?

Supercomputer Peak Performance

Cycle Time Trends

Processor Count Trends

The Multicomputer: an Idealized Parallel Computer

Multicomputer Architecture

Multicomputer Cost Model

How do Real Parallel Computers Fit the Model?

Distributed Memory MIMD Multiprocessor

Shared Memory MIMD Multiprocessor

Distributed Shared Memory (DSM)

Workstation Clusters

SIMD

A Simple Parallel Programming Model

Properties

Parallel Algorithm Example

Summary: Parallel Computers and Computation

Outline

Parallel Algorithm Design

A Design Methodology

Partitioning

Domain Decomposition

Functional Decomposition

Functional Decomposition (2)

Communication

Communication Structures

Global Communication

Agglomeration

Agglomeration Strategies

Mapping

Load Balancing Techniques

Example: Grid Problem

Cyclic Mapping

Block-Cyclic Mapping

Manager/Worker Structure

Example: Atmosphere Model

Atmosphere Model: Governing Equations

Atmosphere Model: Numerical Methods

Atmosphere Model: Partition

Atmosphere Model: Communication

Atmosphere Model: Agglomeration

Atmosphere Model: Mapping

Summary

Email: jake@npac.syr.edu

Home Page: http://www.npac.syr.edu/users/jake