Scientific Software Design: The Object-Oriented Way

by Damian Rouson , Jim Xia , Xiaofeng Xu

• ISBN13: 9780521888134
• ISBN10: 0521888131
• Format: Hardcover
• Copyright: 2011-04-29
• Publisher: Cambridge University Press

Summary

This book concerns software design. Although many current discussions of scientific programming focus on scalable performance, this book focuses on scalable design. The authors analyze how the structure of a package determines its developmental complexity according to such measures as bug search times and documentation information content. The work presents arguments for why these issues impact solution cost and time more than does scalable performance. The final chapter explores the question of scalable execution and shows how scalable design relates to scalable execution. The book's focus is on program organization, which has received considerable attention in the broader software engineering community, where graphical description standards for modeling software structure and behavior have been developed by computer scientists. These discussions might be enriched by engineers who write scientific codes. This book aims to bring such scientific programmers into discussion with computer scientists. The authors do so by introducing object-oriented software design patterns in the context of scientific simulation.

Author Biography

DR. Damian Rouson is the manager of the Reacting Flow Research Department at Sandia National Laboratories. He was formerly Section Head of Combustion Modeling and Scaling at die U.S. Naval Research Laboratory. He was assistant professor of mechanical engineering at the City University of New York, visiting assistant professor at the University of Maryland, and Special Scientist at the University of Cyprus. Damian Rouson received his Bachelors in Mechanical Engineering at Howard University and his Masters and PhD in Mechanical Engineering from Stanford University. DR. Jim Xia is-currently a software designer and tester at the IBM Canada Lab in Markham, Ontario, Canada. He received his PhD in Physics from the University of Western Ontario in 1997 DR. Xiaoffeng Xu is currently an Analyst at General Motors Corp. in Pontiac, Michigan. At this job he performs airflow and combustion CFD analysis to support base engine designs. He received his PhD in Mechanical Engineering (2003) from Iowa State, University, in Ames and is the author or co-author of 39 referred publications

Table of Contents

List of Figures	p. xi
List of Tables	p. xv
Preface	p. xvii
Acknowledgments	p. xxi
Disclaimer	p. xxii
	p. 1
Development Costs and Complexity	p. 3
Introduction	p. 3
Conventional Scientific Programming Costs	p. 7
Conventional Programming Complexity	p. 11
Alternative Programming Paradigms	p. 19
How Performance Informs Design	p. 24
How Design Informs Performance	p. 25
Elements of Style	p. 27
The Object-Oriented Way	p. 31
Nomenclature	p. 31
Object-Oriented Analysis and Design	p. 33
Encapsulation and Information Hiding	p. 36
Wrapping Legacy Software	p. 40
Composition, Aggregation, and Inheritance	p. 48
Static and Dynamic Polymorphism	p. 53
Oop Complexity	p. 54
More on Style	p. 55
Scientific Oop	p. 57
Abstract Data Type Calculus	p. 57
Analysis-Driven Design	p. 66
Design Metrics	p. 69
Complexity Theory	p. 71
Information Theory	p. 73
Still More on Style	p. 77
The Tao of Soop	p. 79
SOOP To Nuts And Bolts	p. 83
Design Patterns Basics	p. 85
Essentials	p. 85
Foundations	p. 86
Building Architecture	p. 87
Software Architecture	p. 93
Scientific Software Architecture	p. 98
Canonical Contexts	p. 99
The Lorenz Equations: A Chaotic Dynamical System	p. 100
Quantum Vortex Dynamics in a Superfluid	p. 102
Burgers' Equation: Shock Formation and Dissipation	p. 104
The Object Pattern	p. 107
The Problem	p. 107
The Solution	p. 108
Fortran implementation	p. 110
C++ Style and Tools	p. 116
C++ Implementation of Vortex	p. 122
The Consequences	p. 127
Related Patterns	p. 127
The Abstract Calculus Pattern	p. 129
The Problem	p. 129
The Solution	p. 130
Fortran Implementation	p. 131
C++ Implementation	p. 137
The Consequences	p. 140
Related Patterns	p. 141
The Strategy and Surrogate Patterns	p. 143
The Problem	p. 143
The Solution	p. 144
Fortran Implementation	p. 146
C++Implementation	p. 155
The Consequences	p. 164
Related Patterns	p. 164
The Puppeteer Pattern	p. 167
The Problem	p. 167
The Solution	p. 169
Fortran Implementation	p. 170
Ac++Tool:2D Allocatable Arrays	p. 185
C++Implementation	p. 191
The Consequences	p. 199
Related Patterns	p. 201
Factory Patterns	p. 202
The Problem	p. 202
The Solution	p. 203
Fortran Implementation	p. 205
C++ Implementation	p. 216
The Consequences	p. 226
Related Patterns	p. 227
Gumbo Soop	p. 229
Formal Constraints	p. 231
Why Be Formal?	p. 231
Problem Statement	p. 233
Side Effects in Abstract Calculus	p. 237
Formal Specification	p. 238
Modeling Arrays with Ocl	p. 238
Hermeticity	p. 239
Economy	p. 242
The Shell Pattern	p. 244
An Assertion Utility for Fortran	p. 245
Case Study: A Fluid Turbulence Solver	p. 247
Mixed-Language Programming	p. 251
Automated Interoperability Tools	p. 251
Manual Interoperability: C++/Fortran 95	p. 254
Case Study: ForTrilinos and CTrilinos	p. 259
C Interoperability in Fortran 2003	p. 261
Method Invocation	p. 269
Tabulation, Construction, and Destruction	p. 273
Polymorphism and Hierarchy Mirroring	p. 278
Discussion	p. 281
Multiphysics Architectures	p. 285
Toward a Scalable Abstract Calculus	p. 285
Amdahl's Law and Parallel Efficiency	p. 286
Automatic Parallelization: Compiler Technology	p. 292
Directive-Based Parallelization: Openmp	p. 295
Library-Based Parallelization: ForTrilinos and Mpi	p. 297
Intrinsic Parallelization: Fortran 2008 Coarrays	p. 309
Case Studies: Multiphysics Modeling	p. 317
Quantum Turbulence in Superfluid Liquid Helium	p. 318
Lattice-Boltzman Biofiuid Dynamics	p. 320
Particle Dispersion in Magnetohy drodynamics	p. 325
Radar Scattering in the Atmospheric Boundary Layer	p. 326
The Morfeus Framework	p. 330
Mathematical Background	p. 335
Interpolation	p. 335
Lagrange Interpolation	p. 335
Linear Solvers	p. 337
Gaussian Elimination	p. 337
Lu Decomposition	p. 342
Nonlinear Solvers	p. 343
Newton's Method in Id	p. 343
Newton's Method in Multiple Dimensions	p. 344
Partial Differential Equations	p. 345
The Heat Equation	p. 345
The Burgers Equation	p. 346
Numerical Analysis	p. 347
Finite Differences	p. 347
Numerical Methods for Differential Equations	p. 350
Unified Modeling Language Elements	p. 357
Use Case Diagrams	p. 357
Class Diagrams	p. 359
Classes	p. 359
Relationships	p. 362
Object Diagrams	p. 366
Sequence Diagrams	p. 367
Interactions and Messages	p. 368
Sequence Diagrams	p. 368
The Object Constraint Language	p. 370
The Context of OCL Expression	p. 370
Initial Value Rule	p. 371
Specifying Object Invariants	p. 371
Adding Pre- and Postconditions to Operations	p. 371
Bibliography	p. 373
Index	p. 379
Table of Contents provided by Ingram. All Rights Reserved.