Contents

• List of Figures
• List of Tables
• 1 Introduction
• 2 Coarse Grain Parallel Machines
• 3 High Performance Fortran
  • 3.1 Fortran 90 Features
  • The HPF Extensions to Fortran 90
    • 3.2.1 Compiler Directives
    • 3.2.2 Parallel Constructs
    • 3.2.3 Intrinsics
    • 3.2.4 Extrinsics
  • 3.3 Anatomy of a typical HPF compiler
  • 3.4 HPF and Fortran 95
  • 3.5 HPF-2.0 -The Next High Performance Fortran Standard
    • 3.5.1 Data mapping
    • 3.5.2 Data and Task Parallelism
    • 3.5.3 Extensions to the HPF Intrinsic and Library Procedures
    • 3.5.4 Miscellaneous
• 4 Application Classification
• 5 Embarrassingly Parallel Applications
  • 5.1 Conversion between Random Numbers
  • 5.2 Data Analysis
  • 5.3 Unindexed Search
  • 5.4 Methods for Stock Pricing Models
  • 5.5 Summary
• 6 Synchronous Applications
  • 6.1 Full Matrix Algorithms
    • 6.1.1 Gaussian Elimination
    • 6.1.2 Matrix and Vector Multiplications
  • 6.2 Structured Sparse Linear Systems
    • 6.2.1 Single Grid Methods
  • 6.3 Unstructured Sparse Linear Systems
  • 6.4 Crystalline Monte Carlo
    • 6.4.1 Potts Model
  • 6.5 Quenching Algorithms
  • 6.6 Computational Fluid Dynamics
    • 6.6.1 Poisson Equation using ADI
    • 6.6.2 Panel Methods
  • 6.7 Particle Dynamics
    • 6.7.1 Direct Methods
  • 6.8 Image and Signal Processing
    • 6.8.1 Windowing Applications
    • 6.8.2 Histogramming Applications
    • 6.8.3 Image Transformations
    • 6.8.4 Fast Fourier Transform
  • 6.9 Neural Networks
    • 6.9.1 Hopfield Networks
    • 6.9.2 Learning Vector Quantizers
  • 6.10 Evolutionary Computation
  • 6.11 Graph Algorithms
    • 6.11.1 All-pairs shortest path algorithm
    • 6.11.2 Minimum spanning tree
    • 6.11.3 Single Source Shortest Path
  • 6.12 Summary of Synchronous Applications
• 7 Conclusions
• A Example Codes
  • A.1 NAS EP Benchmark - Tabulation of Random Numbers
  • A.2 Hough Transformation
  • A.3 Binomial Tree Stock Market Simulation
  • A.4 Gaussian Elimination
  • A.5 Sweep Over an Unstructured Mesh
  • A.6 2-Dimensional Potts Model
  • A.7 2-D Binary Phase Quenching of Cahn Hilliard Cook Equation
  • A.8 Solution of 2-D Poisson Equation by the Alternating Direct Implicit Method
    • A.8.1 Main program for the ADI code
    • A.8.2 Routine to solve a tridiagonal matrix, for the ADI code
    • A.8.3 Routine to find the Right Hand Side, for the ADI code
  • A.9 Direct N-Body Solver
  • A.10 Image Template Matching
  • A.11 2-Dimensional Fast Fourier Transform
    • A.11.1 Main Program for the 2-D FFT Code
    • A.11.2 1-Dimensional FFT routine for the 2-D FFT Code
  • A.12 One Dimensional Fast Fourier Transform
  • A.13 2-Dimensional Convolution
  • A.14 Hopfield Neural Network
• References