Contents

• Contents
• 1. Introduction
• 2. Usage
  • 2.1 Roadmap
  • 2.2 Using the C Preprocessor
  • 2.3 Solvers
    • 2.3.1 Common Arguments
    • 2.3.2 Stationary Methods
    • 2.3.3 Krylov Subspace Solvers
    • 2.3.4 Example using cpp variables
    • 2.3.5 Example using hardwired values
    • 2.3.6 Custom preconditioning via reverse communication
  • 2.4 JTpack77 as a Toolkit
• 3. Design Goals and Implementation Details
  • 3.1 Adherence to standards and use of extensions
  • 3.2 Naming of routines, files, etc.
  • 3.3 Use of parameter vectors and cpp variables
  • 3.4 Automatic determination of environment
  • 3.5 Automatic dependency generation
  • 3.6 Dynamic memory
  • 3.7 Input / Output
• 4. Storage Parameters
  • 4.1 _JT_ storage_ unknown_ - unknown format
  • 4.2 _JT_ storage_ full_ - full (conventional) format
  • 4.3 _JT_ storage_ EISPACK_ - EISPACK format
  • 4.4 _JT_ storage_ LAPACK_ - LAPACK format
  • 4.5 _JT_ storage_ blocktridiaggeneral_ - general block-tridiagonal format
  • 4.6 _JT_ storage_ blocktridiagsymmetric_ - symmetric block-tridiagonal format
  • 4.7 _JT_ storage_ blocktridiagdiagoff_ - block-tridiagonal with diagonal off-diagonal blocks
  • 4.8 _JT_ storage_ blocktridiagdiagoffsymmetric_ - symmetric, block-tridiagonal with diagonal off-diagonal blocks
  • 4.9 _JT_ storage_ ELL_ - ELLPACK-ITPACK, or row-packed format
  • 4.10 _JT_ storage_ ELLtranspose_ - ELLPACK-ITPACK transpose
  • 4.11 _JT_ storage_ COO_ - coordinate format
  • 4.12 _JT_ storage_ RSS_ - row sequential format
  • 4.13 _JT_ storage_ CSS_ - column sequential format
• 5. Control Parameters
  • 5.1 Integer Parameters
    • 5.1.1 iparm(_JT_ safety_) - toggles extra input checking
    • 5.1.2 iparm(_JT_ iter_) - number of iterations completed
    • 5.1.3 iparm(_JT_ out_) - output mode
    • 5.1.4 iparm(_JT_ luout_) - logical unit for output
    • 5.1.5 iparm(_JT_ luerr_) - logical unit for error messages
    • 5.1.6 iparm(_JT_ itmax_) - maximum number of iterations to allow
    • 5.1.7 iparm(_JT_ stop_) - stopping test
    • 5.1.8 iparm(_JT_ norm_) - norm
    • 5.1.9 iparm(_JT_ resid_) - determines how often to compute and/or use true residual (b - Ax) and how often to compute error estimate based on true residual (errt)
    • 5.1.10 iparm(_JT_ method_) - determines which method to use if multiple methods are implemented in one routine
    • 5.1.11 iparm(_JT_ prein_) - indicates whether or not a preconditioner will be passed in
    • 5.1.12 iparm(_JT_ pre_) - controls preconditioning
    • 5.1.13 iparm(_JT_ steps_) - number of Jacobi or SSOR steps to take
    • 5.1.14 iparm(_JT_ nold_) - number of previous vectors to use in truncated methods
  • 5.2 Floating-point Parameters
    • 5.2.1 rparm(_JT_ eps_) - convergence criterion
    • 5.2.2 rparm(_JT_ epspre_) - convergence criterion for preconditioning systems
    • 5.2.3 rparm(_JT_ omega_) - relaxation parameter
    • 5.2.4 rparm(_JT_ anorm_) - norm of coefficient matrix
    • 5.2.5 rparm(_JT_ tiny_) - smallest floating-point value to worry about
• Bibliography
• A. Building and Testing JTPACK77
  • A1. Requirements
  • A2. Building The Code
    • A2.1 Code Organization
    • A2.2 Building the Library
    • A2.3 Building the Test Code
  • A3. Running the Test Suite
  • A4. Cleanup
• Index

Abstract:

Although this appears as the abstract, it's actually a to-do list for the manual.

• beef up Chapter 1
• more on design (memory allocation)
• finish Quick Start section
• description of test suite output