PGHPF Compiler User's Guide - Contents

• Table Of Contents
• Preface
  • Audience Description
  • Compatibility and Conformance to Standards
  • Organization
  • Hardware and Software Constraints
  • Conventions
  • Related Publications
• 1 Getting Started
  • 1.1 Overview
  • 1.2 Invoking PGHPF
    • 1.2.1 PGHPF Command-line Syntax
    • 1.2.2 Compiler Command-line Options
  • 1.3 Filename Conventions
    • 1.3.1 Input Files
    • 1.3.2 Output Files
• 2 PGHPF Compiler Options
  • -M pghpf_flag
• 3 PGHPF Runtime Options
  • 3.1 Runtime Options and Environment Variables
    • 3.1.1 Specify Command Line Options in the Environment
    • 3.1.2 RPM Runtime Library
      • Number of Processors -np
      • Specify Hosts -host
      • Specifying Alternate Hostname -curhost
      • Specify Debugging -g and -debugger
      • Specify Minimum Unbuffered Message Size -minxfer
      • Specify NFS Mount Path for Automounted Systems -mount
      • Specify Remote Shell -rsh
      • Print Execution Information
    • 3.1.3 Shared Memory Systems - RPM
      • Global Heap Size -heapz
    • 3.1.4 RPM1 Runtime Library
    • 3.1.5 PVM Runtime Library
      • PVM - Number of Processors -np
      • Print Execution Information
    • 3.1.6 MPI Runtime Library
      • Print Execution Information
    • • -
  • 3.2 Print Execution Information -stat
    • 3.2.1 CPU Statistics -stat
      • Memory Statistics
      • Message Statistics
      • All Statistics
      • -
• 4 PGHPF Libraries
  • 4.1 Libraries and Startup Routines
• 5 Data Distribution
  • 5.1 Run-time Model - Overview
  • 5.2 HPF Data Distribution
    • 5.2.1 Propagating Data Distribution
      • Default Distribution
      • Explicit HPF Distribution
      • Distributing Allocatable Arrays
      • Distribution of Procedure Arguments
      • Distribution of Compiler Created Temporaries
    • 5.2.2 Computation Partitioning
  • 5.3 InterProcedural Analysis (IPA) Overview
  • 5.4 Static Data Initialization
• 6 Runtime and Execution Model
  • 6.1 Communication Optimization
    • 6.1.1 Replicated Data
    • 6.1.2 Overlap Shift Communications
      • Overlap Shift Optimization
    • 6.1.3 Copy Section Communications
    • 6.1.4 Indirection References and Scheduling
  • 6.2 Program Performance Measurement
• 7 PGHPF Parallelism
  • 7.1 Array Assignment Parallelism
  • 7.2 WHERE Statement Parallelism
    • 7.2.1 WHERE Construct Parallelism
  • 7.3 FORALL Statement and Construct Parallelism
  • 7.4 Library Routines and Intrinsics
    • 7.4.1 3F Library Routines
  • 7.5 INDEPENDENT Loops
    • 7.5.1 Restrictions on Parallelization
    • 7.5.2 INDEPENDENT Clauses
      • The NEW Clause
      • The ON HOME Clause
      • The REDUCTION Clause
    • 7.5.3 Procedure Calling
    • 7.5.3 Independent Performance
  • 7.6 Inlining
    • 7.6.1 Switches
      • The inline Switch
      • The extract Switch
      • Compiler Information Switches
      • Extract Directories
      • Inlining Transformations
    • 7.6.2 Examples
• 8 Extrinsics
  • 8.1 Extrinsic F77_LOCAL Support
    • 8.1.1 Common Routines
      • Get number of processors
      • Get my processor number
      • Translate PGHPF processor number to processor grid coordinates
      • Translate processor grid coordinates to PGHPF processor number
    • 8.1.2 Generic routines
      • Send/receive non-character data
      • Send/receive Fortran character data
    • 8.1.3 MPI
      • Translate processor number to MPI processor identifier
    • 8.1.4 PVM
      • Translate processor number to PVM processor identifier
  • 8.2 Extrinsic HPF_LOCAL Support
  • 8.3 Extrinsic F77_SERIAL Support
  • 8.4 Extrinsic HPF_SERIAL Support
  • 8.5 Extrinsic HPF_CRAFT Support
• 9 Modules
  • 9.1 Modules with Generic Interfaces
  • 9.2 LIB3F Procedures
• A Messages
  • A.1 Diagnostic Messages
  • A.2 Compiler Error Messages
    • A.2.1 Message List
• B Driver Customization
  • B.1 Environment Variable
    • B.1.1 Setting Environment Variables
• Index