IBM Books

Operation and Use, Volume 1 Using the Parallel Operating Environment

Contents

Tables

About this book

  • Who should read this book
  • How this book is organized
  • Overview of contents
  • Conventions and terminology used in this book
  • Abbreviated names
  • How to send your comments
  • National language support (NLS)
  • Setting POE environment variables
  • What's new in Parallel Environment 3.2?
  • New PE Benchmarker tools
  • Improved parallel checkpointing capabilities
  • MPI enhancements
  • DPCL is now an open source offering
  • Removal of pedb debugger support
  • Removal of VT trace collection support
  • Commands no longer supported

    • Chapter 1. Introduction

  • PE Version 3.2 Migration information
  • AIX compatibility
  • Parallel tracing removed
  • Checkpoint/Restart function
  • pedb removed
  • 64-bit application support
  • Fortran application support

    • Chapter 2. Executing parallel programs

  • Executing parallel programs using POE
  • Step 1: Compile the program
  • Step 2: Copy files to individual nodes
  • Step 3: Set up the execution environment
  • Step 4: Start the program marker array analysis tool
  • Step 5: Invoke the executable
  • Controlling program execution
  • Specifying develop mode
  • Making POE wait for processor nodes
  • Making POE ignore arguments
  • Managing standard input, output, and error
  • Determining which nodes will participate in parallel I/O
  • Checkpointing and restarting programs
  • POE user authorization
  • Supported security methods
  • Using DCE user authorization
  • Using AIX user authorization
  • Running POE with MALLOCDEBUG

    • Chapter 3. Managing POE jobs

  • Multi-task core file
  • Specifying the format of corefiles or suppressing corefile generation
  • Generating standard AIX corefiles
  • Writing corefile information to standard error
  • Generating lightweight corefiles
  • Stopping a POE job
  • Cancelling and killing a POE job
  • Detecting remote node failures
  • Considerations for using the SP switch and SP switch 2
  • Scenarios for allocating nodes with LoadLeveler
  • Submitting a batch POE job using IBM LoadLeveler
  • Submitting an interactive POE job using an IBM LoadLeveler command file
  • Generating an output LoadLeveler job command file
  • Running programs under the C shell
  • Using MP_CSS_INTERRUPT
  • Support for performance improvements
  • Interrupt mode control
  • Rcvncall improvements
  • Parallel file copy utilities

    • Chapter 4. Monitoring program execution using the Program Marker Array

  • Step 1: Call PM Array parallel utility functions
  • Step 2: Compile the program
  • Step 3: Set the number of lights
  • Step 4: Open the PM array window
  • Step 5: Invoke the program and monitor its execution
  • Displaying details of a light
  • Displaying task output
  • Step 6: Close the PM Array window

    • Appendix A. Parallel environment commands

  • mcp
  • mcpgath
  • mcpscat
  • mpamddir
  • mpcc
  • mpcc_r
  • mpCC
  • mpCC_r
  • mpiexec
  • mpxlf
  • mpxlf_r
  • mpxlf90
  • mpxlf90_r
  • mpxlf95
  • mpxlf95_r
  • pmarray
  • poe
  • poeckpt
  • poekill
  • poerestart

    • Appendix B. POE environment variables and command-line flags

    Notices

  • Trademarks
  • Acknowledgments

    • Glossary

    Bibliography

  • Information formats
  • Finding documentation on the World Wide Web
  • Accessing PE documentation online
  • RS/6000 SP publications
  • SP planning publications
  • SP software publications
  • AIX publications
  • DCE publications
  • Red books
  • Non-IBM publications

    • Index

    [ Top of Page | Previous Page | Next Page | Table of Contents | Index ]