Performance: Book Announcement (Thomas Fahringer)
3 Sep 1996 22:01:59 -0400

          From comp.compilers

Related articles
Performance: Book Announcement (1996-09-03)
| List of all articles for this month |

From: (Thomas Fahringer)
Newsgroups: comp.compilers
Date: 3 Sep 1996 22:01:59 -0400
Organization: Vienna University, Austria
Keywords: books, parallel, Fortran

            * Announcing my new book on *
            * Automatic Performance Prediction of Parallel Programs *

                                  -------> WEB site <>.

TITLE: Automatic Performance Prediction of Parallel Programs

AUTHOR: Thomas Fahringer, Ph.D.
Assistant Professor
Institute for Software Technology and Parallel Systems
                            University of Vienna, Austria

PUBLISHING COMPANY: Kluwer Academic Publishers, Boston, USA

ORDER FORM: contact author for details


This book describes a unified approach to the problem of automatically
estimating the performance of parallel computer programs. The author focuses
primarily on distributed memory multiprocessor systems, although large
portions of the analysis can be applied to shared memory architectures as
well. The author introduces a novel and very practical approach for
predicting some of the most important performance parameters of parallel
programs, including work distribution, number of transfers, amount of data
transferred, network contention, transfer time, computation time and number
of cache misses. This approach is based on advanced compiler analysis that
carefully examines loop iteration spaces, procedure calls, array subscript
expressions, communication patterns, data distributions and optimizing code
transformations at the program level; and the most important machine
specific parameters including cache characteristics, communication network
indices, and benchmark data for computational operations at the machine

The material has been fully implemented as part of P3T, which is an
integrated automatic performance estimator of the Vienna Fortran Compilation
System (VFCS), a state-of-the-art parallelizing compiler for Fortran77,
Vienna Fortran and a subset of High Performance Fortran (HPF) programs. A
large number of experiments using realistic HPF and Vienna Fortran code
examples demonstrate highly accurate performance estimates, and the ability
of the described performance prediction approach to successfully guide both
programmer and compiler in parallelizing and optimizing parallel programs.

A graphical user interface is described and displayed that visualizes each
program source line together with the corresponding parameter values. P3T
uses color-coded performance visualization to immediately identify hot spots
in the parallel program. Performance data can be filtered and displayed at
various levels of detail. Colors displayed by the graphical user interface
are visualized in greyscale.

Automatic Performance Prediction of Parallel Programs also includes coverage
of fundamental problems of automatic parallelization for distributed memory
multicomputers, a description of the basic parallelization strategy and a
large variety of optimizing code transformations as included under VFCS.



1 Introduction
    1.1 Automatic Parallelization for Multiprocessor Systems
    1.2 Motivation for Performance Prediction
    1.3 P3T: Parameter based Performance Prediction Tool
    1.4 Overview

2 Model
    2.1 Introduction
    2.2 Sequential Programs
    2.3 Parallel Programs
    2.4 Basic Parallelization Strategy
    2.5 Optimizing Compiler Transformations
    2.6 Using P3T and Weight Finder under VFCS
    2.7 Summary

3 Sequential Program Parameters
    3.1 Introduction
    3.2 Sequential Program Parameters
    3.3 Instrumentation
    3.4 Optimization
    3.5 Adaptation of Profile Data
    3.6 Summary

4 Parallel Program Parameters
    4.1 Introduction
    4.2 Work Distribution
    4.3 Number of Transfers
    4.4 Amount of Data Transferred
    4.5 Transfer Time
    4.6 Network Contention
    4.7 Number of Cache Misses
    4.8 Computation Time
    4.9 Summary

5 Experiments
    5.1 Introduction
    5.2 Implementation Status
    5.3 Estimation Accuracy of P3T
    5.4 Usefulness of P3T
    5.5 Graphical User Interface of P3T
    5.6 Summary

6 Related Work
    6.1 Performance Prediction Techniques
    6.2 Classification of Performance Estimators

7 Conclusions
    7.1 Contributions
    7.2 Future Research

A Appendix
    A.1 Intersection and Volume Computation of Convex n-dimensional Polytopes
    A.2 Notation




Kluwer Academic Publishers, Boston

Date of publishing: March 1996
296 pp.
ISBN: 0-7923-9708-8
NLG: 190.00
USD: 102.50
GBP: 72.75

Thomas Fahringer, Ph.D. e-mail:
Assistant Professor tel (office): +43 1 310 56 08 - 86
University of Vienna tel (sec): +43 1 310 56 08 - 71
Institute for Software Technology fax: +43 1 310 56 08 - 88
and Parallel Systems WWW:
Liechtensteinstr. 22
A-1090 Vienna, Austria

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.