PerfDMF

University of Oregon

Copyright  2005 University of Oregon Performance Research Lab

-------------------------------------------------------------------------------

Table of Contents

1. Introduction

    Prerequisites
    Installation

2. Using PerfDMF

    perfdmf_createapp
    perfdmf_createapp
    perfdmf_loadtrial

Chapter1.Introduction

Table of Contents

Prerequisites
Installation

PerfDMF (Performance Data Management Framework) is a an API/Toolkit that sits
atop a DBMS to manage and anaylize performance data. The API is available in
its native Java form as well as C.

Prerequisites

 1. A supported database (currently, PostgreSQL, MySQL, or Oracle).

 2. Java 1.4.

Installation

The PerfDMF utilities and applications are installed as part of the standard
TAU release. Shell scripts are installed in the TAU bin directory to configure
and run the utilities. It is assumed that the user has installed TAU and run
TAU's configre and 'make install'.

 1. Create a database. This step will depend on the user's chosen database.

      * PostgreSQL:

        $ createdb -O perfdmf perfdmf

        Or, from psql

        psql=# create database perfdmf with owner = perfdmf;

      * MySQL: From the MySQL prompt

        mysql> create database perfdmf;

      * Oracle: It is recommended that you create a tablespace for perfdmf:

        create tablespace perfdmf
        datafile '/path/to/somewhere' size 500m reuse;

        Then, create a user that has this tablespace as default:

        create user amorris identified by db;
        grant create session to amorris;
        grant create table to amorris;
        grant create sequence to amorris;
        grant create trigger to amorris;
        alter user amorris quota unlimited on perfdmf;
        alter user amorris default tablespace perfdmf;

        PerfDMF is set up to use the Oracle Thin Java driver. You will have to
        obtain this jar file for your database. In our case, it was ojdbc14.jar

 2. Configure PerfDMF. To configure PerfDMF, run the perfdmf_configure from the
    TAU bin directory.

    The configuration program will prompt the user for several values. The
    default values will work for most users. When configuration is complete, it
    will connect to the database and test the configuration. If the
    configuration is valid and the schema is not found (as will be the case on
    initial configuration), the schema will be uploaded. Be sure to specify the
    correct schema for your database.

Chapter2.Using PerfDMF

Table of Contents

perfdmf_createapp
perfdmf_createapp
perfdmf_loadtrial

The easiest way to interact with PerfDMF is to use ParaProf which provides a
GUI interface to all of the database information. In addition, the following
commandline utilities are provided.

perfdmf_createapp

This utility creates applications with a given name

$ perfdmf_createapp -n "New Application"
Created Application, ID: 24

perfdmf_createapp

This utility creates experiments with a given name, under a specified
application

$ perfdmf_createexp -a 24 -n "New Experiment"
Created Experiment, ID: 38

perfdmf_loadtrial

This utility uploads a trial to the database with a given name, under a
specified experiment

Usage: perfdmf_loadtrial -e <experiment id> -n <name>
                                     [options] <files>

Required Arguments:

  -e, --experimentid <number>    Specify associated experiment
                                       ID for the trial
  -n, --name <text>              Specify the name of the trial

Optional Arguments:

  -f, --filetype <filetype>   Specify type of performance data,
                               options are: profiles (default), pprof,
                               dynaprof, mpip, gprof, psrun, hpm,
                               packed, cube, hpc
  -t, --trialid <number>         Specify trial ID
  -i, --fixnames                 Use the fixnames option for gprof

Notes:
  For the TAU profiles type, you can specify either a specific set of
  profile files on the commandline, or you can specify a directory
  (by default the current directory).  The specified directory will
  be searched for profile.*.*.* files, or, in the case of
  multiple counters, directories named MULTI_* containing profile data.

Examples:

  perfdmf_loadtrial -e 12 -n "Batch 001"
    This will load profile.* (or multiple counters directories MULTI_*)
    into experiment 12 and give the trial the name "Batch 001"

  perfdmf_loadtrial -e 12 -n "HPM data 01" perfhpm*
    This will load perfhpm* files of type HPMToolkit into experiment
    12 and give the trial the name "HPM data 01"


