Tutorial on Performance Evaluation and Benchmarking

10/16/97


Click here to start


Table of Contents

Tutorial on Performance Evaluation and Benchmarking

Outline

Outline

Tutorial Goals

Why do we benchmark?

Application Performance

Performance Prediction

System Optimization

Goals of Performance Modeling

Goals of Performance Modeling

Terms

Execution Time

Execution Time (continued)

Execution Time (Continued)

Communication Speed and Granularity

Theoretical Peak Performance

Why Never Obtained

Whatís Wrong With Speedup T1/Tp ?

Speedup

PPT Slide

Strassenís Algorithm for Matrix Multiply

Strassenís Algorithm

Strassenís Algorithm

Strassenís Algorithm

Speciality/Instability Ratio

Communication Performance

Communication Performance

Communication Latency

Log(P) Communication Parameters

Log(P) Communication Parameters

Caveats

Discussion of and

Hockney Parameter

Machine Architecture

Amdahlís Law

Message Passing Factors

Shared Memory Factors

Simple Performance Model

Benchmarking Methods

Current Status of Performance Modeling

Problems with Existing Benchmarks

Problems with Existing Benchmarks

Vision

Early Benchmark Efforts

Performance Evaluation

Performance Evaluation

Algorithms and Architecture

Algorithm Issues

Blocking

Loop Unrolling

Limitations and Pitfalls

Future Development

The NAS Parallel Benchmarks (NPB)

Linpack Benchmark

Linpack Benchmark

TOP 500

PPT Slide

Single numbers

The Perfect Club

SPEC Benchmark

The NAS Parallel Benchmarks

SPEC Suite 2 -- (SPEC95)

SPEC Organization

SPEC HPC -High Performance Group

Livermore Loops

Euroben

Genesis

RAPS Benchmarks

URLs

Author: Jack J Dongarra

Email: dongarra@cs.utk.edu

Home Page: http://www.netlib.org/utk/people/JackDongarra/