Expression Templates Revisited: A Performance Analysis of Current Methodologies

Authors:
Klaus Iglberger;Georg Hager;Jan Treibig;Ulrich Rüde
Affiliations:
klaus.iglberger@zisc.uni-erlangen.de;georg.hager@rrze.uni-erlangen.de and jan.treibig@rrze.uni-erlangen.de;-;ulrich.ruede@informatik.uni-erlangen.de
Venue:
SIAM Journal on Scientific Computing
Year:
2012

Citing 8
Cited 2

Curiously recurring template patterns

C++ Report
Expression templates

C++ gems
Object-oriented software construction (2nd ed.)

Object-oriented software construction (2nd ed.)
C++ Templates

C++ Templates
C++ Expression Templates Performance Issues in Scientific Computing

IPPS '98 Proceedings of the 12th. International Parallel Processing Symposium on International Parallel Processing Symposium
Advanced expression templates programming

Computing and Visualization in Science
Introduction to High Performance Computing for Scientists and Engineers

Introduction to High Performance Computing for Scientists and Engineers
LIKWID: A Lightweight Performance-Oriented Tool Suite for x86 Multicore Environments

ICPPW '10 Proceedings of the 2010 39th International Conference on Parallel Processing Workshops

Performance patterns and hardware metrics on modern multicore processors: best practices for performance engineering

Euro-Par'12 Proceedings of the 18th international conference on Parallel processing workshops
Automating embedded analysis capabilities and managing software complexity in multiphysics simulation, Part I: Template-based generic programming

Scientific Programming - A New Overview of the Trilinos Project --Part 1

Quantified Score

Hi-index	0.00

Visualization

Abstract

In the last decade, expression templates (ETs) have gained a reputation as an efficient performance optimization tool for C++ codes. This reputation builds on several ET-based linear algebra frameworks focused on combining both elegant and high-performance C++ code. However, on closer examination the assumption that ETs are a performance optimization technique cannot be maintained. In this paper we compare the performance of several generations of ET-based frameworks. We analyze different ET methodologies and explain the inability of some ET implementations to deliver high performance for dense and sparse linear algebra operations. Additionally, we introduce the notion of “smart” ETs, which truly allow for a combination of high performance code with the elegance and maintainability of a domain-specific language.