A comparison of two performance portability metrics
Author:
- 4 August 2023
- journal article
- research article
- Published by Wiley in Concurrency and Computation: Practice and Experience
- Vol. 35 (25)
Abstract
Summary: The rise in the demand for new performance portability frameworks for heterogeneous computing systems has brought with it a number of proposals of workable metrics for evaluating the performance portability of applications. The aim of this article is twofold. First, we analyze the underlying principles of the criteria and definition of the revised Ⴔ metric and show that these principles are partially correct and suffer from a lack of a solid and clear performance portability model. We prove mathematically and demonstrate it practically that the principles are only correct for the architectural efficiency approach based on throughputs but are incorrect for the popular application efficiency approach based on run‐times. Second, we are examining whether the Ⴔ and metrics meet the requirements of consistency, proportionality, and lossless information. We use examples from the scientific literature to show the reader that the Ⴔ metric loses information while the metric and other metrics do not lose information.Keywords
This publication has 22 references indexed in Scilit:
- Implications of a metric for performance portabilityFuture Generation Computer Systems, 2019
- Performance Portability Challenges for Fortran Applications2018
- An Empirical Roofline Methodology for Quantitatively Assessing Performance Portability2018
- Heterogeneous CPU-GPU Execution of Stencil Applications2018
- A Beginner’s Guide to Estimating and Improving Performance Portability2018
- GPU-STREAM v2.0: Benchmarking the Achievable Memory Bandwidth of Many-Core Processors Across Diverse Parallel Programming ModelsLecture Notes in Computer Science, 2016
- RooflineCommunications of the ACM, 2009
- Inference by Eye: Confidence Intervals and How to Read Pictures of Data.American Psychologist, 2005
- The Graphical Presentation of a Collection of MeansJournal of the Royal Statistical Society Series A: Statistics in Society, 1995
- Characterizing computer performance with a single numberCommunications of the ACM, 1988