The ELAPS Framework: Experimental Linear Algebra Performance Studies

TL;DR

ELAPS is an open-source framework that simplifies performance experimentation for dense linear algebra, enabling researchers and developers to analyze how various factors affect performance across diverse computing platforms.

Contribution

The paper introduces ELAPS, a versatile, multi-platform environment that streamlines the process of benchmarking and analyzing linear algebra performance with customizable experiments.

Findings

ELAPS supports experiments on laptops, clusters, accelerators, and supercomputers.

It provides detailed metrics and visual reports for performance analysis.

Demonstrated effectiveness in four real-world application scenarios.

Abstract

Optimal use of computing resources requires extensive coding, tuning and benchmarking. To boost developer productivity in these time consuming tasks, we introduce the Experimental Linear Algebra Performance Studies framework (ELAPS), a multi-platform open source environment for fast yet powerful performance experimentation with dense linear algebra kernels, algorithms, and libraries. ELAPS allows users to construct experiments to investigate how performance and efficiency vary depending on factors such as caching, algorithmic parameters, problem size, and parallelism. Experiments are designed either through Python scripts or a specialized GUI, and run on the whole spectrum of architectures, ranging from laptops to clusters, accelerators, and supercomputers. The resulting experiment reports provide various metrics and statistics that can be analyzed both numerically and visually. We demonstrate the use of ELAPS in four concrete application scenarios and in as many computing environments, illustrating its practical value in supporting critical performance decisions.