Improving Energy Saving of One-sided Matrix Decompositions on CPU-GPU Heterogeneous Systems

TL;DR

This paper introduces BSR, a framework combining fault tolerance overclocking and DVFS to significantly reduce energy consumption in matrix decompositions on CPU-GPU systems while maintaining performance.

Contribution

The paper proposes BSR, an energy-saving framework that integrates ABFT-OC overclocking and DVFS for efficient matrix decompositions on heterogeneous systems.

Findings

BSR saves up to 11.7% more energy than existing methods.

BSR achieves up to 14.1% Energy * Delay^2 reduction.

BSR enables performance improvements up to 1.43x without extra energy cost.

Abstract

One-sided dense matrix decompositions (e.g., Cholesky, LU, and QR) are the key components in scientific computing in many different fields. Although their design has been highly optimized for modern processors, they still consume a considerable amount of energy. As CPU-GPU heterogeneous systems are commonly used for matrix decompositions, in this work, we aim to further improve the energy saving of one-sided matrix decompositions on CPU-GPU heterogeneous systems. We first build an Algorithm-Based Fault Tolerance protected overclocking technique (ABFT-OC) to enable us to exploit reliable overclocking for key matrix decomposition operations. Then, we design an energy-saving matrix decomposition framework, Bi-directional Slack Reclamation(BSR), that can intelligently combine the capability provided by ABFT-OC and DVFS to maximize energy saving and maintain performance and reliability. Experiments show that BSR is able to save up to 11.7% more energy compared with the current best energy saving optimization approach with no performance degradation and up to 14.1% Energy * Delay^2 reduction. Also, BSR enables the Pareto efficient performance-energy trade-off, which is able to provide up to 1.43x performance improvement without costing extra energy.