A GPU-based hyperbolic SVD algorithm

TL;DR

This paper presents a GPU-accelerated Jacobi HSVD algorithm that significantly improves speed and accuracy for hyperbolic SVD computations, with potential for hybrid CPU-GPU parallelism.

Contribution

It introduces a novel GPU-based Jacobi HSVD algorithm that enhances performance and accuracy over existing methods, and discusses hybrid parallelism strategies.

Findings

Significant speedup over sequential and MPI-parallelized algorithms

Improved accuracy in hyperbolic SVD computations

Potential for hybrid CPU-GPU parallelism

Abstract

A one-sided Jacobi hyperbolic singular value decomposition (HSVD) algorithm, using a massively parallel graphics processing unit (GPU), is developed. The algorithm also serves as the final stage of solving a symmetric indefinite eigenvalue problem. Numerical testing demonstrates the gains in speed and accuracy over sequential and MPI-parallelized variants of similar Jacobi-type HSVD algorithms. Finally, possibilities of hybrid CPU--GPU parallelism are discussed.