EigenG

Overview

EigenG is a port of the eigenvalue library EigenExa in the form of running on a single GPU in the CUDA environment. It is based on the traditional Householder tridiagonalization, divide-and-conquer method, and back-transform. Although it had not been upgraded since the publication at PPAM2013, part of the kernel, including the Batched version, has been updated in FY2022 considering the advent of the H100 processor, which follows the A100.

Downloads

• EigenG version 3.1 (tar.gz, 66KB) (October 25, 2022)

Publications

• Toshiyuki Imamura, Susumu Yamada, and Masahiko Machida: “Eigen-G: GPU-based eigenvalue solver for real-symmetric dense matrices”, Proceedings of PPAM2013, Lecture Notes in Computer Science (LNCS), Vol.8384, p.673-682, doi:10.1007/978-3-642-55224-3, 2014.