Mixed precision thin SVD algorithms based on the Gram matrix

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Carson, Erin, Ma, Yuxin, Shao, Meiyue
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908881986256896
author Carson, Erin
Ma, Yuxin
Shao, Meiyue
author_facet Carson, Erin
Ma, Yuxin
Shao, Meiyue
contents In this work, we present a mixed precision algorithm that leverages the Gram matrix and Jacobi methods to compute the singular value decomposition (SVD) of tall-and-skinny matrices. By constructing the Gram matrix in higher precision and coupling it with a Jacobi algorithm, our theoretical analysis and numerical experiments both indicate that the singular values computed by this mixed precision thin SVD algorithm attain high relative accuracy. In practice, our mixed precision thin SVD algorithm yields speedups of over 10x on a single CPU and about 2x on distributed memory systems when compared with traditional thin SVD methods.
format Preprint
id arxiv_https___arxiv_org_abs_2603_11953
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Mixed precision thin SVD algorithms based on the Gram matrix
Carson, Erin
Ma, Yuxin
Shao, Meiyue
Numerical Analysis
65F15, 65F25, 65G50, 65Y05, 65Y20
In this work, we present a mixed precision algorithm that leverages the Gram matrix and Jacobi methods to compute the singular value decomposition (SVD) of tall-and-skinny matrices. By constructing the Gram matrix in higher precision and coupling it with a Jacobi algorithm, our theoretical analysis and numerical experiments both indicate that the singular values computed by this mixed precision thin SVD algorithm attain high relative accuracy. In practice, our mixed precision thin SVD algorithm yields speedups of over 10x on a single CPU and about 2x on distributed memory systems when compared with traditional thin SVD methods.
title Mixed precision thin SVD algorithms based on the Gram matrix
topic Numerical Analysis
65F15, 65F25, 65G50, 65Y05, 65Y20
url https://arxiv.org/abs/2603.11953