Mixed-Precision Orthogonalization-Free Projection Methods for Eigenvalue and Singular Value Problems
SIAM Journal on Scientific Computing · 30.06.2026 · 48 (3) · B544-B568
DOI: 10.1137/25m1756922 ↗Kurz gefasst
Fachartikel in der Zeitschrift SIAM Journal on Scientific Computing, online erschienen am 19.06.2026. Verfasst von Tianshi Xu und 4 weiteren Autor:innen. Die Arbeit nennt 31 Literaturangaben. Gefördert von: National Science Foundation.
Abstract (Originalsprache)
Mixed-precision arithmetic offers significant computational advantages for large-scale matrix computation tasks, yet preserving accuracy and stability in eigenvalue problems and the singular value decomposition remains challenging. This paper introduces an approach that eliminates orthogonalization requirements in traditional Rayleigh–Ritz projection methods. The proposed method employs nonorthogonal bases computed at reduced precision, resulting in bases computed without inner products. A primary focus is on maintaining the linear independence of the basis vectors. Through extensive evaluation with both synthetic test cases and real-world applications, we demonstrate that the proposed approach achieves the desired accuracy while fully taking advantage of mixed-precision arithmetic.
Förderung
- National Science Foundation
Zitieren
APA
Xu, T., Zhang, Z., Chen, J., Saad, Y., & Xi, Y. (2026). Mixed-Precision Orthogonalization-Free Projection Methods for Eigenvalue and Singular Value Problems. SIAM Journal on Scientific Computing, 48(3), B544-B568. https://doi.org/10.1137/25m1756922
Vancouver
Xu T, Zhang Z, Chen J, Saad Y, Xi Y. Mixed-Precision Orthogonalization-Free Projection Methods for Eigenvalue and Singular Value Problems. SIAM Journal on Scientific Computing. 2026;48(3):B544-B568. doi:10.1137/25m1756922
BibTeX
@article{Xu2026,
author = {Xu, Tianshi and Zhang, Zechen and Chen, Jie and Saad, Yousef and Xi, Yuanzhe},
title = {Mixed-Precision Orthogonalization-Free Projection Methods for Eigenvalue and Singular Value Problems},
journal = {SIAM Journal on Scientific Computing},
year = {2026},
volume = {48},
number = {3},
pages = {B544--B568},
doi = {10.1137/25m1756922}
}
EUVIMED