Mixed-Precision SVD on GPUs via Ogita–Aishima Iterative Refinement | Research

Research Labs

All Research Labs

Spatial Intelligence

Applied Research

Autonomous Vehicles

Deep Imagination

Publications

AI Playground

New and Featured

AI Art Gallery

NGC Demos

Research Areas

AI & Machine Learning

3D Deep Learning

Computer Vision

Robotics

All Areas

Careers

Academic Collaborations

Government Collaborations

Graduate Fellowship

Internships

Research Openings

Research Scientists

Meet the Team

Licensing

Skip to main content

Search

Search

Enter the terms you wish to search for.

Publications

Mixed-Precision SVD on GPUs via Ogita–Aishima Iterative Refinement

Mixed-Precision SVD on GPUs via Ogita–Aishima Iterative Refinement

The Ogita–Aishima iterative refinement algorithm for SVD is based on matrix–matrix multiplication, making it an attractive approach for GPUs. We extend the algorithm to complex arithmetic, and introduce robust cluster handling based on Wedin’s perturbation theory and a per-pair version of Ogita–Aishima’s convergence analysis. Our implementation refines an FP32-quality SVD computed with cuSolver’s cusolverDnXgesvdp to an FP64-quality SVD. We validate the correctness of our mixed-precision implementation by testing on a variety of singular value distributions. In all cases the accuracy is comparable to the FP64 SVD routines [D,Z]GESVD. We observe speedups of up to 5.2× over FP64-quality cusolverDnXgesvdp on the RTX PRO 6000 Blackwell Workstation Edition GPU.

Authors

Angelika Schwarz (NVIDIA)

Rasmus Munk Larsen (NVIDIA)

Samuel Rodriguez (NVIDIA)

Christopher Baker (NVIDIA)

Lev Kruglyak (NVIDIA)

Publication Date

Monday, June 29, 2026

Research Area

Algorithms and Numerical Methods

High Performance Computing

Uploaded Files

Mixed-Precision SVD on GPUs via Ogita–Aishima Iterative Refinement.pdf479.61 KB