Reduced and Mixed-Precision Block Householder QR Factorization
Abstract
Recent advances in hardware architecture have highlighted the importance of reduced-precision arithmetic for improving performance and reducing energy consumption in large-scale scientific applications. However, uniformly low-precision methods can suffer from numerical instability, especially in fundamental linear algebra routines. Mixed-precision strategies therefore play an important role in balancing efficiency, stability, and accuracy. In this paper, we focus on the block Householder QR decomposition using the WY representation and study its behavior in reduced- and mixed-precision settings. Starting from a uniform half-precision variant, we show that performing norm computations with fp32 accumulation is necessary to avoid overflow and underflow on large matrices. Based on this observation, we introduce mpWYQR, a mixed-precision algorithm that increases the precision of only selected computations and avoids most of the precision conversions required by the existing mixed-precision blocked QR algorithm (mpBQR3) introduced in [L. M. Yang, A. Fox, and G. Sanders, SIAM J. Sci. Comp., 43.3 (2021), pp. A1723--A1753]. Our analysis shows that mpWYQR accuracy is governed by low-precision unit roundoff, and that its error bound exceeds that of mpBQR3 by only a constant factor, with identical heuristic bounds. In experiments across a range of matrix sizes and condition numbers, we show that mpWYQR retains mpBQR3 accuracy while running up to 18× faster and using less memory. We also discuss the performance cost of computing inner products in higher precision, and the library support needed for an efficient GPU implementation.
Related
- Right Precision, Right Place: Runtime Adaptations for Mixed-Precision Cholesky Factorization
- Adaptive accuracy-driven splitting in Ozaki matrix multiplication and application to block LU factorization
- Piñeiro mixed orthogonal polynomials and bidiagonal factorization
- Finite-Precision Symmetric Krylov Methods: Exact Rounding Examples, Block Paige Identities,and a Variable-Block Lanczos Model
- A GPU-Accelerated Blocked Adaptive Randomized Range Finder Based on an Implicit Householder QR Decomposition
BibTeX
@misc{benner2026reduced,
title = {{Reduced and Mixed-Precision Block Householder QR Factorization}},
author = {Peter Benner and Martin Köhler and Xiaobo Liu and Eda Oktay},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.38537},
eprint = {2610.38537},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.38537},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Reduced and Mixed-Precision Block Householder QR Factorization
AU - Peter Benner
AU - Martin Köhler
AU - Xiaobo Liu
AU - Eda Oktay
PY - 2026
DA - 2026/10/28
PB - arXiv
DO - 10.48550/arxiv.2610.38537
UR - https://arxiv.org/abs/2610.38537
AB - Recent advances in hardware architecture have highlighted the importance of reduced-precision arithmetic for improving performance and reducing energy consumption in large-scale scientific applications. However, uniformly low-precision methods can suffer from numerical instability, especially in fundamental linear algebra routines. Mixed-precision strategies therefore play an important role in balancing efficiency, stability, and accuracy. In this paper, we focus on the block Householder QR decomposition using the WY representation and study its behavior in reduced- and mixed-precision settings. Starting from a uniform half-precision variant, we show that performing norm computations with fp32 accumulation is necessary to avoid overflow and underflow on large matrices. Based on this observation, we introduce mpWYQR, a mixed-precision algorithm that increases the precision of only selected computations and avoids most of the precision conversions required by the existing mixed-precision blocked QR algorithm (mpBQR3) introduced in [L. M. Yang, A. Fox, and G. Sanders, SIAM J. Sci. Comp., 43.3 (2021), pp. A1723--A1753]. Our analysis shows that mpWYQR accuracy is governed by low-precision unit roundoff, and that its error bound exceeds that of mpBQR3 by only a constant factor, with identical heuristic bounds. In experiments across a range of matrix sizes and condition numbers, we show that mpWYQR retains mpBQR3 accuracy while running up to 18× faster and using less memory. We also discuss the performance cost of computing inner products in higher precision, and the library support needed for an efficient GPU implementation.
ER -