A framework for trackable and modular rounding error analysis

DOI10.48550/ARXIV.2610.35420MaRDI QIDQ7408789arXivFDO

Authors Erin Carson, Meiyue Shao, Yuxin Ma

Publication date 26 October 2026

Published in arXiv

Tldr We propose a trackable and modular rounding error analysis for modern numerical computations with evolving arithmetic formats, precisions, and computational kernels, introducing a new notation that handles division errors without restriction on the number of iterations.

Full work available at URL https://doi.org/10.48550/ARXIV.2610.35420

arXiv classification

math.NA

Cites work

• A modular framework for the backward error analysis of GMRES
• Accuracy and Stability of Numerical Algorithms
• Error-free transformations of matrix multiplication by using fast routines of matrix multiplication and its applications
• Householder Orthogonalization with a Nonstandard Inner Product
• Improved backward error bounds for LU and Cholesky factorizations
• Mixed precision algorithms in numerical linear algebra
• Mixed precision block fused multiply-add: error analysis and application to GPU tensor cores
• On the backward stability of s-step GMRES
• On the numerical stability of sketched GMRES
• Pracniques: further remarks on reducing truncation errors
• Relative-error bounds for the LU decomposition via the GTH algorithm
• Rounding Errors in Algebraic Processes
• Simulating Low Precision Floating-Point Arithmetic

This page was built for publication: A framework for trackable and modular rounding error analysis

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7408789)