Numerical Reproducibility and Parallel Computations: Issues for Interval Algorithms

DOI10.1109/TC.2014.2322593zbMATH Open1364.65108arXiv1312.3300MaRDI QIDQ5268095FDOQ5268095

Publication date: 20 June 2017

Published in: IEEE Transactions on Computers (Search for Journal in Brave)

Abstract: What is called "numerical reproducibility" is the problem of getting the same result when the scientific computation is run several times, either on the same machine or on different machines, with different types and numbers of processing units, execution environments, computational loads etc. This problem is especially stringent for HPC numerical simulations. In what follows, the focus is on parallel implementations of interval arithmetic using floating-point arithmetic. For interval computations, numerical reproducibility is of course an issue for testing and debugging purposes. However, as long as the computed result encloses the exact and unknown result, the inclusion property, which is the main property of interval arithmetic, is satisfied and getting bit for bit identical results may not be crucial. Still, implementation issues may invalidate the inclusion property. Several ways to preserve the inclusion property are presented, on the example of the product of matrices with interval coefficients.

Full work available at URL: https://arxiv.org/abs/1312.3300

Mathematics Subject Classification ID

Parallel numerical computation (65Y05) Interval and finite arithmetic (65G30) Numerical algorithms for computer arithmetic, etc. (65Y04)

Cited In (2)

This page was built for publication: Numerical Reproducibility and Parallel Computations: Issues for Interval Algorithms

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5268095)