Fault-tolerant numerical iterative algorithms at scale
- 3 October 2025
- journal article
- Published by SAGE Publications in The International Journal of High Performance Computing Applications
- Vol. 40 (1) , 63-79
Abstract
This work investigates how to protect numerical iterative algorithms from all types of errors that can strike at scale: fail-stop errors (a.k.a. failures) and silent errors, striking both as computation errors and memory bit-flips. We combine various techniques: detectors for computation errors, checksums for memory errors, and checkpoint/restart for failures. The objective is to minimize the expected time per iteration of the algorithm. We design a hierarchical pattern that combines and interleaves all these fault-tolerance mechanisms, and we determine the optimal periodic pattern that achieves this objective. We instantiate these results for the performance analysis of the Preconditioned Conjugate Gradient (PCG) algorithm: we report several scenarios where the optimal pattern dramatically decreases the overhead due to error mitigation.Keywords
This publication has 41 references indexed in Scilit:
- Challenges for Mantle Convection Simulations at the Exa-Scale: Numerics, Algorithmics and Software2023
- Detection and correction of silent errors in the conjugate gradient algorithmNumerical Algorithms, 2022
- A Large-Scale Study of Failures on Petascale SupercomputersJournal of Computer Science and Technology, 2018
- Tutorial: Physics and modeling of Hall thrustersJournal of Applied Physics, 2017
- Coping with recall and precision of soft error detectorsJournal of Parallel and Distributed Computing, 2016
- A backward/forward recovery approach for the preconditioned conjugate gradient methodJournal of Computational Science, 2016
- Abstract robust coarse spaces for systems of PDEs via generalized eigenproblems in the overlapsNumerische Mathematik, 2013
- Algorithm-based fault tolerance applied to high performance computingJournal of Parallel and Distributed Computing, 2009
- A higher order estimate of the optimum checkpoint interval for restart dumpsFuture Generation Computer Systems, 2006
- Hydrostatic, quasi‐hydrostatic, and nonhydrostatic ocean modelingJournal of Geophysical Research: Oceans, 1997