Coping with recall and precision of soft error detectors
详细信息 查看全文
- 作者：Leonardo Bautista-GomezaAuthor Vitae ; Anne BenoitbAuthor Vitae ; Auré ; lien CavelanbAuthor Vitae ; Saurabh K. RainacAuthor Vitae ; Yves Robertb ; dAuthor Vitae ; Hongyang Sunb ; hongyang.sun@ens-lyon.fr" class="auth_mail" title="E-mail the corresponding authorAuthor Vitae
- 关键词：Fault tolerance ; High-performance computing ; Silent data corruption ; Partial verification ; Recall and precision ; Exascale
- 刊名：Journal of Parallel and Distributed Computing
- 出版年：2016
- 出版时间：December 2016
- 年：2016
- 卷：98
- 期：Complete
- 页码：8-24
- 全文大小：1004 K
文摘
- class="listitem" id="list_l000005">
- class="label">•
Resilience algorithms to cope with silent errors for HPC applications.
- class="label">•
Characterization of optimal patterns using partial error detectors.
- class="label">•
Imprecise detectors offer limited usefulness.
- class="label">•
Optimization problem is NP-complete with multiple detector types.
- class="label">•
Construction of an FPTAS and a greedy approximation algorithm.