Deadlock-free incremental replay of message-passing programs

被引：4
作者：
Zambonelli, F
Netzer, RHB
机构：
[1] Univ Modena, Dipartimento Sci Ingn, I-41100 Modena, Italy
[2] IP Performance Inc, N Scituate, RI 02857 USA
来源：
JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING | 2001年 / 61卷 / 05期
基金：
美国国家科学基金会;
关键词：
distributed debugging; checkpointing; incremental replay; adaptive logging; deadlock;
D O I：
10.1006/jpdc.2001.1703
中图分类号：
TP301 [理论、方法];
学科分类号：
081202 ;
摘要：
To support incremental replay of message-passing applications. processes must periodically checkpoint and the content of some messages must be logged, to break dependencies of the current slate of the execution on past events. This paper shows that known adaptive logging algorithms are likely to introduce deadlocks in replay, and we introduce a new algorithm that: (i) prevents deadlocks in replay and (ii) enables the tuning of its behavior to meet specific user needs. (C) 2001 Academic Press.
引用
收藏
页码：667 / 678
页数：12
相关论文
共 47 条
12345