首页> 外文会议>International conference on computational science and its applications >Detection of Probabilistic Dangling References in Multi-core Programs Using Proof-Supported Tools
【24h】

Detection of Probabilistic Dangling References in Multi-core Programs Using Proof-Supported Tools

机译:使用校验工具检测多核计划中的概率悬垂引用

获取原文

摘要

This paper presents a new technique for detection of probabilistic dangling references in multi-core programs. The technique has the form of a simply structured type system and provides a suitable framework for proof-carrying code applications like mobile code applications that have limited resources. The type derivation of each individual analysis serves as a proof for the correctness of the analysis. The type system is designed to analyze parallel programs with structured concurrent constructs: fork-join constructs, conditionally spawned cores, and parallel loops. For a given program S, a probabilistic threshold p_(ms), and a probabilistic reference analysis for S, if S is well-typed in our proposed type system then all computational paths with probabilities greater than or equal to p_(ms) will contain no dangling pointers at run time. The soundness of the presented type system is proved in this paper with respect to a probabilistic operational semantics to our model language.
机译:本文介绍了一种用于检测多核计划中概率悬空引用的新技术。该技术具有简单的结构化系统的形式,并为携带有限资源的移动代码应用程序提供适当的携带代码应用程序的合适框架。每个单独分析的类型导出用作分析正确性的证据。类型系统旨在分析具有结构化并发构造的并行程序:FORK-JOIN构造,条件生成的核和并行环路。对于给定的程序S,概率阈值P_(MS)和S的概率参考分析,如果S在我们所提出的类型系统中符合S良好的类型,那么概率大于或等于P_(MS)的所有计算路径都将包含在运行时没有悬垂的指针。本文以概率的操作语义对我们的模型语言证明了所绘制类型系统的声音。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号