CAJ | 학술논문

针对当前最优秀的基于包含的指针分析在线优化技术即横向传播和纵向传播方法的不足之处，提出一个改进的算法。首先简要介绍基于包含的指针分析方法及其在线优化技术；其次描述横向传播和纵向传播方法，进而说明它们存在的一些缺陷；然后针对这些缺陷提出一个改进的算法，该算法能够在不损失时间开销的条件下以更小的内存开销实现横向传播，能在纵向传播过程中检测出约束图中所有的强连通分量；最后通过一个实例分析及使用CIL工具进行的实验验证了该改进算法的优越性。
침대당전최우수적기우포함적지침분석재선우화기술즉횡향전파화종향전파방법적불족지처，제출일개개진적산법。수선간요개소기우포함적지침분석방법급기재선우화기술；기차묘술횡향전파화종향전파방법，진이설명타문존재적일사결함；연후침대저사결함제출일개개진적산법，해산법능구재불손실시간개소적조건하이경소적내존개소실현횡향전파，능재종향전파과정중검측출약속도중소유적강련통분량；최후통과일개실례분석급사용CIL공구진행적실험험증료해개진산법적우월성。
In view of the deficiency of the state-of-the-art technique of inclusion-based online pointer analysis optimisation technology, i.e.the transverse propagation and longitudinal propagation algorithm,we propose an improved algorithm.First we briefly introduce the inclusion-based pointer analysis technology and its online optimisation technology;and then we describe the transverse propagation and longitudinal propagation algorithm and further,some of their drawbacks.After that,according to these defects we propose an improved algorithm,which is able to achieve transverse propagation with smaller memory overhead and almost equivalent time cost to the original algorithm,and can detect all the strongly connected components in constraint graph detected in the process of longitudinal propagation. Finally we verify the superiority of the improved algorithm via an example analysis and the experiment using CIL tool.