DocumentCode :
2226769
Title :
Thread-Sensitive Pointer Analysis for Inter-Thread Dataflow Detection
Author :
Qian, Ju ; Xu, Baowen
Author_Institution :
Sch. of Comput. Sci. & Eng., Southeast Univ., Nanjing
fYear :
2007
fDate :
21-23 March 2007
Firstpage :
157
Lastpage :
163
Abstract :
Inter-thread dataflows are of great importance in analyzing concurrent programs. Suffering from the complexity of currency, it is still difficult to identify them from Java programs precisely and efficiently. A fundamental problem in inter-thread dataflow analysis is pointer analysis. Currently, pointer analyses can be classified into context-sensitive ones and context-insensitive ones. They are either too time consuming or too imprecise and hence not suitable for scalable and high precision inter-thread dataflow analysis. To change the situation, this paper proposes a thread-sensitive pointer analysis algorithm which only uses sequences of thread starts as calling contexts. As in inter-thread dataflow analysis, the major objective is analyzing interthread dataflows not intra-thread ones, the coarse-grained context-sensitivity can not only effectively speed up the analysis, but also preserve the analyze precision as much as possible
Keywords :
Java; concurrency control; data flow analysis; software metrics; Java programs; concurrent program analysis; currency complexity; interthread dataflow analysis; interthread dataflow detection; thread-sensitive pointer analysis; Algorithm design and analysis; Change detection algorithms; Computer science; Concurrent computing; Data analysis; Data engineering; Information analysis; Java; Software quality; Yarn; call graph; concurrency; dataflow; inter-thread; pointer analysis;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Future Trends of Distributed Computing Systems, 2007. FTDCS '07. 11th IEEE International Workshop on
Conference_Location :
Sedona, AZ
ISSN :
1071-0483
Print_ISBN :
0-7695-2810-4
Type :
conf
DOI :
10.1109/FTDCS.2007.34
Filename :
4144626
Link To Document :
بازگشت