Chavan, Y L
(2015)
Implementing Andersen Alias Analysis in LLVM.
Masters thesis, Indian Institute of Technology Hyderabad.
Abstract
Alias Analysis information is a prerequisite for most of the program analysis and the quality of this information can greatly boost up the precision and performance of the program. Many recent works in this area have focused on high precision and scalability, but still they lack an efficient and scalable alias analysis. The inclusion based or Andersen style alias analysis is widely used to compute such information. Andersen alias analysis is flow insensitive which make it highly precise and motivate many modern optimizing compilers to deploy it. In this work, we have improved and implemented the Andersen alias analysis in the latest version of LLVM.
Actions (login required)
|
View Item |