1. Inicio
  2. Atrás
Título: A practical MHP information computation for concurrent Java programs /
Autores: Li, Lin, 1971-
Fecha: 2005
Publicador: McGill University - MCGILL
Fuente:
Tipo: Electronic Thesis or Dissertation
Tema: Computer Science.
Descripción: With the development of multi-processors, multi-threaded programs and programming languages have become more and more popular. This requires extending the scope of program analysis and compiler optimization from traditional, sequential programs to concurrent programs.
Naumovich et al. proposed May Happen in Parallel (MHP) analysis that determines which program statements may be executed concurrently. From this information, compiler optimization improvements, as well as analysis data on potential program problems such as dataraces can be analyzed or discovered.
Unfortunately, MHP analysis has some limitations with respect to practical use. In this thesis we present an implementation of MHP analysis for Java that attempts to address some of the practical implementation concerns of the original work. We describe a design that incorporates techniques for aiding a feasible implementation and expanding the range of acceptable inputs.
The MHP analysis requires a particular internal representation in order to run. By using a combination of techniques, we are able to compact that representation, and thus significantly improve MHP execution time without, affecting accuracy. We also provide experimental results showing the utility and impact of our approach and optimizations using a variety of concurrent benchmarks. The results show that our optimizations are effective, and allow more and larger benchmarks to be analyzed. For some benchmarks, our optimizations have very impressive results, speeding up MHP analysis by several orders of magnitude.
Idioma: en