Automated Reasoning Support for Invasive Interactive Parallelization

摘要: To parallelize a sequential source code, parallelization strategy must be defined that transforms the code into an equivalent parallel version. Since parallelizing compilers can sometimes transform loops and other well-structured codes ones automatically, we are interested in finding solution to semi-automatically not able mostly because of weakness classical data control dependence analysis, order simplify process transforming for programmers.Invasive Interactive Parallelization (IIP) hypothesizes by using anintelligent system guides user through interactive one boost above direction. The intelligent system's guidance relies on analysis pre-defined transformation sequences. support its main hypothesis, IIP suggests encode sequences terms strategies dictate default ways various patterns facts which have been obtained both from directly user.In this project, investigate how automated reasoning supportthe method with acceptable performance but faster than manual parallelization. We looked at two special problem areas: Divide conquer algorithms codes. Our focus is four legacy C programs such as: Quick sort, Merge Jacobi Matrix multipliation summation OpenMP MPI environment developing assistance tool provides users assistanceneeded code.