A Novel Speculative Multithreading Parallelization Method in Chip Multiprocessor Systems

Chip multiprocessors have become mainstream processors in recent years. In this paper, we propose a novel speculative multithreading parallelization method for parallelizing a sequential program into multiple threads. In our method, extended program dependence graph and topological sort are used to analyze code dependency. With the help of dynamic profiling, we exploit thread level parallelism in loops and the remaining code regions by analyzing control and data speculation in general purpose applications. Using an initial automatic compiler implementation and a validated processor model, this paper demonstrates significant gains using speculation for 4-core chip multiprocessor models running a variety of applications.