Managing interprocedural optimization
Managing interprocedural optimization
A Compiler Optimization Algorithm for Shared-Memory Multiprocessors
IEEE Transactions on Parallel and Distributed Systems
Loop Transformations for Restructuring Compilers: The Foundations
Loop Transformations for Restructuring Compilers: The Foundations
High Performance Compilers for Parallel Computing
High Performance Compilers for Parallel Computing
Partitioning and Labeling of Loops by Unimodular Transformations
IEEE Transactions on Parallel and Distributed Systems
Dependence Uniformization: A Loop Parallelization Technique
IEEE Transactions on Parallel and Distributed Systems
Hi-index | 0.00 |
Since most program execution time is consumed in loop structure, we propose a data dependency elimination method which extracts maximum parallelism from the sequential loop. This procedure automatically transforms nested loops to nested DOALL loops, thereby reduces data dependency effectively. Conventional studies for extracting the parallelism are focused mostly on a single procedure transition. But almost all programs have potential parallelisms among multiple procedures. Procedure call parallelisms, with only a single loop structure or procedure call, almost all are concerned with the uniform code within the uniform data dependency distance. In this study, we propose an algorithm which can be applied to uniform, nonuniform, and complex data dependency distance among multiple procedures. For the performance evaluation, we compare loop extraction, loop embedding, and procedure cloning using CRAY-T3E. The results show that the proposed algorithm is effective.