Genjiang Zhu, Xie Li, Zhongxiu Sun
Abstract
This paper presents a path-based method for identifying parallelism in C++ programs and proposes an execution model which trades off the efficiency and parallelism. Based on the path information provided at compiler time or computed at run time, the model reduces the rollbacks while enhances the parallelism.
OpenAlex reports 5 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
This paper presents a path-based method for identifying parallelism in C++ programs and proposes an execution model which trades off the efficiency and parallelism. Based on the path information provided at compiler time or computed at run time, the model reduces the rollbacks while enhances the parallelism.
Key concepts: Computer science, Parallelism (grammar), Compiler, Parallel computing, Path (computing), Data parallelism, Instruction-level parallelism, Critical path method