Compiling thread-level parallel programs with a C-Compiler

Recent advances in multithreaded shared memory architectures have created a need for efficient and easy-to-use thread-level parallel (TLP) programming languages and tools supporting the synchronous shared memory model. In our previous work we have developed e, a fine-grained TLP programming language for multiprocessor architectures realizing such a model. The language uses a familiar c-like syntax and supports shared and private variables, arbitrary hierarchical groups of threads, and synchronous control structures. This allows a programmer to use various advanced TLP programming techniques like data parallelism, divide-and-conquer technique, different blocking techniques, and both synchronous and asynchronous programming style. In this paper we describe how an experimental compiler for e supporting virtual instruction-level parallelism exploitation can be set up by using a standard c-compiler. We will also shortly evaluate the compiler with real parallel programs on our scalable Eclipse network on chip architecture