Factory: An Object-Oriented Parallel Programming Substrate for Deep Multiprocessors

Abstract. Recent advances in processor technology such as Simultaneous Multithreading (SMT) and Chip Multiprocessing (CMP) enable parallel processing on a single die. These processors are used as building blocks of shared-memory multiprocessor systems, or clusters of multiprocessors. New programming languages and tools are necessary due to the complexities introduced by systems with multigrain, multilevel execution capabilities. This paper introduces Factory, an object-oriented parallel programming substrate which allows programmers to express multigrain parallelism, but alleviates them from having to manage it. Factory is written in C++ without introducing any extensions to the language. Because it leverages existing C++ constructs to express arbitrarily nested parallel computations, it is highly portable and does not require extra compiler support. Moreover, Factory offers programmability and performance comparable to already established multithreading substrates.

• Type: Conference
• Year: 2005
• Where: HPCC
• Authors: Scott Schneider, Christos D. Antonopoulos, Dimitrios S. Nikolopoulos
• Keywords: Enable Parallel Processing, HPCC 2005, Parallel Programming Substrate, Shared-memory Multiprocessor Systems