WIT Press


PCor: A Prototype For Resource Oriented Computing

Price

Free (open access)

Paper DOI

10.2495/HPC020251

Volume

27

Pages

Published

2002

Size

914 kb

Author(s)

A. Pina, V. Oliveira, C. Moreira & A. Alves

Abstract

pCor: a protoype for resource oriented computing A. Pina1, V. Oliveira1, C. Moreira1 & A. Alves1 1Departamento de Informatica, Universidade do Minho, Portugal Abstract In this paper we present COR a resource oriented computing model that address the question of how to integrate user-level fine-grained multithreading, communication and coordination into a cluster of symmetrical multiprocessor computers. To support the design of complex distributed application using the proposed paradigm we built pCoR a run-time system which has new areas that represents extensions to the strict shared memory and message passing models supported by other platforms: remote operations, dynamic domains, communication ports, multithreading management, shared memory, replication and partition are some of its distinguished features. In addition, it provides a thread-safe transport communication layer to take advantage of modem high-performance commodity hardware/software like Myrinet network. 1 Introduction The emerging field of parallel computing illustrated by the increased importance of clustering technologies [1] puts strong demands to the problem of organizing computation, communication and coordination. On large-scale projects where the communication patterns cannot be determined in advance the main problem is correctness, due to the non-determinism and concurrent access to resources. Another issue that poses additional requirements to the task of parallel programming and that can have a great impact upon performance is the need to establish a wised relationship between problem specification and decomposition onto the hardware of a parallel computer. The rise of popular public domain software such as PVM [2] and MPI [3] had a major influence on the wide use of parallel computers, ranging from high performance supercomputer to clusters of workstations or home computers due to

Keywords