The Efficient load balancing in the parallel Computer

Abstract: Recently Cpu workload hardware Technology and multiprocessor Service are developing rapidly. Mathematically computation must be completed within a certain time period .Through the mathematical Analysis our Study Identify various useful result that should be interest in system designer .If all the parallel computer are not same type means not same configuration then proper load balance not occur so some computer finish their work earlier than other and sit ideal which degrade the performance of multicomputer system. .we introduce new Algorithm for program development on multicomputer environment called ODDA (optimize data distribution algorithm) to address the proper load balancing purpose . This algorithm work in the master slave modal. The main processor called the master processor which dynamically assign the workload to the remaining cooperative slave processor in run time environment. The efficient parallel computer has 3 element - collection of device , a Network connect to these computer and s/w that enable used to share data between them. MPI and PVM is the best Software used for message passing interfacing computation between them. Multiprocessor computation is the best way of resource handling and scheduling strategy . This paper contain Dynamic approach load balancing for process migration using process level environment employed parallel programs.

Keywords: Master processor , slave processor , Data distribution , Task manager , MPI, heterogeneous platform .