The use of grid computing is growing these days. Grid computing not only uses dedicated systems but also systems already in use, thereby improving resource utilization. In such a scenario, there is a need to develop performance evaluation techniques (Liang et al., 2004) that not only include execution time but also incorporate factors such as scale up, speed up and other important factors which are elaborated in this paper. These factors have also been tested in a grid developed with shared resources (processors) under various conditions. This paper elaborates the grid developed, the applications tested, the factors used and finally presents results in the form of graphs.

Grid computing (Bart et al., 2005) has become more and more essential and extremely useful nowadays. To get a better idea of what grid computing is, it can be compared with the power grid. Generally, one does not know where and how the electricity is generated. The energy is generated somewhere and pooled in the power grid and distributed to the places in need of it. A computational grid also works similarly. Users gain access to the computational resources without having the knowledge of where the resource is located or the underlying technologies used.

Grid computing can be generally defined as the process of integrating various technologies to provide better solutions. This forms a part of distributed computing. A distributed system is one which provides a single system view to a collection of machines. Independent and physically scattered computing resources falling under the roof of a single large computer is called a grid. The computing resources may be geographically separated and may be heterogeneous. Heterogeneity can be in the architecture of the nodes used or the operating system they use.

The components of the grid must lie on a network for the resources to communicate. The basic difference between a network and a distributed system is that a network just provides a pipe for the resources to communicate, whereas in distributed systems, the communication between the nodes takes place in a more coordinated fashion.

Generally, open standards and protocols are used as they provide mechanisms needed to facilitate communication between components developed by different vendors. They also allow the developers to concentrate on the business logic rather than on the method of programming communication routines.

Computer Sciences Journal, Computational Grid Systems, Parallel CPU Capacity, Resource Management Tool, Quality of Service, QoS, Resource Balancing, Red Hat Enterprise Linux 5, RHEL5, Matrix Multiplication, Distributed System, Computational Fluid Dynamics, Genetic Algorithm, GA, Matrix Multiplication, Darwin's Evolution Theory, RHEL Machines, Resource Management.