high performance computing on graphics processing units: hgpu.org

In 1980’s time, people believed that computer would help to create more faster and efficient processors. But parallel processing challenged the idea. It joined two or more computers together to solve a problem jointly. It was a trend in 1990 to move away from expansive super computers towards network computers like PCs or Workstations. It has improvised the performance components for PC, Networks or Workstations. These are making the network or the cluster of computers more cost-effective and consequently leading to the low cost commodity supercomputing. PCs, Workstations or SMPs are becoming high performance materials. They are not only low cost, but they also generating parallel programming environments. They are helping to build commodity hardware, fast LAN, standard software components like UNIX, MPI and PVM parallel programming environment. These systems are scalable. These can be tuned to available budget and computational needs and allow efficient execution of both sequential and parallel applications. Clusters use intelligent mechanisms for dynamic and network based sharing which response to resource based requirements and their availability. These mechanisms support scalability of cluster performance and allow flexible use of workstations. Presently the cluster or network-wide available resources are supposed to be larger than available resources and the workstations. These intelligent mechanisms also allow the clusters to support multiuser and time sharing parallel execution environments where it is necessary to share the resources. The idea of exploiting such a significant computational capability in networks have gained enthusiastic acceptance within the high performance computer community. This is motivated by the desire to minimize the economical risk and to build a consumer based off shelf technology. Cluster computing is presently considered to bring the wave of the future where bigger computing problems would be solved very easily.