Review of Approaches in High Performance Computing
Nenad Anchev, Blagoj Atanasovski, Sasko Ristov and Marjan Gusev University Sts Cyril and Methodius Faculty of Computer Science and Engineering RugjerBoshkovikj 16 Skopje, Macedonia
Abstract: High performance computing is a preferred area of study in computing which has potential impact in the future computing and hence using many approaches, the studies have been carried out. In the first approach a cluster of systems linked in LAN is used to bring a single system. In the second one in the grid system, a union of many computer resources are linked in multiple locations. When using the data flow approach, the issues are solved, while the von Nuemann principles is found to be less effective. Easily accessible dataflow engines have given ideas to advocate the developments. We in this work analysed the existing many approaches on high performance computing with the special observation about the dataflow use in the HPC.
Keywords: HPC, Data flow, FPGA, Grid Review of Approaches in High Performance Computing
[1] Francis, Matthew (2012-04-02). "Future telescope array drives development of exabyte processing". Retrieved 2012-10-24. [2] Baker, Mark, et al. "Cluster computing and applications." Encyclopedia of Computer Science and Technology 45.Supplement 30 (2002): 87-125. [3] The image is published under the Creative Commons licence, taken from en.wikipedia.org/wiki/ File:CUDA_processing_flow_(En).PNG as seen on 6.4.2013. [4] The image is taken from www.maxeler.com/technology/dataflow-computing/ as seen on 10.4.2013 [5] NVIDIA CUDA Home Pagewww.nvidia.com/object/cuda_home_new.html
Author Rights 
