SD-MARC: A New Multi-Processor Architecture
A. Somdip Dey · 2012
In modern day, HPC (High Performance Computing) is applied to do massive scientific or huge computational works, but HPC architecture mostly consist of computer clusters (series of computers connected to each other to solve a single problem / task). Thus it is problematic for many institutions or organizations to maintain a computer cluster for doing huge computation. So, with the advancement of technology and efficient multiprocessor architecture, it is possible to integrate thousands of processors in one computer system and apply that system for doing the huge computation. This paper basically provides the idea of a new model to apply multiple processors (in order of n (n x n) processors) in one computer system Architecture and how to implement the model to compute different problems faster than they can compute in reality; provided: n>=4 and n=2m, where m=0, 1, 2, 3, 4 ........ N.