A novel two and a half dimensional distributed computing architecture

The demand for high performance distributed computing systems to perform various sophisticated processing tasks has been multiplied over the past decades. With the reducing costs of circuit components, the development of computer architectures has been navigated to become more distributive and specialized. In the paper, a novel distributed computing architecture is presented, in particular, to enhance the utilization of hardware resources, the computational efficiency and the versatility of the principle architecture. Through the realization of LU factorization with partial pivoting, the granularity of processing cells, the versatility and the performance of the novel architecture are clarified