New mapping scheme of matrix algorithms on the Biswapped network

Biswapped Network (BSN) is built of 2n copies of an basis network and total nodes are 2n2. In this paper, we develop new algorithms by improving Group Submatrix Mapping (GSM) to multiply two vectors, a vector and a matrix, and two matrices on BSN, and their complexity are analyzed. Algorithm analysis proves that our suggested scheme is more suitable for large-scale matrix multiplication on BSN than other schemes. © 2013 by Binary Information Press.