Split and Merge algorithm for SVD

This algorithm posses embarassingly paralel structure. Basic approach is to divide a given matrix by rows and applying svd on the submatrices and combining them accordingly to obtain u,v and d.

For complete information refer to "A split-and-merge approach fr singular value decomposition of large scale matrices" by Faming Liang, Runmin Shi and Qianxing Mo. Statistics and it's interface Volume9(2016) 453-459