LSRN: A Parallel Iterative Solver for Strongly Over- or Under-Determined Systems
Abstract
We describe a parallel iterative least squares solver named \texttt{LSRN} that is based on random normal projection. \texttt{LSRN} computes the min-length solution to , where with or , and where may be rank-deficient. Tikhonov regularization may also be included. Since is only involved in matrix-matrix and matrix-vector multiplications, it can be a dense or sparse matrix or a linear operator, and \texttt{LSRN} automatically speeds up when is sparse or a fast linear operator. The preconditioning phase consists of a random normal projection, which is embarrassingly parallel, and a singular value decomposition of size , where is moderately larger than 1, e.g., . We prove that the preconditioned system is well-conditioned, with a strong concentration result on the extreme singular values, and hence that the number of iterations is fully predictable when we apply LSQR or the Chebyshev semi-iterative method. As we demonstrate, the Chebyshev method is particularly efficient for solving large problems on clusters with high communication cost. Numerical results demonstrate that on a shared-memory machine, \texttt{LSRN} outperforms LAPACK's DGELSD on large dense problems, and MATLAB's backslash (SuiteSparseQR) on sparse problems. Further experiments demonstrate that \texttt{LSRN} scales well on an Amazon Elastic Compute Cloud cluster.
Cite
@article{arxiv.1109.5981,
title = {LSRN: A Parallel Iterative Solver for Strongly Over- or Under-Determined Systems},
author = {Xiangrui Meng and Michael A. Saunders and Michael W. Mahoney},
journal= {arXiv preprint arXiv:1109.5981},
year = {2012}
}
Comments
19 pages