Scalability of <i>k</i>-Tridiagonal Matrix Singular Value Decomposition
Andrei Tănăsescu,
Mihai Carabaş,
Florin Pop,
Pantelimon George Popescu
Affiliations
Andrei Tănăsescu
Computer Science and Engineering Department, Faculty of Automatic Control and Computer Science, University Politehnica of Bucharest, Splaiul Independentei 313, 060042 Bucharest, Romania
Mihai Carabaş
Computer Science and Engineering Department, Faculty of Automatic Control and Computer Science, University Politehnica of Bucharest, Splaiul Independentei 313, 060042 Bucharest, Romania
Florin Pop
Computer Science and Engineering Department, Faculty of Automatic Control and Computer Science, University Politehnica of Bucharest, Splaiul Independentei 313, 060042 Bucharest, Romania
Pantelimon George Popescu
Computer Science and Engineering Department, Faculty of Automatic Control and Computer Science, University Politehnica of Bucharest, Splaiul Independentei 313, 060042 Bucharest, Romania
Singular value decomposition has recently seen a great theoretical improvement for k-tridiagonal matrices, obtaining a considerable speed up over all previous implementations, but at the cost of not ordering the singular values. We provide here a refinement of this method, proving that reordering singular values does not affect performance. We complement our refinement with a scalability study on a real physical cluster setup, offering surprising results. Thus, this method provides a major step up over standard industry implementations.
