next up previous contents index
Next: Generalized Eigenvalue and Singular Up: Standard Eigenvalue and Singular Previous: Nonsymmetric Eigenproblems (NEP)   Contents   Index

Singular Value Decomposition (SVD)

The singular value decomposition of an m-by-n matrix A is given by

\begin{displaymath}
A = U \Sigma V ^T, \quad (A=U\Sigma V ^H \quad \mbox{in the complex case})
\end{displaymath}

where U and V are orthogonal (unitary) and $\Sigma$ is an m-by-n diagonal matrix with real diagonal elements, $\sigma _ i $, such that

\begin{displaymath}
\sigma_1 \ge \sigma_2 \ge \ldots \ge \sigma_{\min (m,n)} \ge 0 .
\end{displaymath}

The $\sigma _ i $ are the singular values of A and the first min(m,n) columns of U and V are the left and right singular vectors of A.

The singular values and singular vectors satisfy:

\begin{displaymath}
A v_i = \sigma_i u_i \quad \mbox{and} \quad
A^T u_i = \sigma_i v_i \quad ({\rm or} \quad
A^H u_i = \sigma_i v_i \quad )
\end{displaymath}

where ui and vi are the ith columns of U and V respectively.

There are two types of driver routines for the SVD. Originally LAPACK had just the simple driver described below, and the other one was added after an improved algorithm was discovered.


Table 2.5: Driver routines for standard eigenvalue and singular value problems
Type of Function and storage scheme Single precision Double precision
problem   real complex real complex
SEP simple driver SSYEV CHEEV DSYEV ZHEEV
  divide and conquer driver SSYEVD CHEEVD DSYEVD ZHEEVD
  expert driver SSYEVX CHEEVX DSYEVX ZHEEVX
  RRR driver SSYEVR CHEEVR DSYEVR ZHEEVR
  simple driver (packed storage) SSPEV CHPEV DSPEV ZHPEV
  divide and conquer driver SSPEVD CHPEVD DSPEVD ZHPEVD
  (packed storage)        
  expert driver (packed storage) SSPEVX CHPEVX DSPEVX ZHPEVX
  simple driver (band matrix) SSBEV CHBEV DSBEV ZHBEV
  divide and conquer driver SSBEVD CHBEVD DSBEVD ZHBEVD
  (band matrix)        
  expert driver (band matrix) SSBEVX CHBEVX DSBEVX ZHBEVX
  simple driver (tridiagonal matrix) SSTEV   DSTEV  
  divide and conquer driver SSTEVD   DSTEVD  
  (tridiagonal matrix)        
  expert driver (tridiagonal matrix) SSTEVX   DSTEVX  
  RRR driver (tridiagonal matrix) SSTEVR   DSTEVR  
NEP simple driver for Schur factorization SGEES CGEES DGEES ZGEES
  expert driver for Schur factorization SGEESX CGEESX DGEESX ZGEESX
  simple driver for eigenvalues/vectors SGEEV CGEEV DGEEV ZGEEV
  expert driver for eigenvalues/vectors SGEEVX CGEEVX DGEEVX ZGEEVX
SVD simple driver SGESVD CGESVD DGESVD ZGESVD
  divide and conquer driver SGESDD CGESDD DGESDD ZGESDD


next up previous contents index
Next: Generalized Eigenvalue and Singular Up: Standard Eigenvalue and Singular Previous: Nonsymmetric Eigenproblems (NEP)   Contents   Index
Susan Blackford
1999-10-01