Singular Value Decomposition: LAPACK Computational Routines

Developer Reference for Intel® oneAPI Math Kernel Library for Fortran

Download PDF

ID 766686

Date 4/28/2026

Version

Public

Singular Value Decomposition: LAPACK Computational Routines

This topic describes LAPACK routines for computing the singular value decomposition (SVD) of a general m -by- n matrix A :

A = UΣV^{H} .

In this decomposition, U and V are unitary (for complex A ) or orthogonal (for real A ); Σ is an m -by- n diagonal matrix with real diagonal elements σ _i :

σ_{1}≥σ_{2}≥ ... ≥σ_{min(m, n)}≥ 0 .

The diagonal elements σ _i are singular values of A . The first min(m, n) columns of the matrices U and V are, respectively, left and right singular vectors of A . The singular values and singular vectors satisfy

Av_{i} = σ_{i}u_{i} and A^{H}u_{i} = σ_{i}v_{i}

where u_i and v_i are the i -th columns of U and V , respectively.

To find the SVD of a general matrix A , call the LAPACK routine ?gebrd or ?gbbrd for reducing A to a bidiagonal matrix B by a unitary (orthogonal) transformation: A = QBP^{H} . Then call ?bdsqr , which forms the SVD of a bidiagonal matrix: B = U_{1}ΣV_{1}^{H} .

Thus, the sought-for SVD of A is given by A = UΣV^{H} =(QU_{1})Σ(V_{1}^{H}P^{H}) .

Table “Computational Routines for Singular Value Decomposition (SVD)” lists LAPACK routines (FORTRAN 77 interface) that perform singular value decomposition of matrices.

Table “Computational Routines for Singular Value Decomposition (SVD)” lists LAPACK routines that perform singular value decomposition of matrices. The corresponding routine names in the Fortran 95 interface are the same except that the first character is removed.

Computational Routines for Singular Value Decomposition (SVD)

Operation	Real matrices	Complex matrices
Reduce `A` to a bidiagonal matrix `B` : `A = QBP^{H}` (full storage)	?gebrd	?gebrd
Reduce `A` to a bidiagonal matrix `B` : `A = QBP^{H}` (band storage)	?gbbrd	?gbbrd
Generate the orthogonal (unitary) matrix `Q` or P	?orgbr	?ungbr
Apply the orthogonal (unitary) matrix `Q` or P	?ormbr	?unmbr
Form singular value decomposition of the bidiagonal matrix `B` : `B = UΣV^{H}`	?bdsqr ?bdsdc	?bdsqr

You can use the SVD to find a minimum-norm solution to a (possibly) rank-deficient least squares problem of minimizing ||Ax - b||_{2} . The effective rank k of the matrix A can be determined as the number of singular values which exceed a suitable threshold. The minimum-norm solution is

x = V_{k}(Σ_{k})^{-1}c

where Σ _k is the leading k -by- k submatrix of Σ , the matrix V_k consists of the first k columns of V = PV_{1} , and the vector c consists of the first k elements of U^{H}b = U_{1}^{H}Q^{H}b .

Select Your Language

Using Intel.com Search

Quick Links

Recent Searches

Advanced Search

Only search in

Developer Reference for Intel® oneAPI Math Kernel Library for Fortran

Singular Value Decomposition: LAPACK Computational Routines