sla.svd {RScaLAPACK}R Documentation

Singular Value Decomposition of a Matrix

Description

Compute the singular-value decomposition of a rectangular matrix.

Usage

sla.svd(A, nu=NULL, nv=NULL, NPROWS=0, NPCOLS=0, MB=16, RFLAG=1, SPAWN=1)

Arguments

A

A matrix whose SVD decomposition is to be computed

nu

The number of left singular vectors to be computed. Must be either 0 or min(nrow(x), ncol(x)).

nv

The number of right singular vectors to be computed. Must be either 0 or min(nrow(x), ncol(x)).

NPROWS

Number of Process Rows in the Process Grid.

NPCOLS

Number of Process Cols in the Process Grid.

MB

Block Size.

RFLAG

Flag saying whether the Process Grid should be released after computation.

SPAWN

Flag saying whether a new Process Grid should be spawned.

Details

If the number of processor rows and columns are both zero, one processor is used. If the number of processor rows is nonzero and the number of processor columns is zero, then the processor rows is taken to be a number of processors, and a grid is made accordingly.

Author(s)

Nagiza Samatova (samatovan@ornl.gov), Guruprasad Kora (koragh@ornl.gov), Srikanth Yoginath (yoginathsb@ornl.gov), David Bauer (bauerda@ornl.gov)

References

http://www.netlib.org/scalapack/

http://mathworld.wolfram.com/SingularValueDecomposition.html

See Also

sla.solve the SCALAPACK version of the R function solve

sla.qr the SCALAPACK version of the R function qr

La.svd the LAPACK implementation of singular value decomposition.

Examples

library(RScaLAPACK)
rnorm(64*64)->x
dim(x)=c(64,64)
sla.svd(x, NPROWS=2, NPCOLS=2, MB=16)

[Package RScaLAPACK version 0.6.1 Index]