pbdDMAT (version 0.4-2)

eigen2: eigen2

Description

Compute eigenvalues and, optionally, eigenvectors of a real symmetric matrix by seraching over ranges of values or ranges of indices.

Usage

eigen2(x, range = c(-Inf, Inf), range.type = "interval",
  only.values = FALSE, abstol = 1e-08, orfac = 0.001)

Arguments

x

symmetric, numeric ddmatrix.

range

A set of interval endpoints, i.e. a numeric pair. Controls the set of values over which the eigenvalue search occurs.

range.type

Controls whether interval range refers to a set of possible values for the eigenvalues, or a set of indices for the eigenvalues. Options are "interval" and "index".

only.values

logical. Determines whether only the eigenvalues should be computed, or if the eigenvectors should as well.

abstol

The absolute error tolerance for the eigenvalues.

orfac

Specifies which eigenvectors should be reorthogonalized. Eigenvectors that correspond to eigenvalues which are within tol=orfac*norm(A)of each other are to be reorthogonalized.

Value

Returns a distributed matrix.

Details

This new method computes selected eigenvalues and, optionally, eigenvectors of a real symmetric matrix. Eigenvalues and eigenvectors can be selected by specifying either a range of values or a range of indices for the desired eigenvalues.

Examples

# NOT RUN {
# Save code in a file "demo.r" and run with 2 processors by
# > mpiexec -np 2 Rscript demo.r

library(pbdDMAT, quiet = TRUE)
init.grid()
init.grid()

comm.set.seed(seed=1234, diff=TRUE)

x <- crossprod(ddmatrix("rnorm", 10, 3, bldim=2))
y <- as.matrix(x)

comm.print(eigen(y))

### Look for eigenvalues in the range 0 to 10
ev <- eigen2(x, range=c(0, 10), only.values=TRUE)
comm.print(ev)

finalize()
# }
# NOT RUN {
# }