Developer Reference for Intel® oneAPI Math Kernel Library for Fortran

ID 766686
Date 3/22/2024
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents

?lasq5

Computes one dqds transform in ping-pong form. Used by ?bdsqr and ?stegr.

Syntax

call slasq5( i0, n0, z, pp, tau, sigma, dmin, dmin1, dmin2, dn, dnm1, dnm2, ieee, eps )

call dlasq5( i0, n0, z, pp, tau, sigma, dmin, dmin1, dmin2, dn, dnm1, dnm2, ieee, eps )

Include Files

  • mkl.fi

Description

The routine computes one dqds transform in ping-pong form: one version for ieee machines, another for non-ieee machines.

Input Parameters

i0

INTEGER. First index.

n0

INTEGER. Last index.

z

REAL for slasq5

DOUBLE PRECISION for dlasq5.

Array, DIMENSION (4n). z holds the qd array. emin is stored in z(4*n0)

to avoid an extra argument.

pp

INTEGER. pp=0 for ping, pp=1 for pong.

tau

REAL for slasq5

DOUBLE PRECISION for dlasq5.

This is the shift.

sigma

REAL for slasq5

DOUBLE PRECISION for dlasq5.

This is the accumulated shift up to the current point.

ieee

LOGICAL. Flag for IEEE or non-IEEE arithmetic.

eps

REAL for slasq5

DOUBLE PRECISION for dlasq5.

This is the value of epsilon used.

Output Parameters

dmin

REAL for slasq5

DOUBLE PRECISION for dlasq5.

Minimum value of d.

dmin1

REAL for slasq5

DOUBLE PRECISION for dlasq5.

Minimum value of d, excluding d(n0).

dmin2

REAL for slasq5

DOUBLE PRECISION for dlasq5.

Minimum value of d, excluding d(n0) and d(n0-1).

dn

REAL for slasq5

DOUBLE PRECISION for dlasq5. Contains d(n0), the last value of d.

dnm1

REAL for slasq5

DOUBLE PRECISION for dlasq5. Contains d(n0-1).

dnm2

REAL for slasq5

DOUBLE PRECISION for dlasq5. Contains d(n0-2).