qr: QR

In float: 32-Bit Floats

QR

Description

QR factorization and related functions.

Usage

## S4 method for signature 'float32'
qr(x, tol = 1e-07, ...)

## S4 method for signature 'ANY'
qr.Q(qr, complete = FALSE, Dvec)

## S4 method for signature 'ANY'
qr.R(qr, complete = FALSE)

## S4 method for signature 'ANY'
qr.qy(qr, y)

## S4 method for signature 'ANY'
qr.qty(qr, y)

Arguments

x	A float matrix.
tol	The tolerance for determining numerical column rank.
...	Ignored.
qr	Output of qr().
complete	Should the complete or truncated factor be returned?
Dvec	Vector of diagonals to use when re-constructing Q (default is 1's).
y	A vector/matrix or right hand sides (int, float, or double).

Details

The factorization is performed by the LAPACK routine sgeqp3(). This should be similar to calling qr() on an ordinary R matrix with the argument LAPACK=TRUE. Calling qr(x, LAPACK=FALSE) on a double precision matrix 'x' (the default) will not be comparable in performance (it is much slower) or numerics to calling qr(s) where 's' is single a float matrix.

Examples

library(float)

x = flrunif(10, 3)
qr(x)

float documentation built on May 29, 2024, 8:13 a.m.