fft fails for lengths 392, 588, 968, 980 .... (PR#1429) - R-devel

Mon, Apr 1, 2002 12:25 PM #

R 1.4.1, Linux and Windows

for(i in 1:1000) {
    X <- rnorm(i)
    XX <- fft(fft(X), inverse=T)/i
    if(max(Mod(XX-X)) > 1e-10) print(i)
}

[1] 392
[1] 588
[1] 968
[1] 980

and I then get a segfault during gc().  The answers are way off, with
imaginary parts 1e10 or more.

These numbers are all multiples of 7^2 or 11^2.

(Based on a report to R-help
Date: Thu, 28 Mar 2002 09:37:34 +0100
From: Gabriel Fricout <Gabriel.FRICOUT@cmm.ensmp.fr>
)

Brian D. Ripley,                  ripley@stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._