Re: empirical_inv() out of memory

[Glenn Golden]

"Dmitri A. Sergatskov" writes:
> Glenn Golden wrote:
> > Running 2.1.50: 
> > 
> > Is the following a reasonable out-of-memory situation? 
> > 
> > 
> >     1>   f = 1 - 10 .** [-1:-1/3:-6];
> >     2>   g = abs(rand(4e6,1)) .** 2;
> >     3>   empirical_inv (f, g)
> >     error: memory exhausted -- trying to return to prompt
> 
> I run it on my computer while watching 'top' in other terminal.
> It went through, but octave process size at its maximum was 1201MB.
> 

Spasibo.  It runs ok on a larger memory machine at my work too.
My question really is, is it reasonable for a distribution analysis
function like empirical_inv() -- which, by its nature is expected
to be passed fairly large datasets -- to exhaust 256M when called
with an argument which is roughly only 1/8 the available memory?
Certainly this is something about which reasonble people can disagree,
but it seemed to me in this case that the memory usage was a bit
extravagant.

Looks like the core function (discrete_inv()) is written in a way
which causes a lot of temps to be created, although it's not obvious
at first glance why it should be using as much as 1.2GB...

Glenn



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------