Re: Minuit2

From: Lorenzo Moneta <Lorenzo.Moneta_at_cern.ch>
Date: Mon, 7 Sep 2009 09:30:10 +0200


Hello Patrick,

  Minuit2 should allocate memory of the order of npar x npar for each iteration, where npar is the number of free parameters. If you have a large number of parameters (let's say > 1000), you would need to use also some other algorithm, which does not need ot store the full Hessian matrix. A possible example of such algorithm is L- BFGS (see http://en.wikipedia.org/wiki/L-BFGS )

   Best Regards

  Lorenzo

On Sep 6, 2009, at 2:50 PM, Patrick Dupre wrote:

> Hello Lorenzo,
>
> I have another problem with Minuit. It just invides all my memory
> space !
> I am not sure of what is due to Minuit and what is not, ie due to my
> function.
> I tried to investigate by using valgrind, but I cannot tell what is
> doing wrong.
> I check all my code, and I cannot guess what could be wrong.
> Is there any thing that I should or could do to desallocate the
> memory, ie. something unusual ?
> During the run, which size of the memory is expected to be used ?
> I regularly pass 100 iterations.
> I am using MIGRAD and HESSE.
>
> Thank.
>
> --
> ---
> =
> =
> =
> =
> ======================================================================
> Patrick DUPRÉ | |
> Department of Chemistry | | Phone: (44)-
> (0)-1904-434384
> The University of York | | Fax: (44)-
> (0)-1904-432516
> Heslington | |
> York YO10 5DD United Kingdom | | email: pd520_at_york.ac.uk
> =
> =
> =
> =
> ======================================================================
Received on Mon Sep 07 2009 - 09:30:17 CEST

This archive was generated by hypermail 2.2.0 : Mon Sep 07 2009 - 17:50:01 CEST