Hello Rooters,
I have noticed something strange in TF1::IntegralMultiple. When you try
to evaluate the integral of simple functions, you obtain very unprecise
results whatever the precision you ask for. And indeed looking at the code
one can notice that :
in H_F1.cxx
// The original algorithm expected a parameter MAXPTS
// where MAXPTS = Maximum number of function evaluations to be allowed.
// Here we set MAXPTS to 2*(the lowest possible value)
Int_t maxpts = 2*irlcls;
The maxpts defined here seems too low, raising it gives better results.
Shouldn't it be a parameter of the function?
Thank you,
______________________________________________________________________________
Regis Terrier
Astroparticle Group | tel: +33 1-44271502
PCC - College de France | fax: +33 1-43546989
11, Place Marcelin Berthelot
75231 Paris Cedex 05 | e-mail: terrier@cdf.in2p3.fr
This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:25 MET