forum.alglib.net
http://forum.alglib.net/

Number of intervals and function evaluations: Gauss-Kronrod
http://forum.alglib.net/viewtopic.php?f=2&t=2042
Page 1 of 1

Author:  jeffj [ Tue May 27, 2014 7:06 pm ]
Post subject:  Number of intervals and function evaluations: Gauss-Kronrod

I'm under the impression that autogkintegrate with autogksmooth uses Gauss-Kronrod quadrature. The purpose of Kronrod rules is to reduce the number of function evaluations. Why does autogkreport.nintervals = 74 while autogkreport.nfev = 2205? Shouldn't there be fewer function evaluations?

Author:  jeffj [ Tue May 27, 2014 8:24 pm ]
Post subject:  Re: Number of intervals and function evaluations: Gauss-Kron

I'll add that I discovered this because I noticed that it took autogkintegrate more than an order of magnitude longer to perform integration than the GNU GSL procedure gsl_integration_qag, and I was looking for the cause. I think the high number of evaluations is the culprit.

Page 1 of 1 All times are UTC
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/