[Bug-gsl] [bug #40116] possible error in integration routines

[Patrick Alken]

URL:
  <http://savannah.gnu.org/bugs/?40116>

                 Summary: possible error in integration routines
                 Project: GNU Scientific Library
            Submitted by: psa
            Submitted on: Thu 26 Sep 2013 09:34:45 PM GMT
                Category: Runtime error
                Severity: 3 - Normal
        Operating System: 
                  Status: None
             Assigned to: None
             Open/Closed: Open
                 Release: 
         Discussion Lock: Any

    _______________________________________________________

Details:

(from address@hidden)

Hello,

I apologize if this is the wrong place. I’m currently working with the
integration utilities
of GSL and I’d like some help about this particular point.

I’m not sure the function defined in the file integration/qpsrt2.c and
invoked, for example, 
at line 418 of integration/qags.c is doing what it should, according to the
original
algorithm, nor according to the comment

        /* test whether the interval to be bisected next is the
           smallest interval. */

at line 415 of qags.c. 

In the QUADPACK version (http://www.netlib.org/quadpack/dqagse.f), at line 360
we can 
see the actual test being performed

        if(dabs(blist(maxerr)-alist(maxerr)).gt.small) go to 90

I think the refining loop may fail at some point, since the loop won’t cycle
when the test
evaluates to true.

I don’t have a buggy case to report but it seems to me that it isn’t
right.


Thank you for any help,
Seneika.





    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?40116>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/