🚀 go-pugleaf

In article <BtK3xJ.253@watserv1.uwaterloo.ca>, alongley@cape.UWaterloo.ca (Allan Longley) writes:
> Hello, I hope you can help.
> 
> I am using the NAG library routine for integrating a stiff system of
> ordinary differential equations D02EBF.  The routine is called multiple 
> times in the code with the same initial state, but is returning different 
> answers ...  very different answers!  I even reset the work-space array
> W(iw) before each use of D02EBF.
> 
> I am programming in FORTRAN77 with NAG Mark14B in AIX on an IBM R6000.  
> 
> This is not good.  Can you help?  Please.
> 
> -----------------------------------------------------------------------------
> Allan Longley, University of Waterloo, Department of Chemical Engineering
> e-mail: alongley@cape.uwaterloo.ca
> voice: (519) 885-1211 x3816
> home: (519) 746-5747
> ------------------------------------------------------------------------------

for everyone's information:

NAG Ltd. operates a Response Centre for technical queries
       tel  +44 865 311744
       fax  +44 865 311755
    e-mail  infodesk@nag.co.uk
      post  NAG Response Centre, NAG Ltd, Wilkinson House
            Jordan Hill Road, OXFORD, UK, OX2 8DR
NAG Inc provide a Technical Support Desk Service for North America
       tel  (708) 971 2337
       fax  (708) 971 2706
and plan a Response Centre due to be launched later this year.

for the original poster's information:

We are not aware of any problems with NAG routine D02EBF and neither have
we been able to duplicate the effects you describe. We would be very
grateful if you could get in touch direct and provide us with more 
precise details/example demonstrating the problem. The NAG Library
is now available at Mark 15 on IBM RS/6000 machines.

Richard.

-- 
** R.W. Brankin -- Numerical Algorithms Group Ltd. -- nagrwb@vax.ox.ac.uk **
**   suggestions for sig. welcome   --   will need   include<std.discl>   **