Bessel big floats



On Thu, Jul 10, 2008 at 10:58 PM, Richard Hennessy <rvh2007 at comcast.net>
wrote:

> I am getting weird results when I try to compute the bessel_k function to
> 40 or 50 places.  I think that the bessel_k function is just returning a
> float and the big float is just random digits from the 17th digit on.  Is
> this the correct analysis?
>

Yes, bessel_k returns a float.  What do you mean when you say "try to
compute ... to 40 or 50 places"?  If it is returning a numerical value for a
bfloat argument, that is a bug.

In the future, please include a brief code snippet demonstrating your
problem.  It will make it easier to diagnose.

              -s