Re: [Boost-bugs] [Boost C++ Libraries] #12066: Accuracy of modified Bessel functions

Subject: Re: [Boost-bugs] [Boost C++ Libraries] #12066: Accuracy of modified Bessel functions
From: Boost C++ Libraries (noreply_at_[hidden])
Date: 2016-03-14 13:01:41

#12066: Accuracy of modified Bessel functions
------------------------------------------+--------------------------
  Reporter:  Pavel Holoborodko <pavel@…>  |      Owner:  johnmaddock
      Type:  Library Submissions          |     Status:  new
 Milestone:  To Be Determined             |  Component:  math
   Version:  Boost 1.61.0                 |   Severity:  Optimization
Resolution:                               |   Keywords:
------------------------------------------+--------------------------

Comment (by Pavel Holoborodko <pavel@…>):

 So far I derived 128-bit coefficients for I0 - the same form of
 approximation can be used, as I used for double (it is a bit more accurate
 near x = 0 than originally proposed by Blair).

 From preliminary tests, I1 needs different approximation from what I used
 for double and from Blair's.
 K0/K1 is on the way as well.

 I will share everything as soon as I finish this - it takes time to run
 search for several different forms of approximation in extended precision
 and do other work on the same computer...

-- 
Ticket URL: <https://svn.boost.org/trac/boost/ticket/12066#comment:7>
Boost C++ Libraries <http://www.boost.org/>
Boost provides free peer-reviewed portable C++ source libraries.