|
From: | Michele |
Subject: | Re: About special functions reogranization |
Date: | Wed, 19 Jul 2017 10:20:53 +0200 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.2.1 |
On 07/14/2017 09:04 PM, Rik wrote:
Dear Rik, thanks for your comments. I already worked on gammainc, you can see the details in [1]. The important thing is that Marco, Nir and I implemented it as an .m file, with a subroutine in C++ (you can find this implementation in the bookmark "gammainc" of my repository [2]). Every special function has more or less the same structure: the domain is split into subdomains and different subfunctions are called. Each subfunction implements an algorithm, the most used ones being series expansions and continued fractions. I think that I could continue my work (e.g. betainc) in this way, finding algorithms that are accurate in their domains. Then, with maybe the help of the community, start to translate the subfunctions from .m files to C++ (I'm not autonomous with C++). In this way, at the end of GSoC we will have accurate algorithms, that in the next time can become also fast. [1] http://gsocspecfun.blogspot.it/2017/06/gammainc.html [2] https://bitbucket.org/M_Ginesi/octave --
Michele Ginesi |
[Prev in Thread] | Current Thread | [Next in Thread] |