Re: [eigen] Initial implementation of tensor support

[ Thread Index | Date Index | More Archives ]

Coming back to the original thread ...

On 30.10.2013 12:00, Christian Seiler wrote:
As I said in the initial thread: I didn't just plan to dump the code
here and leave, I am willing to maintain tensor support in the
foreseeable future. (Obviously, I can't promise anything for the distant
future, but that's life.) If it does make into Eigen's unsupported/,
then I will continue working on it and making it more useful for more

Naturally, you can't promise anything for the distant future -- if all goes well, by that point the code is stable, well-documented and has a big-enough user base to keep it maintained :)

generic applications. First goal would probably be to construct
expression templates so that assignments, binary operations (+, -) and
scalar multiplication work. (Which means basically just copying and
adjusting Eigen's current logic for matrices/vectors.)

Yes, having some actual tensor functionality besides data-holding/accessing would definitely be good ;) I don't know, what maturity level we require before integration. I guess you mostly need to wait if any objections arise. (I'm also not sure who is considered "core-developer" at the moment ...)

And of course, people who use it already should make suggestions or contribute (API-changes, feature requests, ...). And people who plan to use it should try it out.


Dipl.-Inf., Dipl.-Math. Christoph Hertzberg
Cartesium 0.049
Universität Bremen
Enrique-Schmidt-Straße 5
28359 Bremen

Tel: +49 (421) 218-64252

Mail converted by MHonArc 2.6.19+