On Sat, 20 Feb 2021 at 14:28, Kartik Sethi <[email protected]> wrote:
>
> Thanks Oscar Benjamin for the wonderful suggestions. I will start looking 
> into algorithms and methods that have not been implemented for the domain 
> Matrix class yet.
> I would be happy to help in building Domain Matrix class.
>
> Do you think improving and implementing algorithms in Domain Matrix class 
> would be a suitable Gsoc Proposal?

Yes, absolutely. If anyone is interested in improving sympy's matrix
calculations I think that this is the most important thing to do right
now. Many algorithms are already implemented for Matrix but are
basically unusable because they are too slow except in the simplest
cases. If we can make core matrix operations faster then that will
benefit all of them.

In the cases where DomainMatrix is slow the main thing that is slow is
gcd computations. That can be improved by making more use of fraction
free algorithms as in this PR that anyone is welcome to finish:
https://github.com/sympy/sympy/pull/20483

The other side of it is improving the gcd algorithms themselves but I
think that should be a whole separate project.


--
Oscar

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sympy/CAHVvXxTLXFUrcnrTY4cDMytEPSbM9fR0G-gZtap9vRPi0Ogegw%40mail.gmail.com.

Reply via email to