Re: polynomial partial fractions

[Bill Allombert <Bill.Allombert@math.u-bordeaux.fr>]

On Tue, Dec 16, 2014 at 09:17:33PM +1100, Kevin Ryde wrote:
> Karim Belabas <Karim.Belabas@math.u-bordeaux.fr> writes:
> >
> >   C = vector(#P, i, Mod(B \/ PE[i], PE[i]));
> >   [Q, vector(#P, i, padic(lift(A/C[i]), P[i])), PE];
> 
> Nice.  I don't think I knew to go by poly inverses before.  How does
> that rate against a matsolve()?  About the same but less taking apart
> and putting back together polynomials perhaps?

Inverting POLMOD is done using the extended Euclid algorithm which is likely to
be much faster than matsolve() (O(n^2) instead of O(n^3)).

Cheers,
Bill.