Performance improvements
authorAntonio Jimenez Pastor <antonio@ebook.dk-compmath.jku.at>
Wed, 13 Jun 2018 11:45:01 +0000 (13:45 +0200)
committerAntonio Jimenez Pastor <antonio@ebook.dk-compmath.jku.at>
Wed, 13 Jun 2018 11:45:01 +0000 (13:45 +0200)
commit4598b6852a8e25f8e5a8daa886e726854bb3c71a
treeca21480ca8ea5b11b92ec7e8b9b1136f54046c1f
parent4dc011cf370fab2f27d8f2319c5060d07a4ef020
Performance improvements

1 - The method for computing the root of the indicial polynomial now is cached.
This means that operators with the same indicial polynomial take less time to
compute their roots.
2 - Computing th roots of the indicial polynomial when there are parameters is done
using a modular approximation: we evaluate the parameters and when the polynomial
has the same degree we compute the roots of a univariate polynomial. That is a superset
of the roots we want, so we check and we keep those that were roots for the indicial polynomial.

Minor changes:
   matrix.py now use the method is_Matrix instead of sage.matrix.Matrix (deprecated in
SAGE: https://trac.sagemath.org/ticket/24096
ajpastor/misc/bareiss.py
ajpastor/misc/matrix.py
ajpastor/operator/operator.py
releases/diff_defined_functions__0.5.zip
releases/old/diff_defined_functions__0.5__18.06.13_13:45:01.zip [new file with mode: 0644]