More reimplementing matrix_integer_dense using FLINT

[sagetrac-kartikv]

Now with hermite normal forms. Added a new algorithm choice ("flint") to hermite form, and updated cutoffs to account for the fact that pari0 generally beats flint < 125 rows/cols and FLINT wins above that. pari1 and pari4 seem to always be much worse, as is ntl and the built-in padic method, but I only tested up to matrices of a few hundred entries. Also, flint does better for smaller matrices with larger entries, so that should be accounted for as well, but this is a reasonable first pass.

CC: @vbraun

Component: linear algebra

Keywords: matrix, dense

Author: Kartik Venkatram

Branch/Commit: e7fe591

Reviewer: R. Andrew Ohana

Issue created by migration from https://trac.sagemath.org/ticket/19081

[sagetrac-kartikv]

Author: Kartik Venkatram

[sagetrac-kartikv]

Branch: u/kartikv/19081_flint_hnf

[sagetrac-kartikv]

Commit: 4165319

[sagetrac-kartikv]

New commits:

4165319

adding flint as an option for HNF and using it as the default for moderately large matrices

[sagetrac-git]

Changed commit from 4165319 to f11bdf3

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

f11bdf3

now with correctly sized transformation matrices

[ohanar]

Changed reviewer from vbraun to R. Andrew Ohana

[ohanar]

comment:3

The description of the default algorithm doesn't match the code.

[sagetrac-git]

Changed commit from f11bdf3 to e7fe591

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

e7fe591

corrected comment

[sagetrac-kartikv]

comment:5

Sorry about that, too much undoing. Should be right now.

[ohanar]

comment:6

looks good to me

[vbraun]

Changed branch from u/kartikv/19081_flint_hnf to e7fe591