Set up cross polytope with both Vrep and Hrep (if backend supports it)

[kliem]

We set up the cross polytope with precomputed double description, if the backend supports it. Otherwise, the vertices will be used as before (the shorter input is selected automatically by #28880).

Before this ticket:

sage: %time P = polytopes.cross_polytope(8, backend='field')
CPU times: user 14.3 s, sys: 4.01 ms, total: 14.3 s
Wall time: 14.3 s
sage: %timeit P = polytopes.cross_polytope(13, backend='ppl')
1 loop, best of 5: 312 ms per loop

With this ticket:

sage: %time P = polytopes.hypercube(8, backend='field')
CPU times: user 29.8 ms, sys: 11.2 ms, total: 41 ms
Wall time: 40.3 ms
sage: %timeit P = polytopes.cross_polytope(13, backend='ppl')
1 loop, best of 5: 312 ms per loop

sage: %timeit P = polytopes.hypercube(13, backend='field')
10 loops, best of 5: 69.5 ms per loop

Note: This might slow down some backends (e.g. ppl) a bit, as the inequalities are precomputed and then discarded again. However, this seems to be only minor. The speed up for backend field is significant, as nothing is being computed anymore.

CC: @jplab @LaisRast

Component: geometry

Keywords: cross polytope, precomputed data

Author: Jonathan Kliem

Branch/Commit: a14d5a1

Reviewer: Frédéric Chapoton

Issue created by migration from https://trac.sagemath.org/ticket/29323

[kliem]

Commit: dd291b5

[kliem]

New commits:

dd291b5

set up cross polytope with precomputed data if backend supports it

[kliem]

Branch: public/29323

[kliem]

Changed branch from public/29323 to public/29323-reb

[kliem]

Changed commit from dd291b5 to a14d5a1

[kliem]

New commits:

862d44f	set up cross polytope with precomputed data if backend supports it
a14d5a1	fixed failing doctests

[fchapoton]

comment:3

ok

[fchapoton]

Reviewer: Frédéric Chapoton

[vbraun]

Changed branch from public/29323-reb to a14d5a1