Opened 13 months ago
Closed 13 months ago
#32152 closed enhancement (fixed)
Set up pyramid with both Vrep and Hrep
Reported by: | gh-kliem | Owned by: | |
---|---|---|---|
Priority: | major | Milestone: | sage-9.4 |
Component: | geometry | Keywords: | polyhedron, double description, precomputed, pyramid |
Cc: | Merged in: | ||
Authors: | Jonathan Kliem | Reviewers: | Travis Scrimshaw |
Report Upstream: | N/A | Work issues: | |
Branch: | 2b1c64b (Commits, GitHub, GitLab) | Commit: | 2b1c64b047628dfd5179cb98f637da57bb74452f |
Dependencies: | Stopgaps: |
Description
We set up the pyramid over a polyhedron with the double description.
Before:
sage: P = polytopes.permutahedron(6) sage: %time Q = P.pyramid() CPU times: user 99.4 ms, sys: 0 ns, total: 99.4 ms Wall time: 98 ms sage: P = polytopes.hypercube(8) sage: %time Q = P.pyramid() CPU times: user 34.2 ms, sys: 24 µs, total: 34.2 ms Wall time: 33.3 ms sage: P = polytopes.cross_polytope(8) sage: %time Q = P.pyramid() CPU times: user 40.5 ms, sys: 3.96 ms, total: 44.5 ms Wall time: 43.3 ms sage: P = polytopes.hypercube(6, backend='field') sage: %time Q = P.pyramid() CPU times: user 974 ms, sys: 0 ns, total: 974 ms Wall time: 973 ms
After:
sage: P = polytopes.permutahedron(6) sage: %time Q = P.pyramid() CPU times: user 119 ms, sys: 197 µs, total: 120 ms Wall time: 118 ms sage: P = polytopes.hypercube(8) sage: %time Q = P.pyramid() CPU times: user 30.5 ms, sys: 3.92 ms, total: 34.4 ms Wall time: 33.4 ms sage: P = polytopes.cross_polytope(8) sage: %time Q = P.pyramid() CPU times: user 40.4 ms, sys: 0 ns, total: 40.4 ms Wall time: 39.7 ms sage: P = polytopes.hypercube(6, backend='field') sage: %time Q = P.pyramid() CPU times: user 8.04 ms, sys: 0 ns, total: 8.04 ms Wall time: 7.99 ms sage: P = polytopes.hypercube(10, backend='field') sage: %time Q = P.pyramid() CPU times: user 46.4 ms, sys: 98 µs, total: 46.5 ms Wall time: 45.6 ms
Change History (6)
comment:1 Changed 13 months ago by
- Status changed from new to needs_review
comment:2 Changed 13 months ago by
comment:3 Changed 13 months ago by
Yes, there could be a slowdown and especially for something as simple as the pyramid, but not at the moment.
Currently, all polyhedra are immutable and non-interactive and the Vrepresentation and Hrepresentation is always already computed.
With mutable polyhedra (theoretically makes sense with ppl
, normaliz
and polymake
) things will be a bit different. But then just embedding in higher dimension and adding a vertex would be the better approach anyway. So I would propose overwriting the pyramid method in the mutable case anyway.
comment:4 Changed 13 months ago by
- Reviewers set to Travis Scrimshaw
- Status changed from needs_review to positive_review
Okay, I understand. Thank you for the explanation. LGTM.
comment:5 Changed 13 months ago by
Thank you.
comment:6 Changed 13 months ago by
- Branch changed from u/gh-kliem/double_description_pyramid to 2b1c64b047628dfd5179cb98f637da57bb74452f
- Resolution set to fixed
- Status changed from positive_review to closed
Could there be a slowdown if we need to compute the number of inequalities, say if a polytope was given in terms of its vertices? Do we even want to bother checking if one or both of the representations have been done (and then choose accordingly)?
I will have the same comment on #32150 and #32151.