BasisLieHighestWeight: Add kwarg compute_polytope to keep track of the vertices of the polytope of the monomial basis
#5597
Conversation
janikapeters
force-pushed
the
jp/polytope_in_blhw
branch
from
ce06556 to
f5307ea
Compare
| end | ||
|
|
||
| function reduce_to_vertices(possible_vertices::Vector{PointVector{QQFieldElem}}) | ||
| return collect(vertices(convex_hull(possible_vertices))) |
There was a problem hiding this comment.
I would guess that most of the additional time is spent in this repeated convex_hull and collect calls.
Could you try to instead of keeping a list of vertices to keep the polytope itself? This would also simplify some operations, as you can use the Minkowski sum directly
There was a problem hiding this comment.
Tried it out and in smaller cases it is a bit slower if we store the polytope, in a little bigger cases like :C,3,[1,6,0] where this version takes less than 0.5 seconds but if we store the polytope the process dies because it takes too much memory. So I guess we shouldn't do this.
Maybe we can try to not reduce everytime but still store the vertices and not the polytope?
| operators::Vector{RootSpaceElem}, # monomial x_i is corresponds to f_operators[i] | ||
| monomial_ordering_input::Union{AbsGenOrdering,Symbol}, | ||
| monomial_ordering_input::Union{AbsGenOrdering,Symbol}; | ||
| compute_polytope=false, |
There was a problem hiding this comment.
Please check if there is a performance difference between this and the use of a Val-type for the flag, see https://docs.julialang.org/en/v1/manual/types/#%22Value-types%22 for details (or ask me)
Co-authored-by: Lars Göttgens <[email protected]>
2 participants
No description provided.