You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In oscar-system/Oscar.jl#4581 I am trying to optimize conversion between GAP and OSCAR finite field elems (FFE), and one of the bottlenecks is the construction of FqFieldElem from a vector of integer coefficients (a Vector{Int}).
This ends up calling the following Nemo function:
function (a::FqField)(b::Vector{<:IntegerUnion})
da =degree(a)
db =length(b)
da == db ||error("Coercion impossible")
returna(parent(defining_polynomial(a))(b))
end
So this first constructs a polynomial from the Vector{Int} and then converts that into an FqFieldElem. It would be nice if this intermediate step could be skipped.
However I am not sure how feasible that is without "cheating" (i.e. doing different things depending on which underlying field implementation is used).
The text was updated successfully, but these errors were encountered:
In oscar-system/Oscar.jl#4581 I am trying to optimize conversion between GAP and OSCAR finite field elems (FFE), and one of the bottlenecks is the construction of
FqFieldElem
from a vector of integer coefficients (aVector{Int}
).This ends up calling the following Nemo function:
So this first constructs a polynomial from the
Vector{Int}
and then converts that into anFqFieldElem
. It would be nice if this intermediate step could be skipped.However I am not sure how feasible that is without "cheating" (i.e. doing different things depending on which underlying field implementation is used).
The text was updated successfully, but these errors were encountered: