Tommy Hofmann

Results 702 comments of Tommy Hofmann

hm, where did that ``` function Base.setprecision(a::Generic.MatSpaceElem{QadicFieldElem}, N::Int) b = deepcopy(a) setprecision!(b, N) return b end ``` move to @joschmitt?

Yes and no. Try using `Native.GF(2)` (once Nemo is loaded).

Hm, yes, it is quite bad, but a bit puzzling. @fieker: The AbstractAlgebra code is not that sophisticated in comparison to Nemo/flint, which goes through `n_mulmod_preinv`. Is this just the...

I tried 30 bits and got the same timings. We could in fact just copy the code from AA to Nemo.

@albinahlback since you interested in flint being fast, do you have an idea what could be going on? How long does ```julia julia> @btime begin K = Native.GF(1099511627791) a =...

Hm, I tried to produce the assembly code, but I noticed something strange. It seems that there is something dodgy going one due to some problems in the benchmark setup....

It depends a bit on the definition, but most functions related to factoring of fmpz_poly are mathematically wrong. They always give results for the polynomial considered as an element of...

@fredrik-johansson, do you have an idea what is happening here? This is just calling `arb_root`: ``` julia> RR = ArbField(64) Real Field with 64 bits of precision and error bounds...

Thanks! I will probably do an `is_zero` check in Nemo for now.

Hm, the check in Nemo is not quite correct. At the moment it is ``` x < 0 && throw(DomainError("x must be positive")) ``` but it should be ``` is_zero(x)...