set gfp_mat or nmod_mat entry with Int value

[simonbrandhorst]

Before

julia> F = Hecke.Native.GF(17)
Finite field of characteristic 17

julia> A = F[1 2; 3 4];

julia> @btime A[1,1] = UInt(2);
  19.783 ns (0 allocations: 0 bytes)

julia> @btime A[1,1] = 2;
  44.078 ns (2 allocations: 32 bytes)

After

julia> @btime A[1,1] = 2;
  25.585 ns (0 allocations: 0 bytes)

Maybe not optimal, but in any case it is an improvement.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.30%. Comparing base (3a96b0c) to head (3b29c7e).
⚠️ Report is 2 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2301      +/-   ##
==========================================
- Coverage   88.36%   88.30%   -0.07%     
==========================================
  Files         102      102              
  Lines       37510    37526      +16     
==========================================
- Hits        33147    33136      -11     
- Misses       4363     4390      +27     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Tobias271828]

That's a good idea – optimising the code! Thanks!

I don’t quite understand your code, though: what do you need the function _nmod_reduce_int for? The method mod(::Int, ::UInt) returns a value >= 0 of type UInt. Hence, it should be enough to just paste this code into nmod_mat.jl:

@inline function setindex!(a::T, u::Int, i::Int, j::Int) where T <: Zmodn_mat
  @boundscheck _checkbounds(a, i, j)
  R = base_ring(a)
  setindex_raw!(a, mod(u, R.n), i, j)
end

It works on my machine, and I obtained this benchmarking result by using it.

julia> F = Native.GF(17)
Finite field of characteristic 17

julia> A = F[1 2; 3 4];

julia> @b A[1,1] = UInt(2)
15.700 ns

julia> @b A[1,1] = 2
14.675 ns

[simonbrandhorst]

You are of course right. Thanks!