Speed up dynamic degree orders

[gdalle]

• Store buckets in a contiguous vector with specified limits instead of a vector of vectors

[gdalle]

Benchmark full coloring

using SparseArrays, SparseMatrixColorings
using Chairmarks, StableRNGs

for n in (1000, 3000), p in (0.0003, 0.001, 0.003, 0.01)
    As = [sprand(StableRNG(i), Bool, n, n, p) for i in 1:10]
    problem = ColoringProblem()
    algo = GreedyColoringAlgorithm(DynamicLargestFirst())
    bench = @b fast_coloring.($As, $(Ref(problem)), $(Ref(algo))) seconds = 3
    @info "Benchmark: n=$n, p=$p"
    display(bench)
end

This branch:

[ Info: Benchmark: n=1000, p=0.0003
393.125 μs (362 allocs: 588.125 KiB)
[ Info: Benchmark: n=1000, p=0.001
656.292 μs (362 allocs: 645.938 KiB)
[ Info: Benchmark: n=1000, p=0.003
3.128 ms (362 allocs: 808.422 KiB)
[ Info: Benchmark: n=1000, p=0.01
19.676 ms (362 allocs: 1.362 MiB)
[ Info: Benchmark: n=3000, p=0.0003
2.776 ms (362 allocs: 1.834 MiB)
[ Info: Benchmark: n=3000, p=0.001
11.834 ms (362 allocs: 2.321 MiB)
[ Info: Benchmark: n=3000, p=0.003
55.021 ms (362 allocs: 3.726 MiB)
[ Info: Benchmark: n=3000, p=0.01
435.350 ms (392 allocs: 8.782 MiB)

Main branch:

[ Info: Benchmark: n=1000, p=0.0003
622.584 μs (583 allocs: 968.344 KiB)
[ Info: Benchmark: n=1000, p=0.001
1.244 ms (777 allocs: 996.750 KiB)
[ Info: Benchmark: n=1000, p=0.003
4.543 ms (1526 allocs: 1.380 MiB)
[ Info: Benchmark: n=1000, p=0.01
31.623 ms (5934 allocs: 2.639 MiB)
[ Info: Benchmark: n=3000, p=0.0003
4.389 ms (909 allocs: 2.717 MiB)
[ Info: Benchmark: n=3000, p=0.001
16.675 ms (1939 allocs: 3.789 MiB)
[ Info: Benchmark: n=3000, p=0.003
86.568 ms (6665 allocs: 6.467 MiB)
[ Info: Benchmark: n=3000, p=0.01
741.368 ms (34668 allocs: 14.959 MiB)

Benchmark just ordering

using SparseArrays, SparseMatrixColorings
using Chairmarks, StableRNGs

for n in (1000, 3000), p in (0.0003, 0.001, 0.003, 0.01)
    As = [sprand(StableRNG(i), Bool, n, n, p) for i in 1:10]
    gs = SparseMatrixColorings.BipartiteGraph.(As)
    order = DynamicLargestFirst()
    bench = @b SparseMatrixColorings.vertices.($gs, $(Ref(Val(2))), $(Ref(order))) seconds =
        3
    @info "Benchmark: n=$n, p=$p"
    display(bench)
end

This branch:

[ Info: Benchmark: n=1000, p=0.0003
321.125 μs (242 allocs: 324.875 KiB)
[ Info: Benchmark: n=1000, p=0.001
491.416 μs (242 allocs: 326.562 KiB)
[ Info: Benchmark: n=1000, p=0.003
2.114 ms (242 allocs: 332.422 KiB)
[ Info: Benchmark: n=1000, p=0.01
14.323 ms (242 allocs: 372.922 KiB)
[ Info: Benchmark: n=3000, p=0.0003
2.185 ms (242 allocs: 956.203 KiB)
[ Info: Benchmark: n=3000, p=0.001
8.269 ms (242 allocs: 963.234 KiB)
[ Info: Benchmark: n=3000, p=0.003
39.502 ms (242 allocs: 1001.156 KiB)
[ Info: Benchmark: n=3000, p=0.01
327.865 ms (272 allocs: 1.218 MiB)

Main branch:

[ Info: Benchmark: n=1000, p=0.0003
559.833 μs (463 allocs: 705.094 KiB)
[ Info: Benchmark: n=1000, p=0.001
1.090 ms (657 allocs: 677.375 KiB)
[ Info: Benchmark: n=1000, p=0.003
3.524 ms (1406 allocs: 936.641 KiB)
[ Info: Benchmark: n=1000, p=0.01
26.086 ms (5814 allocs: 1.640 MiB)
[ Info: Benchmark: n=3000, p=0.0003
3.841 ms (789 allocs: 1.817 MiB)
[ Info: Benchmark: n=3000, p=0.001
12.936 ms (1819 allocs: 2.409 MiB)
[ Info: Benchmark: n=3000, p=0.003
70.679 ms (6545 allocs: 3.718 MiB)
[ Info: Benchmark: n=3000, p=0.01
635.092 ms (34548 allocs: 7.395 MiB)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 100.00%. Comparing base (5c9b8c7) to head (997ca35).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##              main      #220   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           15        15           
  Lines         1642      1680   +38     
=========================================
+ Hits          1642      1680   +38     

☔ 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.

[amontoison]

All tests passed and you have good results in the benchmarks.
I merge the PR 👍