Better specialization

[MilesCranmer]

PR to see if the lack of OperatorEnum specialization was what caused the 10% performance regression.

[github-actions]

Benchmark Results

	master	1124700...	master/11247009f97ddc...
search/multithreading	26.4 ± 1.4 s	26.2 ± 0.76 s	1.01
search/serial	33 ± 0.2 s	32.3 ± 0.47 s	1.02
utils/best_of_sample	0.802 ± 0.25 μs	0.872 ± 0.29 μs	0.92
utils/check_constraints_x10	11.3 ± 2.9 μs	11.3 ± 2.9 μs	0.999
utils/compute_complexity_x10/Float64	2.21 ± 0.11 μs	2.19 ± 0.11 μs	1.01
utils/compute_complexity_x10/Int64	2.25 ± 0.11 μs	2.19 ± 0.11 μs	1.03
utils/compute_complexity_x10/nothing	1.47 ± 0.09 μs	1.37 ± 0.11 μs	1.07
utils/optimize_constants_x10	31.2 ± 6.7 ms	30.6 ± 7.1 ms	1.02
time_to_load	0.906 ± 0.0043 s	0.912 ± 0.0075 s	0.994

Benchmark Plots

A plot of the benchmark results have been uploaded as an artifact to the workflow run for this PR.
Go to "Actions"->"Benchmark a pull request"->[the most recent run]->"Artifacts" (at the bottom).

[MilesCranmer]

Nope; doesn't seem to help at all...