40% speedup (for default settings) via more parallelism inside workers

[MilesCranmer]

Added parallelism to various loops over the populations.

[github-actions]

Benchmark Results

	master	a34f40a...	master/a34f40af0bca17...
search/multithreading	25.9 ± 1.2 s	17.6 ± 0.49 s	1.48
search/serial	32.2 ± 0.055 s	32.3 ± 0.045 s	0.995
utils/best_of_sample	0.781 ± 0.21 μs	0.742 ± 0.19 μs	1.05
utils/check_constraints_x10	11.4 ± 2.9 μs	11.4 ± 2.9 μs	1.01
utils/compute_complexity_x10/Float64	2.2 ± 0.11 μs	2.23 ± 0.12 μs	0.987
utils/compute_complexity_x10/Int64	2.19 ± 0.11 μs	2.2 ± 0.11 μs	0.995
utils/compute_complexity_x10/nothing	1.48 ± 0.09 μs	1.48 ± 0.091 μs	1
utils/optimize_constants_x10	30.1 ± 6.5 ms	30.6 ± 6.5 ms	0.984
time_to_load	0.885 ± 0.0017 s	0.88 ± 0.0037 s	1.01

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).