Re-use allocations in mutation loop

[MilesCranmer]

This re-uses allocations in the mutation loop. This should hopefully result in a nice speed up from reduced need for garbage collection.

[github-actions]

Benchmark Results

	master	19eeb11...	master/19eeb111d0cbbb...
search/multithreading	17.2 ± 0.94 s	18.7 ± 0.35 s	0.921
search/serial	29 ± 0.36 s	29.5 ± 0.46 s	0.983
utils/best_of_sample	1.61 ± 0.43 μs	1.73 ± 1.3 μs	0.93
utils/check_constraints_x10	11.3 ± 3 μs	11.1 ± 2.8 μs	1.02
utils/compute_complexity_x10/Float64	2.06 ± 0.11 μs	2.09 ± 0.12 μs	0.986
utils/compute_complexity_x10/Int64	2 ± 0.11 μs	2.03 ± 0.11 μs	0.985
utils/compute_complexity_x10/nothing	1.43 ± 0.14 μs	1.46 ± 0.13 μs	0.979
utils/insert_random_op_x10	5.66 ± 1.8 μs	5.74 ± 1.9 μs	0.986
utils/next_generation_x100	0.343 ± 0.093 ms	0.428 ± 0.097 ms	0.801
utils/optimize_constants_x10	0.0367 ± 0.0085 s	0.0351 ± 0.0082 s	1.04
utils/randomly_rotate_tree_x10	5.27 ± 0.59 μs	5.18 ± 0.56 μs	1.02
time_to_load	1.98 ± 0.016 s	2.02 ± 0.024 s	0.98

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).

[coveralls]

Pull Request Test Coverage Report for Build 12323432167

Details

• 7 of 7 (100.0%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.01%) to 95.173%

---

Totals
Change from base Build 12271094085:	0.01%
Covered Lines:	3135
Relevant Lines:	3294

---

💛 - Coveralls

[MilesCranmer]

Hm. There’s no speedup… not sure it’s worth the added complexity?

Edit: made the interface have a fallback, it seems ok now. My guess is that the perf gains will be on larger systems with larger trees where the extra copies hurt