Use inlined rand from sampler for selecting level

[Tortar]

Let's see if this has an effect, maybe since benchmarks in JuliaLang/julia#57771 (comment)

[github-actions]

Benchmark Results

	ca5d92d	b806003	ca5d92d / b806003
TTFX excluding time to load	8.6 ± 0 ms	8.69 ± 0 ms	1.03,0.976,0.99
code size in bytes	1.33e+04 ± 0 h	1.33e+04 ± 0 h	0.997
code size in lines	459 ± 0 h	459 ± 0 h	1
code size in syntax nodes	3.44e+03 ± 0 h	3.45e+03 ± 0 h	0.997
constructor n=100 σ=0.1	6.26 ± 0.53 μs	6.01 ± 0.34 μs	0.974,1.03,1.04
constructor n=100 σ=1.0	6.47 ± 0.28 μs	6.38 ± 0.31 μs	0.965,1.07,1.01
constructor n=100 σ=10.0	7.02 ± 0.36 μs	6.76 ± 0.46 μs	0.99,1.02,1.04
constructor n=100 σ=100.0	6.39 ± 0.5 μs	6.36 ± 1.2 μs	1.03,0.989,1
constructor n=1000 σ=0.1	0.0435 ± 0.0018 ms	0.0466 ± 0.0033 ms	1,1.02,0.933
constructor n=1000 σ=1.0	0.0455 ± 0.0013 ms	0.046 ± 0.002 ms	1,1.02,0.988
constructor n=1000 σ=10.0	0.0526 ± 0.0016 ms	0.0522 ± 0.0025 ms	1.02,1.03,1.01
constructor n=1000 σ=100.0	0.06 ± 0.0024 ms	0.0594 ± 0.0012 ms	1,0.98,1.01
constructor n=10000 σ=0.1	0.453 ± 0.036 ms	0.432 ± 0.029 ms	1.01,1.01,1.05
constructor n=10000 σ=1.0	0.468 ± 0.045 ms	0.44 ± 0.025 ms	1,1.11,1.06
constructor n=10000 σ=10.0	0.468 ± 0.14 ms	0.483 ± 0.033 ms	1,1.02,0.969
constructor n=10000 σ=100.0	0.609 ± 0.032 ms	0.583 ± 0.022 ms	1.01,1.01,1.05
delete ∘ rand n=100 σ=0.1	4.72 ± 0.19 μs	4.55 ± 0.2 μs	1.04
delete ∘ rand n=100 σ=1.0	5.01 ± 0.2 μs	4.85 ± 0.19 μs	1.04,1.04,1.03
delete ∘ rand n=100 σ=10.0	5.2 ± 0.19 μs	5.02 ± 0.19 μs	1.03,1.03,1.04
delete ∘ rand n=100 σ=100.0	9.29 ± 0.75 μs	9.01 ± 0.76 μs	1.03
delete ∘ rand n=1000 σ=0.1	0.0468 ± 0.00087 ms	0.0451 ± 0.00079 ms	1.04
delete ∘ rand n=1000 σ=1.0	0.0511 ± 0.0011 ms	0.0494 ± 0.00093 ms	1.04
delete ∘ rand n=1000 σ=10.0	0.0524 ± 0.0011 ms	0.0506 ± 0.00083 ms	1.04,1.03,1.04
delete ∘ rand n=1000 σ=100.0	0.0611 ± 0.0013 ms	0.0596 ± 0.0014 ms	1.03
delete ∘ rand n=10000 σ=0.1	0.508 ± 0.012 ms	0.495 ± 0.0095 ms	1.03
delete ∘ rand n=10000 σ=1.0	0.549 ± 0.014 ms	0.536 ± 0.014 ms	1.03,1.04,1.02
delete ∘ rand n=10000 σ=10.0	0.554 ± 0.02 ms	0.537 ± 0.014 ms	1.03
delete ∘ rand n=10000 σ=100.0	0.558 ± 0.025 ms	0.536 ± 0.013 ms	1.03,1.04,1.04
empty constructor	2.05 ± 0.49 μs	1.89 ± 0.33 μs	2.66,3.23,1.09
intermixed_h n=100 σ=0.1	11.7 ± 0.96 μs	11.4 ± 0.99 μs	1.02,1.04,1.02
intermixed_h n=100 σ=1.0	12.2 ± 1.1 μs	11.7 ± 1 μs	1.01,1.06,1.04
intermixed_h n=100 σ=10.0	11.9 ± 1.8 μs	11.8 ± 2.1 μs	1.03,0.989,1.01
intermixed_h n=100 σ=100.0	13.4 ± 1.7 μs	14 ± 1.9 μs	0.987,1.02,0.954
intermixed_h n=1000 σ=0.1	0.11 ± 0.0094 ms	0.107 ± 0.0089 ms	1.02,1.01,1.03
intermixed_h n=1000 σ=1.0	0.114 ± 0.0084 ms	0.112 ± 0.0093 ms	1.02
intermixed_h n=1000 σ=10.0	0.108 ± 0.0094 ms	0.112 ± 0.0086 ms	1.03,1.08,0.966
intermixed_h n=1000 σ=100.0	0.132 ± 0.02 ms	0.119 ± 0.014 ms	1.02,1.03,1.11
intermixed_h n=10000 σ=0.1	1.24 ± 0.22 ms	1.21 ± 0.24 ms	1.1,1.01,1.02
intermixed_h n=10000 σ=1.0	1.2 ± 0.18 ms	1.17 ± 0.16 ms	1.04,1.03,1.03
intermixed_h n=10000 σ=10.0	1.11 ± 0.17 ms	1.14 ± 0.21 ms	1.03,0.981,0.97
intermixed_h n=10000 σ=100.0	1.16 ± 0.21 ms	1.19 ± 0.18 ms	1.05,1.02,0.978
pathological 1	0.0456 ± 0.00029 μs	0.0453 ± 0.00019 μs	1.01,1.02,1.01
pathological 1′	0.208 ± 0.0021 μs	0.201 ± 0.0017 μs	1.03,1.01,1.04
pathological 2	0.0629 ± 0.00028 μs	0.0623 ± 0.00032 μs	1.01
pathological 2′	0.183 ± 0.0015 μs	0.183 ± 0.0015 μs	0.997,1.02,1
pathological 3	18.7 ± 0.29 ns	17.2 ± 0.23 ns	1.08,1.08,1.09
pathological 4	0.0626 ± 0.00019 μs	0.0617 ± 0.00028 μs	1.01,1,1.01
pathological 4′	0.235 ± 0.0018 μs	0.227 ± 0.0017 μs	1.02,1.1,1.04
pathological 5a	0.0448 ± 0.00028 μs	0.0444 ± 0.00024 μs	1.1,1.01,1.01
pathological 5b	0.0448 ± 0.00026 μs	0.0442 ± 0.0003 μs	1.01
pathological 5b′	0.468 ± 0.0068 μs	0.461 ± 0.011 μs	1.02,1.01,1.02
sample n=100 σ=0.1	28 ± 0.66 ns	26.5 ± 0.67 ns	1.06,1.07,1.06
sample n=100 σ=1.0	0.0325 ± 0.0023 μs	31.4 ± 2.3 ns	1.04,1.05,1.03
sample n=100 σ=10.0	21.4 ± 5.3 ns	19.9 ± 4.6 ns	1.09,1.07,1.07
sample n=100 σ=100.0	18.6 ± 4.5 ns	17.2 ± 4 ns	1.07,1.11,1.08
sample n=1000 σ=0.1	26.6 ± 5.6 ns	24.4 ± 4.4 ns	1.06,1.06,1.09
sample n=1000 σ=1.0	0.0354 ± 0.0025 μs	0.0341 ± 0.0022 μs	1.05,1.04,1.04
sample n=1000 σ=10.0	22.3 ± 6.2 ns	20.7 ± 5.4 ns	1.04,1.08,1.08
sample n=1000 σ=100.0	18.5 ± 4.3 ns	17.1 ± 4 ns	1.09,1.08,1.08
sample n=10000 σ=0.1	0.0345 ± 0.0013 μs	0.0336 ± 0.0018 μs	1.06,1.06,1.03
sample n=10000 σ=1.0	0.0403 ± 0.0045 μs	0.0369 ± 0.0029 μs	1.05,1.05,1.09
sample n=10000 σ=10.0	23.5 ± 6.1 ns	22.1 ± 6.6 ns	1.1,1.03,1.06
sample n=10000 σ=100.0	19.3 ± 4.2 ns	17.3 ± 3.6 ns	1.06,1.07,1.12
summarysize n=100 σ=0.1	1.13e+05 ± 0 h	1.13e+05 ± 0 h	1
summarysize n=100 σ=1.0	1.13e+05 ± 0 h	1.13e+05 ± 0 h	1
summarysize n=100 σ=10.0	1.13e+05 ± 0 h	1.13e+05 ± 0 h	1
summarysize n=100 σ=100.0	1.13e+05 ± 0 h	1.13e+05 ± 0 h	1
summarysize n=1000 σ=0.1	1.42e+05 ± 0 h	1.42e+05 ± 0 h	1
summarysize n=1000 σ=1.0	1.42e+05 ± 0 h	1.42e+05 ± 0 h	1
summarysize n=1000 σ=10.0	1.42e+05 ± 0 h	1.42e+05 ± 0 h	1
summarysize n=1000 σ=100.0	1.42e+05 ± 0 h	1.42e+05 ± 0 h	1
summarysize n=10000 σ=0.1	1e+06 ± 0 h	1e+06 ± 0 h	1
summarysize n=10000 σ=1.0	1e+06 ± 0 h	1e+06 ± 0 h	1
summarysize n=10000 σ=10.0	1e+06 ± 0 h	1e+06 ± 0 h	1
summarysize n=10000 σ=100.0	1e+06 ± 0 h	1e+06 ± 0 h	1
update ∘ rand n=100 σ=0.1	0.0865 ± 0.0023 μs	0.0836 ± 0.0022 μs	1.04,1.04,1.03
update ∘ rand n=100 σ=1.0	0.0928 ± 0.0029 μs	0.0901 ± 0.0029 μs	1.03
update ∘ rand n=100 σ=10.0	0.0999 ± 0.0037 μs	0.0979 ± 0.0039 μs	1.03,1.02,1.02
update ∘ rand n=100 σ=100.0	0.228 ± 0.026 μs	0.222 ± 0.023 μs	1.02,1.02,1.03
update ∘ rand n=1000 σ=0.1	0.0872 ± 0.0024 μs	0.0841 ± 0.0024 μs	1.04
update ∘ rand n=1000 σ=1.0	0.0937 ± 0.0036 μs	0.0909 ± 0.0029 μs	1.03
update ∘ rand n=1000 σ=10.0	0.0992 ± 0.002 μs	0.098 ± 0.0043 μs	1.03,1.03,1.01
update ∘ rand n=1000 σ=100.0	0.222 ± 0.011 μs	0.22 ± 0.0095 μs	1,0.995,1.01
update ∘ rand n=10000 σ=0.1	0.0957 ± 0.0018 μs	0.092 ± 0.0011 μs	1.04,1.03,1.04
update ∘ rand n=10000 σ=1.0	0.0988 ± 0.00087 μs	0.0965 ± 0.0018 μs	1.03,1.04,1.02
update ∘ rand n=10000 σ=10.0	0.0985 ± 0.0016 μs	0.0969 ± 0.0023 μs	1.04,1.04,1.02
update ∘ rand n=10000 σ=100.0	0.198 ± 0.0045 μs	0.199 ± 0.0028 μs	0.995,0.994,0.997
time_to_load	0.075 ± 0.0017 s	0.0738 ± 0.0021 s	0.966,0.999,1.02

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).

[Tortar]

It does