feat: reduce size of segments containing fixed-size-list arrays

[danking]

I also verified this on some fixed-size-list data I have locally and, indeed, I see more reasonable segment sizes (~1MiB).

[codspeed-hq]

CodSpeed Performance Report

Merging this PR will degrade performance by 20.01%

_{Comparing dk/better-segment-sizes-for-fixed-size-list (ad235e9) with develop (83664a0)}

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

Summary

⚡ 10 improved benchmarks
❌ 7 regressed benchmarks
✅ 1144 untouched benchmarks
⏩ 1323 skipped benchmarks¹

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	canonical_into_non_nullable[(10000, 1, 0.0)]	36.2 µs	24.9 µs	+45.52%
⚡	Simulation	canonical_into_non_nullable[(10000, 1, 0.01)]	41 µs	31.3 µs	+31.28%
⚡	Simulation	canonical_into_non_nullable[(10000, 10, 0.0)]	278.8 µs	194.3 µs	+43.52%
⚡	Simulation	canonical_into_non_nullable[(10000, 1, 0.1)]	57 µs	47.1 µs	+21.09%
⚡	Simulation	canonical_into_non_nullable[(10000, 10, 0.01)]	306 µs	221.3 µs	+38.31%
⚡	Simulation	canonical_into_non_nullable[(10000, 10, 0.1)]	471.6 µs	380.7 µs	+23.86%
❌	Simulation	canonical_into_nullable[(10000, 100, 0.0)]	4.3 ms	5 ms	-12.44%
⚡	Simulation	into_canonical_non_nullable[(10000, 10, 0.0)]	283.7 µs	201.7 µs	+40.66%
❌	Simulation	into_canonical_non_nullable[(10000, 1, 0.01)]	38.3 µs	46.3 µs	-17.25%
❌	Simulation	into_canonical_non_nullable[(10000, 1, 0.1)]	54.4 µs	62.8 µs	-13.37%
⚡	Simulation	into_canonical_non_nullable[(10000, 10, 0.1)]	473 µs	385.2 µs	+22.81%
⚡	Simulation	into_canonical_non_nullable[(10000, 10, 0.01)]	310.6 µs	229.2 µs	+35.47%
❌	Simulation	into_canonical_non_nullable[(10000, 1, 0.0)]	32.3 µs	40.4 µs	-20.01%
❌	Simulation	into_canonical_nullable[(10000, 100, 0.0)]	4.3 ms	5.1 ms	-14.45%
❌	Simulation	into_canonical_nullable[(10000, 10, 0.0)]	459.1 µs	541.3 µs	-15.2%
⚡	Simulation	into_canonical_nullable[(10000, 10, 0.1)]	720.2 µs	632.2 µs	+13.91%
❌	Simulation	into_canonical_nullable[(10000, 100, 0.1)]	6.1 ms	6.9 ms	-12.01%

Footnotes

1. 1323 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[lwwmanning]

this should help with large binary values too, no?

[connortsui20]

Very nice