Hoist Slots

[gatesn]

Supersedes #7240

[codspeed-hq]

Merging this PR will degrade performance by 85.55%

⚡ 14 improved benchmarks
❌ 5 regressed benchmarks
✅ 1103 untouched benchmarks
⏩ 1530 skipped benchmarks¹

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

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	chunked_bool_canonical_into[(10, 1000)]	1,098.4 µs	945.8 µs	+16.13%
⚡	Simulation	chunked_opt_bool_canonical_into[(10, 1000)]	1.3 ms	1.2 ms	+11.7%
⚡	Simulation	chunked_bool_canonical_into[(100, 100)]	134.8 µs	119 µs	+13.24%
⚡	Simulation	chunked_opt_bool_into_canonical[(10, 1000)]	1.5 ms	1.4 ms	+11.88%
❌	Simulation	bench_many_nulls[0.1]	161.2 µs	419.6 µs	-61.59%
❌	Simulation	bench_many_nulls[0.01]	47.7 µs	329.9 µs	-85.55%
❌	Simulation	bench_many_nulls[0.5]	336 µs	492.6 µs	-31.8%
⚡	Simulation	filter_all_true[250000]	10 µs	8.3 µs	+19.67%
⚡	Simulation	filter_all_true[100000]	10 µs	8.3 µs	+19.67%
⚡	Simulation	varbinview_zip_block_mask	3.7 ms	2.9 ms	+27.76%
⚡	Simulation	varbinview_zip_fragmented_mask	7.1 ms	6.4 ms	+11.92%
⚡	Simulation	old_alp_prim_test_between[f32, 32768]	273.1 µs	223.6 µs	+22.15%
❌	Simulation	new_bp_prim_test_between[i64, 32768]	180.1 µs	225.4 µs	-20.08%
❌	Simulation	patched_take_10k_first_chunk_only	274.4 µs	331.5 µs	-17.23%
⚡	Simulation	take_10_contiguous	27.2 µs	24.7 µs	+10.04%
⚡	Simulation	new_alp_prim_test_between[f32, 32768]	187.2 µs	170 µs	+10.1%
⚡	Simulation	decompress_rd[f32, (10000, 0.1)]	93.1 µs	80.4 µs	+15.81%
⚡	Simulation	decompress_rd[f32, (10000, 0.0)]	137.3 µs	123.6 µs	+11.06%
⚡	Simulation	decompress_rd[f32, (10000, 0.01)]	93.1 µs	79.8 µs	+16.76%

---

_{Comparing ngates/slots (931c1a3) with develop (b2a5a70)}

Footnotes

1. 1530 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. ↩