bench: compare uniform and per-chunk bitpacking

[mhk197]

Add a benchmark sweep comparing global-width and per-chunk encoding and decoding across uniform, alternating, drifting, and outlier-heavy distributions. Report compressed sizes alongside timing to make compression and runtime tradeoffs visible.

Size accounting includes the offsets child for v2 arrays and excludes it for uniform arrays, whose v1 wire format stores only the scalar width.

Validation: 435 FastLanes/BtrBlocks tests passed (1 skipped). Focused Clippy passed with all targets, all features, and warnings denied. FastLanes doctests and workspace formatting checks passed. CUDA runtime tests were not run. The focused u32 decoding benchmark completed across uniform and varying-width cases (40 samples, 64 iterations per sample).

[codspeed]

Merging this PR will regress 3 benchmarks

⚠️ 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.

⚠️ 4 benchmarks measured no execution time

Nothing ran under measurement, usually because the compiler removed the code under test. These results are not comparable, so they count as unchanged.

Preventing compiler optimizations

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 4 improved benchmarks
❌ 3 regressed benchmarks
✅ 2225 untouched benchmarks
🆕 160 new benchmarks
⏩ 323 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	take_fsl_f16_random[256, 100]	85.4 µs	141.1 µs	-39.52%
❌	Simulation	take_chunked_fsl_sorted[32, 64]	196 µs	252.4 µs	-22.34%
❌	Simulation	take_fsl_random[64, 100]	124.9 µs	143.8 µs	-13.16%
⚡	Simulation	take_fsl_u8_random[256, 100]	98.6 µs	45.8 µs	×2.2
⚡	WallTime	filtered_sink_i64_neon[NineNullsInTen]	19.6 µs	17.6 µs	+11.8%
⚡	WallTime	filtered_owned_i64_neon[NineNullsInTen]	19.5 µs	17.6 µs	+10.92%
⚡	WallTime	scalar_subtract_neon	13.4 µs	12.1 µs	+10.62%
🆕	Simulation	compress_v1[u16, drift]	N/A	291.2 µs	N/A
🆕	Simulation	compress_v1[u16, drift+exc1%]	N/A	829.2 µs	N/A
🆕	Simulation	compress_v1[u16, drift+null10%]	N/A	519.3 µs	N/A
🆕	Simulation	compress_v1[u16, random]	N/A	297.4 µs	N/A
🆕	Simulation	compress_v1[u16, spiky]	N/A	884.5 µs	N/A
🆕	Simulation	compress_v1[u16, spiky+exc1%]	N/A	891.4 µs	N/A
🆕	Simulation	compress_v1[u16, uniform]	N/A	267.5 µs	N/A
🆕	Simulation	compress_v1[u16, uniform+exc1%]	N/A	810.9 µs	N/A
🆕	Simulation	compress_v1[u16, uniform+null10%]	N/A	496.9 µs	N/A
🆕	Simulation	compress_v1[u16, zero_heavy]	N/A	267.9 µs	N/A
🆕	Simulation	compress_v1[u32, drift]	N/A	427.4 µs	N/A
🆕	Simulation	compress_v1[u32, drift+exc1%]	N/A	988.9 µs	N/A
🆕	Simulation	compress_v1[u32, drift+null10%]	N/A	1.5 ms	N/A
...	...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing mk/bitpacked-stack-09-benchmarks (5c86d38) with mk/bitpacked-stack-08-fused-encoder (34b2162)²}

Footnotes

1. 323 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. ↩

2. No successful run was found on mk/bitpacked-stack-08-fused-encoder (ab1d8e1) during the generation of this report, so 86f78ee was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩