Benchmark RowFn filtered fallback

[connortsui20]

Summary

• Tracking Issue: Tracking Issue: Execute RowFn over Vortex arrays #9130
• Comparison target: Replace filter-and-scatter RowFn fallback with filtered valid-row execution #9645

Adds a focused baseline for the RowFn fallback under comparison. FilteredI64 declares DENSE_SAFE = false, so partially valid inputs select valid-only execution, and it declines null-tolerant decoding, so direct valid-row execution is unavailable. The remaining path is filter-and-scatter on develop and filtered-row execution with #9645.

Changes

Benchmarks owned and sink outputs with all-valid, one-null-in-eight, and nine-nulls-in-ten inputs. The partial-validity cases force the fallback, while the all-valid cases bypass it as unchanged controls. All cases run on the native AVX2, AVX-512, and NEON CodSpeed legs.

[codspeed-hq]

Merging this PR will regress 2 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.

⚠️ 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

⚡ 3 improved benchmarks
❌ 2 regressed benchmarks
✅ 2089 untouched benchmarks
🆕 18 new benchmarks
⏩ 206 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	random_i16[0.8]	76.5 µs	94.4 µs	-18.99%
❌	Simulation	cold_misaligned[(16, 64)]	346.8 µs	392.4 µs	-11.61%
⚡	WallTime	arrow_checked_add_u32_neon[16384]	20.4 µs	12.9 µs	+58.94%
⚡	WallTime	words_gather_scalar_avx2[65536]	9.4 µs	8.3 µs	+13.32%
⚡	Simulation	compress_fsst[(500, 64, 8)]	533.5 µs	484.6 µs	+10.1%
🆕	WallTime	filtered_owned_i64_neon[AllValid]	N/A	4.6 µs	N/A
🆕	WallTime	filtered_owned_i64_neon[NineNullsInTen]	N/A	25.8 µs	N/A
🆕	WallTime	filtered_owned_i64_neon[OneNullInEight]	N/A	47.3 µs	N/A
🆕	WallTime	filtered_sink_i64_neon[AllValid]	N/A	4.8 µs	N/A
🆕	WallTime	filtered_sink_i64_neon[NineNullsInTen]	N/A	25.6 µs	N/A
🆕	WallTime	filtered_sink_i64_neon[OneNullInEight]	N/A	47.2 µs	N/A
🆕	WallTime	filtered_owned_i64_avx512[AllValid]	N/A	5.4 µs	N/A
🆕	WallTime	filtered_owned_i64_avx512[NineNullsInTen]	N/A	37.6 µs	N/A
🆕	WallTime	filtered_owned_i64_avx512[OneNullInEight]	N/A	56.5 µs	N/A
🆕	WallTime	filtered_sink_i64_avx512[AllValid]	N/A	5.3 µs	N/A
🆕	WallTime	filtered_sink_i64_avx512[NineNullsInTen]	N/A	37.6 µs	N/A
🆕	WallTime	filtered_sink_i64_avx512[OneNullInEight]	N/A	56.3 µs	N/A
🆕	WallTime	filtered_owned_i64_avx2[AllValid]	N/A	5.5 µs	N/A
🆕	WallTime	filtered_owned_i64_avx2[NineNullsInTen]	N/A	37.4 µs	N/A
🆕	WallTime	filtered_owned_i64_avx2[OneNullInEight]	N/A	56.3 µs	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 ct/row-fn-filtered-fallback-benchmark (f5b5a0b) with develop (21cfecc)}

Footnotes

1. 206 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. ↩