perf: branchless boolean zip kernel

[joseph-isaacs]

Summary

Adds a dedicated, branchless ZipKernel for Bool. Booleans are bit-packed, so selecting if_true where the mask is set and if_false where it is not is a single bitwise blend over the packed words — (true & mask) | (false & !mask) — instead of the generic per-run builder (which degrades to per-element work on fragmented masks). The blend is mask-shape-independent.

It also introduces a shared zip_validity helper in the zip module: the result validity is itself a zip over the two boolean validity bitmaps, so it's built as a (lazy) zip array reusing this kernel. This gives the per-encoding zip kernels one shared validity-selection path, and the recursion terminates immediately because validity bitmaps are always Bool(NonNullable).

This lands first; the primitive and listview branchless zip kernels build on it (their nullable validity selection becomes a fast bool zip rather than the slow generic builder).

Changes

• vortex-array/src/arrays/bool/compute/zip.rs (new): the kernel.
• vortex-array/src/arrays/bool/compute/mod.rs, .../vtable/kernel.rs: register it.
• vortex-array/src/scalar_fn/fns/zip/mod.rs: shared pub(crate) fn zip_validity.
• vortex-array/benches/bool_zip.rs (new): a small divan bench.

Performance (divan, 65,536 bools, median)

case	time
nonnull	~7 µs
nullable	~14 µs

Testing

• Unit tests for non-nullable and Validity::Array inputs spanning the 64-bit mask chunk boundary + remainder.
• Full vortex-array lib suite passes (2933 tests); cargo +nightly fmt and clippy -D warnings (default + all-features) clean.

https://claude.ai/code/session_01N5ivPiCJy7dGQjMEP7ips9

---

Generated by Claude Code

[codspeed-hq]

Merging this PR will not alter performance

⚠️ 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 improved benchmarks
❌ 6 regressed benchmarks
✅ 1503 untouched benchmarks
🆕 2 new benchmarks

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	chunked_bool_canonical_into[(1000, 10)]	31.6 µs	46.6 µs	-32.16%
❌	Simulation	chunked_varbinview_canonical_into[(1000, 10)]	161.9 µs	198.2 µs	-18.33%
❌	Simulation	compare[15]	119.9 µs	145.8 µs	-17.78%
❌	Simulation	chunked_varbinview_into_canonical[(1000, 10)]	177.2 µs	213.5 µs	-17.02%
❌	Simulation	compare[14]	117.5 µs	141.5 µs	-16.97%
❌	Simulation	compare[13]	115.5 µs	137.7 µs	-16.1%
⚡	Simulation	bitwise_not_vortex_buffer_mut[128]	275.3 ns	216.9 ns	+26.89%
⚡	Simulation	bitwise_not_vortex_buffer_mut[1024]	336.9 ns	278.6 ns	+20.94%
⚡	Simulation	bitwise_not_vortex_buffer_mut[2048]	400.6 ns	342.2 ns	+17.05%
⚡	Simulation	compare[5]	76.9 µs	69.2 µs	+11.16%
🆕	Simulation	nonnull	N/A	77.5 µs	N/A
🆕	Simulation	nullable	N/A	135.4 µs	N/A

Tip

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

---

_{Comparing claude/bool-branchless-zip (7ad4b18) with develop (e06d80b)}