Perform fixed width array operations in type width units instead of byte units

[robert3005]

Instead of doing all offset calculation in byte units we do them in type T units
which is less errorprone

---

_{Stack created with GitHub Stacks CLI • Give Feedback 💬}

[codspeed-hq]

Merging this PR will degrade performance by 11.12%

⚠️ 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
❌ 6 regressed benchmarks
✅ 2188 untouched benchmarks
⏩ 218 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	WallTime	dict_canonicalize_gt_u8_avx512[1000000]	419.4 µs	766.7 µs	-45.3%
❌	WallTime	dict_canonicalize_gt_u8_avx512[16000000]	6.8 ms	10.4 ms	-35.07%
❌	Simulation	random_i256[0.8]	292.8 µs	332.4 µs	-11.93%
❌	WallTime	mul_u32_nonnull_avx512	5.6 µs	6.4 µs	-11.47%
❌	Simulation	random_i256[0.5]	204.7 µs	230.2 µs	-11.11%
❌	WallTime	mul_i32_nonnull_avx512	7.1 µs	8 µs	-10.27%
⚡	WallTime	arrow_checked_add_u32_avx2[16384]	21.3 µs	17.6 µs	+20.84%
⚡	WallTime	filtered_owned_i64_avx512[OneNullInEight]	26.1 µs	22.5 µs	+16.45%
⚡	WallTime	dict_canonicalize_gt_u8_neon[1000000]	542.4 µs	487.1 µs	+11.37%

Tip

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

---

_{Comparing rk/fixedwidtht (4ff30dc) with rk/uninitwriteslice (2720cef)}

Footnotes

1. 218 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. ↩