Optimize Dataset sort()

[andrewdalpino]

Dataset::sort() is an O(n²) bubble sort — confirmed at src/Datasets/Dataset.php:441-459. Switch to native usort() (Timsort, O(n log n), C-implemented). 10–100× for n > ~1000.

n	Before (bubble, O(n²))	After (usort, O(n log n))	Speedup
100	0.765 ms ±1.2%	0.384 ms ±1.3%	2.0×
1,000	44.343 ms ±0.7%	2.961 ms ±1.0%	15.0×
2,500	195.562 ms ±1.0%	8.288 ms ±1.2%	23.6×
5,000	759.693 ms ±0.7%	17.776 ms ±1.2%	42.7×

Targetting 3.0 branch as this changes the operation from an exact sort to an unstable sort.

[Copilot]

Pull request overview

This PR replaces Dataset::sort()’s O(n²) bubble sort with PHP’s native usort()-based approach to improve sorting performance on large datasets, and updates the dataset sorting benchmark to cover multiple dataset sizes.

Changes:

• Replaced the manual bubble sort in Rubix\ML\Datasets\Dataset::sort() with a usort() comparator wrapper.
• Updated SortingBench to benchmark multiple dataset sizes via PhpBench parameter providers.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
src/Datasets/Dataset.php	Switches dataset sorting implementation from bubble sort to usort() for O(n log n) behavior.
benchmarks/Datasets/SortingBench.php	Expands the benchmark to run sorting across multiple dataset sizes using param providers.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[apphp]

@andrewdalpino do we need this file in repo: runtime/test-results ?
What do you thing if we add it to .gitignore?