fix(matrix): use bitvec-rs instead of hand-crafted bit manipulation

[samueltardieu]

For performance measurements only, not to merge yet.

[codspeed-hq]

CodSpeed Performance Report

Merging #596 will degrade performances by 45.61%

_{Comparing bitvec-rs (c091b8e) with main (e99545d)}

Summary

❌ 1 regressions
✅ 35 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Benchmark	main	bitvec-rs	Change
❌	transpose_non_square	160.3 µs	294.6 µs	-45.61%

[samueltardieu]

Bad performance, abandoning