v0.6.0
v0.6.0 - 2026-05-08
Performance
- Slice-by-8 CRC-32 Algorithm: Upgraded CRC-32 from Slice-by-4 to Slice-by-8 by adding four new pre-computed lookup tables (
crct4–crct7). The inner loop now processes 8 bytes per iteration instead of 4, with 8 independent table lookups per step, yielding a 36.6% throughput improvement on CRC-32 operations.