Expand description
Fixed-point arithmetic for deterministic computation
All operations use 16.16 fixed-point format (16 bits integer, 16 bits fractional) to ensure bit-exact reproducibility across platforms (x86, ARM, WASM).
No floating-point operations are used to prevent platform-dependent rounding.
Structsยง
- Fixed
- 16.16 fixed-point number (16 bits integer, 16 bits fractional)