apfloat Lib 2.41

Apfloat is a C++ arbitrary precision arithmetic package. Multiplications are done using Fast Fourier Transforms for O(n log n) complexity. The transforms are done as Number Theoretic Transforms to avoid round-off problems.

Three different moduli are used for optimal memory usage. The final result is achieved using the Chinese Remainder Theorem. The algorithms are optimized for very high precision (more than 100 000 digits). The package is written to be easily portable, but also includes assembler optimization in critical parts for various processors for maximum performance.

Completly freeware for non-commercial use.