A free high-performance arithmetic library designed for computational applications, offering fast and efficient mathematical operations for complex algorithms.
If you're looking for C++ arbitrary precision arithmetic package, then you might want to check out Apfloat. The package handles multiplications via Fast Fourier Transforms that are specifically designed for O(n log n) complexity. These transforms are done using Number Theoretic Transforms to escape round-off problems.
Another advantage is the use of three different modules that are optimized for memory usage. The final result is achieved through the Chinese Remainder Theorem. The system is also designed to optimize algorithms for very high precision (more than 100,000 digits).
Apfloat is written to be easily portable, but it also includes specific assembler optimization in critical parts for various processors to offer maximum performance. It's definitely worth paying attention to since it's also completely freeware for non-commercial use.
Another advantage is the use of three different modules that are optimized for memory usage. The final result is achieved through the Chinese Remainder Theorem. The system is also designed to optimize algorithms for very high precision (more than 100,000 digits).
Apfloat is written to be easily portable, but it also includes specific assembler optimization in critical parts for various processors to offer maximum performance. It's definitely worth paying attention to since it's also completely freeware for non-commercial use.
What's New