ulbn: High-precision Number Library

English 简体中文

Features

• C89/C++98 compatible (optional C++20 headers available)
• Minimal hardware assumptions
• Avoids most UB (Undefined Behavior)
• Strict boundary checks
• Custom memory allocator support
• Most functions are optimized
• No external dependencies

Dependencies

No external dependencies needed.

ulbn.h/ulbn.c

Minimum C89/C++98, with macros to suggest optimizations for some code.

Hardware Assumptions:

• Requires CHAR_BIT to be even

ulbn.hpp

Requires the following C++20 features:

• Concepts and constranints
• Three-way comparison
• std::endian
• std::span
• std::format (optional)

According to cppreference, the minimum compiler requirements is:

• GCC 10 (May, 7, 2020)
• Clang 10 (March, 24, 2020)
• MSVC 19.26 (May, 19, 2020)

test

Requires the following assumptions for modern platforms:

• Signed integers are stored in two's complement form
• float has 23 bits of precision (IEEE754)
• double has 52 bits of precision (IEEE754)
• char is 8-bits
• int64_t exists
• The platform is little-endian or big-endian
• <bit>
• std::format (optional)

Roadmap

• High-precision integers
  • Arithmetic
    • Faster multiplicatin
      • Karatsuba algorithm (Toom-2)
      • Toom-3 algorithm
      • Toom-4 algorithm
      • FFT
    • Faster base conversion
      • Faster input
      • Faster output
  • Comparison
  • Bitwise Operation
  • Root
  • Number theory functions
    • GCD/LCM
    • Extended GCD
    • Prime number determination
    • Prime number search
    • Factorization
• High-precision fractions
• High-precision floating point number
• High-precision decimal floating point number