bigfloat

Arbitrary-precision floating-point library.

Written by Bob Jenkins (public domain). Rust translation by Claude in src/.

Documentation for the original C++: https://burtleburtle.net/bob/math/bigfloat.html

Precision

By default 20 base-2^32 digits (~193 decimal digits). Recompile with C_DIGITS and C_LOG adjusted for more or less precision.

Operations

Arithmetic: add, sub, mul, div, sqrt, power, power_int

Transcendental: exp, ln, log, sin, cos, tan, csc, sec, a_sin, a_cos, a_tan

Constants: pi(), e_const()

Random: rand (uniform), rand_norm (normal distribution)

Linear algebra: gaussian_elimination

Conversions: to_double, to_integer, to_fraction, print_decimal, print_hex, print_continued_fraction

Special values: ±0, ±inf, NaN

Example

use bigfloat::BigFloat;

// Solve 2×2 system via Gaussian elimination
let mut m = vec![
    vec![BigFloat::from_int(1), BigFloat::from_int(1), BigFloat::from_int(2)],
    vec![BigFloat::from_int(1), BigFloat::from_int(2), BigFloat::from_int(3)],
];
BigFloat::gaussian_elimination(&mut m, 2, 2);

println!("{}", BigFloat::pi().to_double());   // π to ~193 decimal digits

Building

cargo build --release
cargo test --features bigfloat-test