Skip to main content

Crate tpt_math_exact

Crate tpt_math_exact 

Source
Expand description

Arbitrary-precision rational arithmetic and interval arithmetic.

This crate thin-wraps num_bigint and num_rational to expose exact rational numbers (BigRational), and layers an Interval type on top for rigorous (rounding-error-free) bounds. Because the bounds are exact rationals, interval arithmetic here never accumulates floating-point error.

§Features

  • std (default) — enable the std feature of the wrapped crates.
  • alloc — signal an allocator is available (no_std + alloc).

Re-exports§

pub use certify::certified_solve;
pub use certify::CertifyError;
pub use tpt_math_numeric as numeric;

Modules§

certify
Certified linear solves by the Krawczyk operator.
roots
Error-bounded root isolation for scalar functions, on exact intervals.

Structs§

BigInt
A big signed integer type.
BigUint
A big unsigned integer type.
Interval
A closed interval [lo, hi] over an ordered, arithmetic type.

Type Aliases§

BigRational
Alias for arbitrary precision rationals.
Rational
Exact rational number type alias.