fpmath

fpmath is a pure-Rust floating point library that implements math functions for f32 and f64.

Features

Less than 1 ULP error in most functions.
f32 versions do not internally use f64 arithmetic.
All functions are also implemented for soft-float single-precision and double precision.
no_std
All functions are tested for accuracy (MPFR is used to calculate reference values).
The included generator crate can generate all magic constants used in the algorithms.

Included functions

Absolute value
Copy sign
Rounding (to nearest, towards zero, towards infinity, towards negative infinity)
Exponential in base e, 2 and 10
Logarithm in base e, 2 and 10
Trigonometric (sine, cosine, tangent) in radians, degrees and half-revolutions
Inverse trigonometric (arcsine, arccosine, arctangent) in radians, degrees and half-revolutions
Hyperbolic (sine, cosine, tangent)
Inverse hyperbolic (arcsine, arccosine, arctangent)
Square and cube root
Hypotenuse
Power (floating point and integer exponent)

License

Licensed under either of

Apache License, Version 2.0 (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT)

at your option.