Hi! I was searching for some fully compile-time libraries to evaluate polynomials, and while I found some examples for simple use-cases, I did not see any support for complex multivariable equations. For example, there was no compile-time support to evaluate an equation like this:

3.0 * w * sin(x) * y² * cos(z) +
-1.2 * w³ * tan(x) * exp(y) * z +
0.7 * ln(w) * sqrt(x) * atan(y) * sinh(z) +
1.1 * cosh(w) * x * y * sin(z)

With this motivation, I built const_poly, a crate that lets you evaluate any multivariable equation or polynomial at compile time with high accuracy and zero runtime overhead.

Goals:

• no_std compatible – no heap allocations, no panics.

• Full compile-time evaluation of arbitrarily complex equations with high numerical accuracy (benchmarked at 1e-7).

• Fully documented with code examples, user-friendly macros, benchmarking, and a comprehensive suite of tests.

• Define expressions using variety of mathematical functions, all evaluable at compile time.

Who is this for?

• This library is primarily meant to empower scientific computing and mathematical libraries in rust to perform all numerical approximations entirely at compile time.

• Embedded and no_std environments where heapless, panic-free code is essential.

• Metaprogramming and symbolic math tools that benefit from evaluating complex expressions entirely at compile time.

I love to hear your feedback. Please let me know what you think!

github: GitHub - kmolan/const_poly: Compile-time evaluation of multivariable polynomials in rust

crate: crates.io: Rust Package Registry

1 post - 1 participant

Read full topic