minterpy.jit_compiled

This sub-package collects all just-in-time compiled codes for Minterpy.

Minterpy accelerates numerous performance-critical internal functions through just-in-time (JIT) compilation using the Numba package.

Most JIT-compiled code is decorated with njit instead of jit. This means that the input and output requirements for the compiled functions are explicityly define in advance. While njit is less flexible than jit, it is typically faster and more memory efficient.

Warning

Input validation and error checking are generally not implemented within JIT-compiled functions. Therefore, it is the caller’s responsibility to ensure sure that the input and output requirements are met.

Module	Description
common	Common performance-sensitivity numerical routines
multi_index	Compiled numerical routines related to multi-indices
newton.diff	Compiled numerical routines to differentiate Newton polynomials
newton.eval	Compiled numerical routines to evaluate Newton polynomials
canonical	Compiled numerical routines related to the canonical basis
transformations	Compiled and common numerical routines for basis transformations