Compiled functions

#include <heyoka/expression.hpp>

The cfunc class

template<typename T>
class cfunc

Compiled function.

This class allows to compile just-in-time symbolic functions defined via the expression system. The type T is the floating-point type which will be used for the numerical evaluation of the function.

A tutorial showcasing the use of this class is available.

using in_1d = mdspan<const T, dextents<std::size_t, 1>>

using out_1d = mdspan<T, dextents<std::size_t, 1>>

using in_2d = mdspan<const T, dextents<std::size_t, 2>>

using out_2d = mdspan<T, dextents<std::size_t, 2>>

Views over 1D and 2D input/output arrays.

These are convenience typedefs for dynamically-sized 1D and 2D array views. They are used as input and output arguments in the call operators.

cfunc() noexcept

Default constructor.

The default constructor inits the compiled function in an invalid state in which the only supported operations are copy/move construction/assignment, destruction and the invocation of the is_valid() function.

template<typename ...KwArgs>
explicit cfunc(std::vector<expression> fn, std::vector<expression> vars, const KwArgs&... kw_args)

Main constructor.

This constructor will create a compiled function for the evaluation of the multivariate symbolic vector function fn. The vars argument is a list of variable expressions representing the order in which the inputs of fn are passed to the call operator during evaluation.

[[nodiscard]] bool is_valid() const noexcept