amrex::ParserExecutor< N > Struct Template Reference

Callable wrapper around a compiled parser expression with N variables. More...

#include <AMReX_Parser.H>

Public Member Functions
__host__ __device__ double	operator() () const noexcept
	Evaluate the expression with no arguments.
template<typename... Ts> requires (N > 0 && sizeof...(Ts) == N && !amrex::Same<float,Ts...>::value)
__host__ __device__ double	operator() (Ts... var) const noexcept
	Evaluate the expression with N double-compatible arguments.
template<std::same_as< float >... Ts> requires (N > 0 && sizeof...(Ts) == N)
__host__ __device__ float	operator() (Ts... var) const noexcept
	Evaluate the expression with N float arguments.
__host__ __device__ double	operator() (GpuArray< double, N > const &var) const noexcept
	Evaluate the expression using an explicit array of inputs.
__host__ __device__	operator bool () const
	True when a compiled executor buffer is available.

Public Attributes
char *	m_host_executor = nullptr
	Pointer to host-visible bytecode.
char *	m_device_executor = nullptr
	Pointer to device-visible bytecode (if copied).

Detailed Description

template<int N>
struct amrex::ParserExecutor< N >

Callable wrapper around a compiled parser expression with N variables.

Instances are returned by Parser::compile() / Parser::compileHost() and may be invoked on host or device. Device execution is available only when AMREX_USE_GPU is enabled and a device executor buffer has been created.

Template Parameters

N	Number of variables referenced in the expression.

Member Function Documentation

operator bool()

template<int N>

__host__ __device__ amrex::ParserExecutor< N >::operator bool ( ) const

inlineexplicit

True when a compiled executor buffer is available.

operator()() [1/4]

template<int N>

__host__ __device__ double amrex::ParserExecutor< N >::operator() ( ) const

inlinenoexcept

Evaluate the expression with no arguments.

Enabled only for constant expressions (N == 0).

Returns

Expression value as double precision.

operator()() [2/4]

template<int N>

__host__ __device__ double amrex::ParserExecutor< N >::operator() ( GpuArray< double, N > const &  var ) const

inlinenoexcept

Evaluate the expression using an explicit array of inputs.

Parameters

var	Array storing the N variables in parser variable order.

Returns

Expression value promoted to double precision.

operator()() [3/4]

template<int N>

template<typename... Ts>
requires (N > 0 && sizeof...(Ts) == N && !amrex::Same<float,Ts...>::value)

__host__ __device__ double amrex::ParserExecutor< N >::operator() ( Ts...  var ) const

inlinenoexcept

Evaluate the expression with N double-compatible arguments.

Parameters

var	Scalar arguments supplied in parser variable order.

Returns

Expression value promoted to double precision.

operator()() [4/4]

template<int N>

template<std::same_as< float >... Ts>
requires (N > 0 && sizeof...(Ts) == N)

__host__ __device__ float amrex::ParserExecutor< N >::operator() ( Ts...  var ) const

inlinenoexcept

Evaluate the expression with N float arguments.

Returns a float when all inputs are float.

Parameters

var	Scalar arguments supplied in parser variable order.

Member Data Documentation

m_device_executor

template<int N>

char* amrex::ParserExecutor< N >::m_device_executor = nullptr

Pointer to device-visible bytecode (if copied).

m_host_executor

template<int N>

char* amrex::ParserExecutor< N >::m_host_executor = nullptr

Pointer to host-visible bytecode.

---

The documentation for this struct was generated from the following files:

• /home/runner/work/amrex/amrex/Src/Base/AMReX_BaseFwd.H
• /home/runner/work/amrex/amrex/Src/Base/Parser/AMReX_Parser.H