panzer::Expr Namespace Reference

Contains all symbols which support panzer::Expr::Eval. More...

Classes
class	EvalBase
	Base class for panzer::Expr::Eval, does everything that is independent of the Kokkos::View template parameter. More...
struct	RebindDataType
	Rebinds a Kokkos::View data type to use a new scalar type. More...
struct	RebindDataType< NestedDataType *, NewScalarType >
struct	RebindDataType< NestedDataType[], NewScalarType >
struct	RebindDataType< NestedDataType[N], NewScalarType >
struct	RebindViewType
	Builds on RebindDataType, but acts directly on a Kokkos::View type. More...
struct	RebindViewType< Kokkos::View< DT, VP...>, NewScalarType >
class	Eval
	Interprets mathematical expressions in a string and evaluates them using Kokkos::View objects as values and Kokkos::parallel_for for the operators. This class is mean to support Kokkos-parallel execution of user-provided mathematical expressions. Example uses include evaluating analytic boundary and initial conditions for a PDE problem. The API of this class (namely Eval::read_string or another read_* function inherited from Teuchos::Reader) is only meant to be called once for a given set of evaluation points. Variables should be predefined with point-dependent values all stored in a single Kokkos::View. For example, X coordinates for all points could be stored in a Kokkos::View<double*>, whose extent is the number of points. Values which are the same for all points are still supported, for example the current time. Then if the expression is for example "x^t", this class launches a single parallel_for to raise the coordinate of all points at once to the "t" power. More...
struct	ScalarTernary
struct	ScalarOr
struct	ScalarAnd
struct	ScalarGT
struct	ScalarLT
struct	ScalarGEQ
struct	ScalarLEQ
struct	ScalarEQ
struct	ScalarAdd
struct	ScalarSub
struct	ScalarMul
struct	ScalarDiv
struct	ScalarPow
struct	ScalarNeg
struct	Indexer
struct	Indexer< ViewType, 1, 0 >
struct	Indexer< ViewType, 1, 1 >
struct	Indexer< ViewType, 2, 0 >
struct	Indexer< ViewType, 2, 1 >
struct	Indexer< ViewType, 2, 2 >
struct	MaxRank
struct	MaxRank< T >
struct	ResultType
struct	TernaryResultType
struct	BinaryFunctor
struct	BinaryFunctor< Op, Result, Left, Right, 0 >
struct	BinaryFunctor< Op, Result, Left, Right, 1 >
struct	BinaryFunctor< Op, Result, Left, Right, 2 >
struct	TernaryFunctor
struct	TernaryFunctor< Cond, Left, Right, 1 >
struct	TernaryFunctor< Cond, Left, Right, 2 >
struct	UnaryFunctor
struct	UnaryFunctor< Op, Result, 0 >
struct	UnaryFunctor< Op, Result, 1 >
struct	UnaryFunctor< Op, Result, 2 >
struct	ScalarAbs
struct	ScalarExp
struct	ScalarLog
struct	ScalarSqrt
struct	ScalarSin
struct	ScalarCos
struct	ScalarTan
struct	UnaryFunction

Enumerations
enum	BinaryOpCode {   BinaryOpCode::OR, BinaryOpCode::AND, BinaryOpCode::GT, BinaryOpCode::LT,   BinaryOpCode::GEQ, BinaryOpCode::LEQ, BinaryOpCode::EQ, BinaryOpCode::ADD,   BinaryOpCode::SUB, BinaryOpCode::MUL, BinaryOpCode::DIV, BinaryOpCode::POW }
	Denotes the native binary operators in the Teuchos::MathExpr language. More...

Functions
static const char *	get_op_syntax (BinaryOpCode code)
template<typename DT , typename... VP>
void	set_cmath_functions (Eval< DT, VP...> &eval)
	Add support for functions such as sqrt(), sin(), and cos() More...

Detailed Description

Contains all symbols which support panzer::Expr::Eval.

Enumeration Type Documentation

enum panzer::Expr::BinaryOpCode

strong

Denotes the native binary operators in the Teuchos::MathExpr language.

Enumerator
OR
AND
GT
LT
GEQ
LEQ
EQ
ADD
SUB
MUL
DIV
POW

Definition at line 71 of file Panzer_ExprEval.hpp.

Function Documentation

static const char* panzer::Expr::get_op_syntax ( BinaryOpCode  code )

static

Definition at line 228 of file Panzer_ExprEval.cpp.

template<typename DT , typename... VP>

void panzer::Expr::set_cmath_functions

(

Eval< DT, VP...> &

eval

)

Add support for functions such as sqrt(), sin(), and cos()

After this call, (eval) is able to interpret strings which include calls to these functions:

• abs(x)
• exp(x)
• log(x)
• sqrt(x)
• sin(x)
• cos(x)
• tan(x)

Note

This function is kept separate in case users want to use Expr::Eval with a scalar type that only supports basic operators but doesn't support these higher-level math functions (an integer for example).

Definition at line 709 of file Panzer_ExprEval_impl.hpp.