Teuchos::RawMPITraits< T > Class Template Reference

Templated traits class that allows a datatype to be used with MPI that MPI can directly handle. More...

#include <Teuchos_RawMPITraits.hpp>

List of all members.

Static Public Member Functions
int	adjustCount (const int count)
	Return the adjusted cout of items.
MPI_Datatype	type ()
	Return the raw MPI data type of the template argument.
MPI_Op	sumOp ()
	Return the MPI_Op object for a sum reduction.
MPI_Op	maxOp ()
	Return the MPI_Op object for a max reduction.
MPI_Op	minOp ()
	Return the MPI_Op object for a min reduction.

---

Detailed Description

template<class T>
class Teuchos::RawMPITraits< T >

Templated traits class that allows a datatype to be used with MPI that MPI can directly handle.

A specialization of this traits class should only be created for datatypes that can be directly handled by MPI in some way. Note that this traits class assumes that the datatype T is directly composed of datatypes that MPI can directly handle. This traits interface allows for specializations to create user-defined MPI_Datatype and MPI_Op objects to be returned from their static functions.

Note:

• This class should not compile if it is instantiated by accident.
• It should only be included if MPI is available and the MPI header must be included before this header file.
• Template specializations exist for datatypes: char, int, float, and double.
• A partial template specialization exists for std::complex<T> where it is assumed that the real type T is directly handlable with MPI.
• Only sum reductions are supported for all data types.
• The reductions max and min are only supported by datatypes whee ScalarTraits<T>::isComparable==true which is a compile-time boolean that can be used in template metaprogramming techniques.

Definition at line 68 of file Teuchos_RawMPITraits.hpp.

---

Member Function Documentation

template<class T> int Teuchos::RawMPITraits< T >::adjustCount (  const int  count  )  [inline, static]

Return the adjusted cout of items. Definition at line 71 of file Teuchos_RawMPITraits.hpp.

template<class T> MPI_Datatype Teuchos::RawMPITraits< T >::type (   )  [inline, static]

Return the raw MPI data type of the template argument. Definition at line 73 of file Teuchos_RawMPITraits.hpp.

template<class T> MPI_Op Teuchos::RawMPITraits< T >::sumOp (   )  [inline, static]

Return the MPI_Op object for a sum reduction. Definition at line 75 of file Teuchos_RawMPITraits.hpp.

template<class T> MPI_Op Teuchos::RawMPITraits< T >::maxOp (   )  [inline, static]

Return the MPI_Op object for a max reduction. Definition at line 77 of file Teuchos_RawMPITraits.hpp.

template<class T> MPI_Op Teuchos::RawMPITraits< T >::minOp (   )  [inline, static]

Return the MPI_Op object for a min reduction. Definition at line 79 of file Teuchos_RawMPITraits.hpp.

---

The documentation for this class was generated from the following file:

• Teuchos_RawMPITraits.hpp

---