This class facilitates inline conversion of an input data vector to a different precision level, depending on the underlying type of Real and whether or not the single_precision flag is set. More...

#include <exodusII_io_helper.h>

Public Member Functions
	MappedOutputVector (const std::vector< Real > &vec_in, bool single_precision_in)

	~MappedOutputVector ()=default

void *	data ()

Private Attributes
const std::vector< Real > &	our_data

bool	single_precision

std::vector< double >	double_vec

std::vector< float >	float_vec

Detailed Description

This class facilitates inline conversion of an input data vector to a different precision level, depending on the underlying type of Real and whether or not the single_precision flag is set.

This should be used whenever floating point data is being written to the Exodus file. Note that if no precision conversion has to take place, there should be very little overhead involved in using this object.

Definition at line 751 of file exodusII_io_helper.h.

Constructor & Destructor Documentation

◆ MappedOutputVector()

libMesh::ExodusII_IO_Helper::MappedOutputVector::MappedOutputVector	(	const std::vector< Real > &	vec_in,
		bool	single_precision_in
	)

Definition at line 417 of file exodusII_io_helper.C.

   : our_data(our_data_in),
     single_precision(single_precision_in)
 {
   if (single_precision)
     {
       if (sizeof(Real) != sizeof(float))
         float_vec.assign(our_data.begin(), our_data.end());
     }
  
   else if (sizeof(Real) != sizeof(double))
     double_vec.assign(our_data.begin(), our_data.end());
 }

References double_vec, float_vec, our_data, libMesh::Real, and single_precision.

◆ ~MappedOutputVector()

libMesh::ExodusII_IO_Helper::MappedOutputVector::~MappedOutputVector ( )

default

Member Function Documentation

◆ data()

void * libMesh::ExodusII_IO_Helper::MappedOutputVector::data ( )

Definition at line 433 of file exodusII_io_helper.C.

 {
   if (single_precision)
     {
       if (sizeof(Real) != sizeof(float))
         return static_cast<void*>(float_vec.data());
     }
  
   else if (sizeof(Real) != sizeof(double))
     return static_cast<void*>(double_vec.data());
  
   // Otherwise return a (suitably casted) pointer to the original underlying data.
   return const_cast<void *>(static_cast<const void *>(our_data.data()));
 }