#include <source/patchdata/edge/EdgeData.h>
Inheritance diagram for SAMRAI::pdat::EdgeData< DIM, TYPE >:
Public Types | |
typedef EdgeIterator< DIM > | Iterator |
Public Member Functions | |
EdgeData (const hier::Box< DIM > &box, const int depth, const hier::IntVector< DIM > &ghosts, tbox::Pointer< tbox::Arena > pool=tbox::Pointer< tbox::Arena >((0))) | |
virtual | ~EdgeData () |
virtual void | copy (const hier::PatchData< DIM > &src) |
virtual void | copy2 (hier::PatchData< DIM > &dst) const |
virtual void | copy (const hier::PatchData< DIM > &src, const hier::BoxOverlap< DIM > &overlap) |
virtual void | copy2 (hier::PatchData< DIM > &dst, const hier::BoxOverlap< DIM > &overlap) const |
void | copyDepth (int dst_depth, const EdgeData< DIM, TYPE > &src, int src_depth) |
Fast copy between the source and destination at the specified depths. | |
virtual bool | canEstimateStreamSizeFromBox () const |
virtual int | getDataStreamSize (const hier::BoxOverlap< DIM > &overlap) const |
virtual void | packStream (tbox::AbstractStream &stream, const hier::BoxOverlap< DIM > &overlap) const |
virtual void | unpackStream (tbox::AbstractStream &stream, const hier::BoxOverlap< DIM > &overlap) |
int | getDepth () const |
TYPE * | getPointer (const int axis, const int d=0) |
const TYPE * | getPointer (const int axis, const int d=0) const |
TYPE & | operator() (const EdgeIndex< DIM > &i, const int d=0) |
const TYPE & | operator() (const EdgeIndex< DIM > &i, const int d=0) const |
ArrayData< DIM, TYPE > & | getArrayData (const int axis) |
const ArrayData< DIM, TYPE > & | getArrayData (const int axis) const |
void | fill (const TYPE &t, const int d=0) |
void | fill (const TYPE &t, const hier::Box< DIM > &box, const int d=0) |
void | fillAll (const TYPE &t) |
void | fillAll (const TYPE &t, const hier::Box< DIM > &box) |
void | copyOnBox (const EdgeData< DIM, TYPE > &src, const hier::Box< DIM > &box) |
void | print (const hier::Box< DIM > &box, ostream &os=tbox::plog, int prec=-1) const |
void | print (const hier::Box< DIM > &box, const int d, ostream &os=tbox::plog, int prec=-1) const |
void | printAxis (const int axis, const hier::Box< DIM > &box, ostream &os=tbox::plog, int prec=-1) const |
void | printAxis (const int axis, const hier::Box< DIM > &box, const int d, ostream &os=tbox::plog, int prec=-1) const |
virtual void | getSpecializedFromDatabase (tbox::Pointer< tbox::Database > database) |
virtual void | putSpecializedToDatabase (tbox::Pointer< tbox::Database > database) |
Static Public Member Functions | |
static size_t | getSizeOfData (const hier::Box< DIM > &box, const int depth, const hier::IntVector< DIM > &ghosts) |
Edge data is stored in DIM arrays, each of which contains the data for the edges tangent to a single coordinate direction. Memory allocation is in column-major ordering (e.g., Fortran style) so that the leftmost index runs fastest in memory. For example, a three-dimensional edge data object instantiated with a box [l0:u0,l1:u1,l2:u2] allocates three data arrays dimensioned as
[ l0 : u0 , l1 : u1+1 , l2 : u2+1 , d ] , [ l0 : u0+1 , l1 : u1 , l2 : u2+1 , d ] , [ l0 : u0+1 , l1 : u1+1 , l2 : u2 , d ] , *
The data type TYPE must define a default constructor (that takes no arguments) and also the assignment operator.
|
The edge iterator iterates over the elements on one axis of an edge centered box geometry. This typedef is a convenient link to the EdgeIterator<DIM> class. |
|
The constructor for an edge data object. The box describes the interior of the index space and the ghosts vector describes the ghost cells in each coordinate direction. The depth gives the number of components for each spatial location in the array. If the memory arena is not given, then the standard arena is used. |
|
The virtual destructor for an edge data object. |
|
A fast copy between the source and destination. Data is copied from the source into the destination where there is overlap in the underlying index space. The copy is performed on the interior plus the ghost cell width (for both the source and destination). If copy() does not understand the source object type, then copy2() is called. |
|
A fast copy between the source and destination. Member function copy2() is similar to copy() except that the location of source and destination are reversed and copy2() throws an exception (aka dumps core) if it does not understand the type of the argument. |
|
Copy data from the source into the destination using the designated overlap descriptor. The overlap description will have been computed using the appropriate box geometry objects. If copy() does not understand the source object type, then copy2() is called. |
|
Copy data from the source into the destination using the designated overlap descriptor. Member function copy2() is similar to the copy() member function except that the location of source and destination are reversed and copy2() throws an exception (aka dumps core) if it does not understand the type of the argument. |
|
Fast copy between the source and destination at the specified depths. Data is copied from the source into the destination where there is overlap in the underlying index space. The copy is performed on the interior plus the ghost cell width (for both the source and destination). |
|
Determines whether the patch data subclass can estinate the necessary stream size using only index space information. This routine is defined for the standard built-in types (bool, char, double, float, and int). Implements SAMRAI::hier::PatchData< DIM >. |
|
Calculate the number of bytes needed to stream the data lying in the specified box domain. This routine is defined for the standard built-in types (bool, char, double, float, and int). |
|
Pack data lying on the specified index set into the output stream. |
|
Unpack data from the message stream into the specified index set. |
|
Return the depth (e.g., the number of components in each spatial location) of the array. |
|
Get a pointer to the beginning of a particular component of the edge centered array. |
|
Get a const pointer to the beginning of a particular component of the edge centered array. |
|
hier::Index into the edge data array using an edge index. |
|
hier::Index into the edge data array (via a const reference) using an edge index. |
|
Retrieve the array data corresponding to the specified axis. |
|
Retrieve the const array data corresponding to the specified axis. |
|
Fill all values of component d with the value t. |
|
Fill all values of component d within the box with the value t. |
|
Fill all components with value t. |
|
Fill all components within the box with value t. |
|
Copy data from supplied source over the supplied box. |
|
Calculate the amount of memory needed to represent the data in a edge centered grid. This function assumes that the amount of memory needed for TYPE is sizeof(TYPE). If this is not the case, then a specialized function must be defined. |
|
Print all edge centered data residing in the specified box. If the depth of the array is greater than one, all components are printed. Precision of floating point numbers (i.e., TYPE = float, double, or dcomplex) can be specified using the provided argument. The default is 12 decimal places for double and complex floating point numbers, and the default is 6 decimal places floats. For other types, this is ignored. |
|
Print the specified component of the edge centered data residing in the specified box. Precision of floating point numbers (i.e., TYPE = float, double, or dcomplex) can be specified using the provided argument. The default is 12 decimal places for double and complex floating point numbers, and the default is 6 decimal places floats. For other types, this is ignored. |
|
Print all edge centered data for specified axis index residing in the specified box. If the depth of the data is greater than one, then all components are printed. Precision of floating point numbers (i.e., TYPE = float, double, or dcomplex) can be specified using the provided argument. The default is 12 decimal places for double and complex floating point numbers, and the default is 6 decimal places floats. For other types, this is ignored. |
|
Print specified component for all edge centered data for the specified axis index residing in the specified box. Precision of floating point numbers (i.e., TYPE = float, double, or dcomplex) can be specified using the provided argument. The default is 12 decimal places for double and complex floating point numbers, and the default is 6 decimal places floats. For other types, this is ignored. |
|
Check that class version and restart file version are equal. If so, read data members from the database. Assertions: database must be non-null pointer. Implements SAMRAI::hier::PatchData< DIM >. |
|
Write out the class version number and other data members to the database. Assertions: database must be non-null pointer. Implements SAMRAI::hier::PatchData< DIM >. |