#include <source/patchdata/array/ArrayData.h>
Inheritance diagram for SAMRAI::pdat::ArrayData< DIM, TYPE >:
Public Types | |
| typedef ArrayDataIterator< DIM > | Iterator |
Public Member Functions | |
| ArrayData () | |
| ArrayData (const hier::Box< DIM > &box, const int depth, tbox::Pointer< tbox::Arena > pool=tbox::Pointer< tbox::Arena >((0))) | |
| ~ArrayData () | |
| const hier::Box< DIM > & | getBox () const |
| void | initializeArray (const hier::Box< DIM > &box, const int depth, const tbox::Pointer< tbox::Arena > &pool) |
| bool | isInitialized () const |
| Returns whether array has been properly initialized proper depth and box. | |
| void | undefineData () |
| bool | isStandardType () |
| void | copy (const ArrayData< DIM, TYPE > &src, const hier::Box< DIM > &box) |
| void | copy (const ArrayData< DIM, TYPE > &src, const hier::Box< DIM > &box, const hier::IntVector< DIM > &offset) |
| void | copy (const ArrayData< DIM, TYPE > &src, const hier::BoxList< DIM > &boxes, const hier::IntVector< DIM > &offset) |
| void | copyDepth (int dst_depth, const ArrayData< DIM, TYPE > &src, int src_depth, const hier::Box< DIM > &box) |
| Copy one depth of a source array to a destination depth of another. | |
| int | getDataStreamSize (const hier::BoxList< DIM > &dest_boxes, const hier::IntVector< DIM > &source_offset) const |
| void | packStream (tbox::AbstractStream &stream, const hier::Box< DIM > &dest_box, const hier::IntVector< DIM > &source_offset) const |
| void | packStream (tbox::AbstractStream &stream, const hier::BoxList< DIM > &dest_boxes, const hier::IntVector< DIM > &source_offset) const |
| void | unpackStream (tbox::AbstractStream &stream, const hier::Box< DIM > &dest_box, const hier::IntVector< DIM > &source_offset) |
| void | unpackStream (tbox::AbstractStream &stream, const hier::BoxList< DIM > &boxes, const hier::IntVector< DIM > &source_offset) |
| int | getDepth () const |
| int | getOffset () const |
| TYPE * | getPointer (const int d=0) |
| const TYPE * | getPointer (const int d=0) const |
| TYPE & | operator() (const hier::Index< DIM > &i, const int d) |
| const TYPE & | operator() (const hier::Index< DIM > &i, const int d) const |
| void | fillAll (const TYPE &t) |
| void | fillAll (const TYPE &t, const hier::Box< DIM > &box) |
| void | fill (const TYPE &t, const int d=0) |
| void | fill (const TYPE &t, const hier::Box< DIM > &box, const int d=0) |
| void | getFromDatabase (tbox::Pointer< tbox::Database > database) |
| void | putToDatabase (tbox::Pointer< tbox::Database > database, bool data_only=false) |
| void | getSpecializedFromDatabase (tbox::Pointer< tbox::Database > database) |
| void | putSpecializedToDatabase (tbox::Pointer< tbox::Database > database) |
Static Public Member Functions | |
| static size_t | getSizeOfData (const hier::Box< DIM > &box, const int depth) |
| static bool | canEstimateStreamSizeFromBox () |
The data storage is in (i,...,k,d) order, where i,...,k indicates spatial indices and the d indicates the component at that location. Memory allocation is in column-major ordering (e.g., Fortran style) so that the leftmost index runs fastest in memory.
The data type TYPE must define a default constructor (that takes no arguments) and also the assignment operator. Note that a number of functions only work for standard built-in types (bool, char, double, float, and int). To use this class with other user-defined types, many of these functions will need to be specialized, especially those that deal with message packing and unpacking.
|
|||||
|
The array data iterator iterates over the elements of a box associated with an ArrayData object. This typedef is convenient link to the ArrayDataIterator<DIM> class. |
|
|||||||||
|
The no-arguments constructor creates an empty array data object. The initializeArray() member function must be called before the array can be used. |
|
||||||||||||||||||||
|
The constructor for an array data object. The box describes the spatial extents of the index space and the depth gives the number of data components for each spatial location in the array. tbox::Array memory is allocated from the specified memory pool. |
|
|||||||||
|
The destructor for an array data object releases all memory allocated for the array elements. |
|
|||||||||
|
Return the box over which the array is defined. |
|
||||||||||||||||||||
|
Initialize the array data for the specified box and depth. This routine is normally called to initialize an array that was created by the no-arguments constructor. |
|
|||||||||
|
Returns whether array has been properly initialized proper depth and box. Only initialized arrays can be used. Uninitialized arrays should not be used until initializeArray() is called. |
|
|||||||||
|
Set the array data to some ``undefined'' state. For floats and doubles, this means setting data to signaling NaNs that cause a floating point exception when used in a numerical expression. |
|
||||||||||||||||
|
Return the amount of memory space needed to allocate a box of the specified size and depth. This function will need to be redefined for data types other than the standard bool, char, double, float, and int types. |
|
|||||||||
|
Return true if the TYPE is a standard fundamental data type |
|
||||||||||||||||
|
Copy data from the source array data on the specified index domain. This routine will intersect the specified box against the source and destination boxes to find the region of intersection. |
|
||||||||||||||||||||
|
Copy data shifted in the index space from the source into the destination. The boxes are in the destination index space, and the source data is taken from the boxes shifted into the source index space. This routine will intersect the specified box against the source and destination boxes to find the region of intersection. |
|
||||||||||||||||||||
|
Copy data shifted in the index space from the source into the destination. The boxes are in the destination index space, and the source data is taken from the boxes shifted into the source index space. |
|
||||||||||||||||||||||||
|
Copy one depth of a source array to a destination depth of another. This routine will intersect the specified box against the source and destination boxes to find the region of intersection. |
|
|||||||||
|
Return whether the amount of buffer space in the message stream can be estimated using the box alone (i.e., without specific type information). For built-in types (bool, char, double, float, int, and dcomplex), this routine returns true. For other user-defined data types that may require special handling, the user MUST define a different implementation. |
|
||||||||||||||||
|
Calculate the number of bytes needed to stream the data lying in the specified box domains. This routine is only defined for the built-in types of bool, char, double, float, and int. For all other types, the user must define a specialized implementation. |
|
||||||||||||||||||||
|
Pack data lying on the specified index set into the output stream. The shifted box must lie completely within the index space of the array. If compiled with assertions enabled, the packing routine will abort if the box is not contained within the index space of the array. |
|
||||||||||||||||||||
|
Pack data lying on the specified index set into the output stream. The shifted boxes in the box list must lie completely within the index space of the array. If compiled with assertions enabled, the packing routine will abort if the boxes are not contained within the index space of the array. |
|
||||||||||||||||||||
|
Unpack data from the message stream into the index set lying under the specified index set. The box must lie completely within the index space of the array. If compiled with assertions enabled, the unpacking routine will abort if the box is not contained within the index space of the array. |
|
||||||||||||||||||||
|
Unpack data from the message stream into the index set lying under the specified index set. The boxes in the box list must lie completely within the index space of the array. If compiled with assertions enabled, the unpacking routine will abort if the boxes are not contained within the index space of the array. |
|
|||||||||
|
Return the depth (e.g., the number of components in each spatial location) of the array. |
|
|||||||||
|
Return the offset (e.g., the number of data values for each depth component) of the array. |
|
||||||||||
|
Get a pointer to the beginning of a particular component of the patch data array. |
|
||||||||||
|
Get a const pointer to the beginning of a particular component of the patch data array. |
|
||||||||||||||||
|
hier::Index into the array using an index and the component. |
|
||||||||||||||||
|
hier::Index into the array (via a const reference) using an index and the component. |
|
||||||||||
|
Fill all components with value t. |
|
||||||||||||||||
|
Fill all components within the box with value t. |
|
||||||||||||||||
|
Fill all values of component d with the value t. |
|
||||||||||||||||||||
|
Fill all values of component d within the box with the value t. |
|
||||||||||
|
Check to make sure that the class version and restart file version are equal. If so, read in data from database. This routine calls getSpecializedFromDatabase() to read in the proper data type. Assertions: database must be a non-null pointer. |
|
||||||||||||||||
|
Writes out data for class representation to database. This routine calls putSpecializedToDatabase() to read in the proper data type. The default behavior (boolean argument is false) is to put all data members in database. Otherwise, only the array contents are written out. Assertions: database must be a non-null pointer. |
|
||||||||||
|
Use specialized template method to get the correct behavior when reading in the array of data. |
|
||||||||||
|
Use specialized template method to get the correct behavior when writing out the array of data. |
1.4.2