|
Tpetra parallel linear algebra
Version of the Day
|
Nonowning view of a set of degrees of freedom corresponding to a mesh point in a block vector or multivector. More...
#include <Tpetra_Experimental_BlockView.hpp>
Public Member Functions | |
| LittleVector (Scalar *const A, const LO blockSize, const LO stride) | |
| Constructor. More... | |
| template<class T > | |
| LittleVector (T *const A, const LO blockSize, const LO stride, typename std::enable_if< !std::is_same< Scalar, T >::value &&std::is_convertible< Scalar, T >::value &&sizeof(Scalar)==sizeof(T), int * >::type ignoreMe=NULL) | |
Constructor that takes an impl_scalar_type pointer. More... | |
| Scalar * | getRawPtr () const |
| Pointer to the block's entries. More... | |
| LO | getBlockSize () const |
| The block size (number of degrees of freedom per mesh point). More... | |
| LO | getStride () const |
| Stride between consecutive entries. More... | |
| impl_scalar_type & | operator() (const LO i) const |
| Reference to entry (i) of the vector. More... | |
| template<class LittleVectorType > | |
| void | update (const Scalar &alpha, const LittleVectorType &X) const |
*this := *this + alpha * X. More... | |
| template<class LittleVectorType > | |
| void | assign (const LittleVectorType &X) const |
*this := X. More... | |
| void | scale (const Scalar &alpha) const |
(*this)(i,j) := alpha * (*this)(i,j) for all (i,j). More... | |
| void | fill (const Scalar &alpha) const |
(*this)(i,j) := alpha for all (i,j). More... | |
| template<class LittleVectorType > | |
| void | absmax (const LittleVectorType &X) const |
(*this)(i,j) := max(abs((*this)(i,j)), abs(X(i,j))) for all (i,j). More... | |
| template<class LittleVectorType > | |
| bool | equal (const LittleVectorType &X) const |
| true if and only if all entries of this equal all entries of X. More... | |
| template<class LittleBlockType , class LittleVectorType > | |
| void | matvecUpdate (const Scalar &alpha, const LittleBlockType &A, const LittleVectorType &X) const |
(*this) := (*this) + alpha * A * X (matrix-vector multiply). More... | |
Nonowning view of a set of degrees of freedom corresponding to a mesh point in a block vector or multivector.
| Scalar | The type of entries. |
| LO | The type of local indices. See the documentation of the first template parameter of Map for requirements. |
"Little" means local (not distributed over multiple MPI processes; stored to maximize locality) and small (think length 3, not length 1000).
The Scalar template parameter may be const or nonconst. This is one reason why instance methods below that take a LittleVector accept it as a template parameter: that lets you add a const LittleVector (e.g., LittleVector<const double, int>) to a nonconst LittleVector (e.g., LittleVector<double, int>).
Definition at line 325 of file Tpetra_Experimental_BlockView.hpp.
|
inline |
Constructor.
| A | [in] Pointer to the vector's entries |
| blockSize | [in] Dimension of the vector |
| stride | [in] Stride between consecutive entries |
Definition at line 338 of file Tpetra_Experimental_BlockView.hpp.
|
inline |
Constructor that takes an impl_scalar_type pointer.
| A | [in] Pointer to the vector's entries, as impl_scalar_type* rather than Scalar* |
| blockSize | [in] Dimension of the vector |
| stride | [in] Stride between consecutive entries |
While this constructor is templated on a type T, the intent is that T == impl_scalar_type. (We must template on T rather than using impl_scalar_type directly, because of how std::enable_if works.) The long, complicated std::enable_if expression ensures that this constructor only exists if Scalar differs from impl_scalar_type, but the two types are mutually compatible and have the same size. (They must be bitwise compatible, so that reinterpret_cast makes sense between them.)
Definition at line 361 of file Tpetra_Experimental_BlockView.hpp.
|
inline |
Pointer to the block's entries.
Definition at line 375 of file Tpetra_Experimental_BlockView.hpp.
|
inline |
The block size (number of degrees of freedom per mesh point).
Definition at line 380 of file Tpetra_Experimental_BlockView.hpp.
|
inline |
Stride between consecutive entries.
Definition at line 385 of file Tpetra_Experimental_BlockView.hpp.
|
inline |
Reference to entry (i) of the vector.
impl_scalar_type and not as Scalar, in order to avoid a lot of reinterpret_cast calls in the inner loop of the sparse matrix-vector multiply kernel of Tpetra::Experimental::BlockCrsMatrix. Any pair of types impl_scalar_type, Scalar used here should always be convertible in either direction, so the return type should not pose any issues in practice. Definition at line 399 of file Tpetra_Experimental_BlockView.hpp.
|
inline |
*this := *this + alpha * X.
Definition at line 405 of file Tpetra_Experimental_BlockView.hpp.
|
inline |
*this := X.
Definition at line 414 of file Tpetra_Experimental_BlockView.hpp.
|
inline |
(*this)(i,j) := alpha * (*this)(i,j) for all (i,j).
Definition at line 421 of file Tpetra_Experimental_BlockView.hpp.
|
inline |
(*this)(i,j) := alpha for all (i,j).
Definition at line 429 of file Tpetra_Experimental_BlockView.hpp.
|
inline |
(*this)(i,j) := max(abs((*this)(i,j)), abs(X(i,j))) for all (i,j).
Tpetra uses this operation to implement the ABSMAX CombineMode.
Definition at line 441 of file Tpetra_Experimental_BlockView.hpp.
|
inline |
true if and only if all entries of this equal all entries of X.
Definition at line 450 of file Tpetra_Experimental_BlockView.hpp.
|
inline |
(*this) := (*this) + alpha * A * X (matrix-vector multiply).
Definition at line 465 of file Tpetra_Experimental_BlockView.hpp.
1.8.11