sgpp::datadriven::AlgorithmMultipleEvaluationDistributed< BASIS > Class Template Reference

Abstract implementation for distributed multiple function evaluations. More...

#include <AlgorithmMultipleEvaluationDistributed.hpp>

Public Member Functions
void	mult (sgpp::base::GridStorage &storage, BASIS &basis, sgpp::base::DataVector &source, DataMatrix &x, DataVectorDistributed &result)
	Performs a distributed mass evaluation.
void	mult_transpose (sgpp::base::GridStorage &storage, BASIS &basis, sgpp::base::DataVector &source, DataMatrix &x, DataVectorDistributed &result)
	Performs a distributed transposed mass evaluation.

Detailed Description

template<class BASIS>
class sgpp::datadriven::AlgorithmMultipleEvaluationDistributed< BASIS >

Abstract implementation for distributed multiple function evaluations.

Based on AlgorithmMultipleEvaluation.hpp

In Data Mining to operators are needed: mass evaluation and transposed evaluation, referenced in literature as matrix vector products with matrices B^T (mass evaluation) and B (transposed evaluation).

If there are \(N\) basis functions \(\varphi(\vec{x})\) and \(m\) data points, then B is a (mxN) matrix, with

\[ (B)_{j,i} = \varphi_i(x_j). \]

Member Function Documentation

mult()

template<class BASIS >

void sgpp::datadriven::AlgorithmMultipleEvaluationDistributed< BASIS >::mult ( sgpp::base::GridStorage &  storage, BASIS &  basis, sgpp::base::DataVector &  source, DataMatrix &  x, DataVectorDistributed &  result  )

inline

Performs a distributed mass evaluation.

Parameters

storage	GridStorage object that contains the grid's points information
basis	a reference to a class that implements a specific basis
source	the coefficients of the grid points
x	the d-dimensional vector with data points (row-wise)
result	the result vector of the matrix vector multiplication

References sgpp::datadriven::DataVectorDistributed::getLocalPointer(), sgpp::datadriven::DataVectorDistributed::getLocalRows(), sgpp::base::DataMatrix::getNcols(), sgpp::base::DataMatrix::getRow(), sgpp::datadriven::DataVectorDistributed::isProcessMapped(), and sgpp::datadriven::DataVectorDistributed::localToGlobalRowIndex().

Referenced by sgpp::datadriven::OperationMultipleEvalLinearDistributed::multDistributed(), and sgpp::datadriven::OperationMultipleEvalModLinearDistributed::multDistributed().

mult_transpose()

template<class BASIS >

void sgpp::datadriven::AlgorithmMultipleEvaluationDistributed< BASIS >::mult_transpose ( sgpp::base::GridStorage &  storage, BASIS &  basis, sgpp::base::DataVector &  source, DataMatrix &  x, DataVectorDistributed &  result  )

inline

Performs a distributed transposed mass evaluation.

Parameters

storage	GridStorage object that contains the grid's points information
basis	a reference to a class that implements a specific basis
source	the coefficients of the grid points
x	the d-dimensional vector with data points (row-wise)
result	the result vector of the matrix vector multiplication

References sgpp::datadriven::DataVectorDistributed::add(), sgpp::datadriven::DataVectorDistributed::distribute(), sgpp::datadriven::DataVectorDistributed::getBlockSize(), sgpp::datadriven::DataVectorDistributed::getGlobalRows(), sgpp::base::DataMatrix::getNcols(), sgpp::datadriven::DataVectorDistributed::getProcessGrid(), sgpp::base::DataMatrix::getRow(), sgpp::base::DataVector::getSize(), sgpp::datadriven::DataVectorDistributed::isProcessMapped(), and sgpp::datadriven::DataVectorDistributed::setAll().

Referenced by sgpp::datadriven::OperationMultipleEvalLinearDistributed::multTransposeDistributed(), and sgpp::datadriven::OperationMultipleEvalModLinearDistributed::multTransposeDistributed().

---

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

• datadriven/src/sgpp/datadriven/algorithm/AlgorithmMultipleEvaluationDistributed.hpp