Rcpp_matrix_sum

C++ Function Reference

1 Signature

Rcpp::RObject BigDataStatMeth::Rcpp_matrix_sum(T A, T B)

2 Description

Low-level block-based matrix-vector addition implementation.

3 Parameters

  • A (T): Input matrix
  • B (T): Input vector
  • block_size (``): Size of processing blocks
  • bparal (``): Enable/disable parallel processing
  • threads (``): Number of threads for parallel computation

4 Returns

Eigen::MatrixXd containing the result

5 Details

Core implementation that:Processes matrix in cache-friendly blocksSupports parallel execution through OpenMPOptimizes memory access patternsHandles edge cases for non-uniform block sizes

6 Source Code

File: inst/include/memAlgebra/memSum.hpp • Lines 91-115

inline Rcpp::RObject Rcpp_matrix_sum ( T  A, T  B)
    {
        
        // static_assert(std::is_same<T, Eigen::MatrixXd >::value ||
        //               std::is_same<T, Eigen::Map< Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic, Eigen::RowMajor>> >::value ||
        //               std::is_same<T, Eigen::Map< Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic, Eigen::ColMajor>> >::value ||
        //               std::is_same<T, Rcpp::NumericMatrix >::value,
        //               "Error - type not allowed");
        
        Rcpp::NumericMatrix m = Rcpp::as<Rcpp::NumericMatrix>(A);
        Rcpp::NumericMatrix m2 = Rcpp::as<Rcpp::NumericMatrix>(B);
        
        if( m.rows() == m2.rows() && m.cols() == m2.cols()) {
            Rcpp::NumericVector C = m + m2;
            C.attr("dim") = Rcpp::Dimension( m.rows(), m.cols());
            
            return(C);
            
        } else {
            Rcpp::Rcout<<"Error: non-conformable arguments";
        }
        
        return(R_NilValue);
        
    }

7 Usage Example

#include "BigDataStatMeth.hpp"

// Example usage
auto result = Rcpp_matrix_sum(...);