Update of DenseMatrix::setElement example.

#include <iostream>

#include <TNL/Algorithms/ParallelFor.h>

#include <TNL/Matrices/DenseMatrix.h>

#include <TNL/Devices/Host.h>

#include <TNL/Pointers/SharedPointer.h>

#include <TNL/Pointers/SmartPointersRegister.h>

template< typename Device >

void setElements()

{

   TNL::Matrices::DenseMatrix< double, Device > matrix( 5, 5 );

   TNL::Pointers::SharedPointer< TNL::Matrices::DenseMatrix< double, Device > > matrix( 5, 5 );

   for( int i = 0; i < 5; i++ )

      matrix.setElement( i, i, i );

      matrix->setElement( i, i, i );

   std::cout << "Matrix set from the host:" << std::endl;

   std::cout << *matrix << std::endl;

   auto f = [=] __cuda_callable__ ( int i ) mutable {

      matrix->setElement( i, i, -i );

   };

   TNL::Pointers::synchronizeSmartPointersOnDevice< Device >();

   TNL::Algorithms::ParallelFor< Device >::exec( 0, 5, f );

   std::cout << "Matrix set from its native device:" << std::endl;

   std::cout << *matrix << std::endl;

   std::cout << matrix << std::endl;

}

int main( int argc, char* argv[] )

       * \param column is columns index of the element.

       * \param value is the value the element will be set to.

       * \param thisElementMultiplicator is multiplicator the original matrix element

       *   value is multiplied by before addition of given e value.

       *   value is multiplied by before addition of given \e value.

       */

      __cuda_callable__

      void addElement( const IndexType row,