Optimizing dense-matrix vector multiplication inspired by sparse matrix implementation. (d5984f14) · Commits · TNL / tnl-dev

src/TNL/Matrices/DenseMatrixView.hpp

+25 −2

Original line number	Diff line number	Diff line
		@@ -538,10 +538,33 @@ vectorProduct( const InVector& inVector,
		auto fetch = [=] __cuda_callable__ ( IndexType row, IndexType column, IndexType offset, bool& compute ) -> RealType {
		return valuesView[ offset ] * inVectorView[ column ];
		};
		auto keeper = [=] __cuda_callable__ ( IndexType row, const RealType& value ) mutable {
		auto keeperGeneral = [=] __cuda_callable__ ( IndexType row, const RealType& value ) mutable {
		outVectorView[ row ] = matrixMultiplicator * value + outVectorMultiplicator * outVectorView[ row ];
		};
		this->segments.reduceSegments( begin, end, fetch, std::plus<>{}, keeper, ( RealType ) 0.0 );
		auto keeperDirect = [=] __cuda_callable__ ( IndexType row, const RealType& value ) mutable {
		outVectorView[ row ] = value;
		};
		auto keeperMatrixMult = [=] __cuda_callable__ ( IndexType row, const RealType& value ) mutable {
		outVectorView[ row ] = matrixMultiplicator * value;
		};
		auto keeperVectorMult = [=] __cuda_callable__ ( IndexType row, const RealType& value ) mutable {
		outVectorView[ row ] = outVectorMultiplicator * outVectorView[ row ] + value;
		};

		if( outVectorMultiplicator == 0.0 )
		{
		if( matrixMultiplicator == 1.0 )
		this->segments.reduceSegments( begin, end, fetch, std::plus<>{}, keeperDirect, ( RealType ) 0.0 );
		else
		this->segments.reduceSegments( begin, end, fetch, std::plus<>{}, keeperMatrixMult, ( RealType ) 0.0 );
		}
		else
		{
		if( matrixMultiplicator == 1.0 )
		this->segments.reduceSegments( begin, end, fetch, std::plus<>{}, keeperVectorMult, ( RealType ) 0.0 );
		else
		this->segments.reduceSegments( begin, end, fetch, std::plus<>{}, keeperGeneral, ( RealType ) 0.0 );
		}
		}

		template< typename Real,