Loading Documentation/Tutorials/Matrices/CMakeLists.txt +3 −3 Original line number Diff line number Diff line Loading @@ -104,9 +104,9 @@ ELSE() #### # THe following examples/benchmarks run for very long time ADD_EXECUTABLE( DenseMatrixSetup_Benchmark DenseMatrixSetup_Benchmark_cuda.cpp ) ADD_EXECUTABLE( SparseMatrixSetup_Benchmark SparseMatrixSetup_Benchmark_cuda.cpp ) ADD_EXECUTABLE( MultidiagonalMatrixSetup_Benchmark MultidiagonalMatrixSetup_Benchmark_cuda.cpp ) ADD_EXECUTABLE( DenseMatrixSetup_Benchmark DenseMatrixSetup_Benchmark.cpp ) ADD_EXECUTABLE( SparseMatrixSetup_Benchmark SparseMatrixSetup_Benchmark.cpp ) ADD_EXECUTABLE( MultidiagonalMatrixSetup_Benchmark MultidiagonalMatrixSetup_Benchmark.cpp ) ENDIF() IF( BUILD_CUDA ) Loading Documentation/Tutorials/Matrices/SparseMatrixSetup_Benchmark.cpp +1 −1 Original line number Diff line number Diff line Loading @@ -69,7 +69,7 @@ template< typename Matrix > void setElement_on_host_and_transfer( const int gridSize, Matrix& matrix ) { using RealType = typename Matrix::RealType; using HostMatrix = typename Matrix::Self< RealType, TNL::Devices::Host >; using HostMatrix = typename Matrix::template Self< RealType, TNL::Devices::Host >; const int matrixSize = gridSize * gridSize; TNL::Containers::Vector< int, typename HostMatrix::DeviceType, int > rowCapacities( matrixSize, 5 ); Loading src/Benchmarks/LinearSolvers/tnl-benchmark-linear-solvers.h +1 −1 Original line number Diff line number Diff line Loading @@ -479,7 +479,7 @@ struct LinearSolversBenchmark DeviceType, IndexType, TNL::Matrices::GeneralMatrix, Algorithms::Segments::CSR Algorithms::Segments::CSRDefault >; SharedPointer< CSR > matrixCopy; Matrices::copySparseMatrix( *matrixCopy, *matrixPointer ); Loading src/Benchmarks/SpMV/spmv-legacy.h +33 −21 Original line number Diff line number Diff line Loading @@ -49,7 +49,16 @@ using SlicedEllpackAlias = Matrices::Legacy::SlicedEllpack< Real, Device, Index // Segments based sparse matrix aliases template< typename Real, typename Device, typename Index > using SparseMatrix_CSR = Matrices::SparseMatrix< Real, Device, Index, Matrices::GeneralMatrix, Algorithms::Segments::CSR >; using SparseMatrix_CSR_Scalar = Matrices::SparseMatrix< Real, Device, Index, Matrices::GeneralMatrix, Algorithms::Segments::CSRScalar >; template< typename Real, typename Device, typename Index > using SparseMatrix_CSR_Vector = Matrices::SparseMatrix< Real, Device, Index, Matrices::GeneralMatrix, Algorithms::Segments::CSRVector >; template< typename Real, typename Device, typename Index > using SparseMatrix_CSR_Hybrid = Matrices::SparseMatrix< Real, Device, Index, Matrices::GeneralMatrix, Algorithms::Segments::CSRHybrid >; template< typename Real, typename Device, typename Index > using SparseMatrix_CSR_Adaptive = Matrices::SparseMatrix< Real, Device, Index, Matrices::GeneralMatrix, Algorithms::Segments::CSRAdaptive >; template< typename Device, typename Index, typename IndexAllocator > using EllpackSegments = Algorithms::Segments::Ellpack< Device, Index, IndexAllocator >; Loading Loading @@ -320,7 +329,10 @@ benchmarkSpmvSynthetic( Benchmark& benchmark, benchmarkSpMV< Real, SparseMatrixLegacy_CSR_Adaptive >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrixLegacy_CSR_MultiVector >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrixLegacy_CSR_LightWithoutAtomic >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrix_CSR >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrix_CSR_Scalar >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrix_CSR_Vector >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrix_CSR_Hybrid >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrix_CSR_Adaptive >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, Matrices::Legacy::Ellpack >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrix_Ellpack >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SlicedEllpackAlias >( benchmark, hostOutVector, inputFileName, verboseMR ); Loading src/Benchmarks/SpMV/tnl-benchmark-spmv.h +1 −3 Original line number Diff line number Diff line Loading @@ -63,7 +63,6 @@ std::string getCurrDateTime() timeinfo = localtime( &rawtime ); strftime( buffer, sizeof( buffer ), "%d-%m-%Y--%H:%M:%S", timeinfo ); std::string curr_date_time( buffer ); return curr_date_time; } Loading Loading @@ -134,7 +133,6 @@ main( int argc, char* argv[] ) // prepare global metadata Benchmark::MetadataMap metadata = getHardwareMetadata(); // Initiate setup of benchmarks if( precision == "all" || precision == "float" ) runSpMVBenchmarks< float >( benchmark, metadata, inputFileName, verboseMR ); Loading Loading
Documentation/Tutorials/Matrices/CMakeLists.txt +3 −3 Original line number Diff line number Diff line Loading @@ -104,9 +104,9 @@ ELSE() #### # THe following examples/benchmarks run for very long time ADD_EXECUTABLE( DenseMatrixSetup_Benchmark DenseMatrixSetup_Benchmark_cuda.cpp ) ADD_EXECUTABLE( SparseMatrixSetup_Benchmark SparseMatrixSetup_Benchmark_cuda.cpp ) ADD_EXECUTABLE( MultidiagonalMatrixSetup_Benchmark MultidiagonalMatrixSetup_Benchmark_cuda.cpp ) ADD_EXECUTABLE( DenseMatrixSetup_Benchmark DenseMatrixSetup_Benchmark.cpp ) ADD_EXECUTABLE( SparseMatrixSetup_Benchmark SparseMatrixSetup_Benchmark.cpp ) ADD_EXECUTABLE( MultidiagonalMatrixSetup_Benchmark MultidiagonalMatrixSetup_Benchmark.cpp ) ENDIF() IF( BUILD_CUDA ) Loading
Documentation/Tutorials/Matrices/SparseMatrixSetup_Benchmark.cpp +1 −1 Original line number Diff line number Diff line Loading @@ -69,7 +69,7 @@ template< typename Matrix > void setElement_on_host_and_transfer( const int gridSize, Matrix& matrix ) { using RealType = typename Matrix::RealType; using HostMatrix = typename Matrix::Self< RealType, TNL::Devices::Host >; using HostMatrix = typename Matrix::template Self< RealType, TNL::Devices::Host >; const int matrixSize = gridSize * gridSize; TNL::Containers::Vector< int, typename HostMatrix::DeviceType, int > rowCapacities( matrixSize, 5 ); Loading
src/Benchmarks/LinearSolvers/tnl-benchmark-linear-solvers.h +1 −1 Original line number Diff line number Diff line Loading @@ -479,7 +479,7 @@ struct LinearSolversBenchmark DeviceType, IndexType, TNL::Matrices::GeneralMatrix, Algorithms::Segments::CSR Algorithms::Segments::CSRDefault >; SharedPointer< CSR > matrixCopy; Matrices::copySparseMatrix( *matrixCopy, *matrixPointer ); Loading
src/Benchmarks/SpMV/spmv-legacy.h +33 −21 Original line number Diff line number Diff line Loading @@ -49,7 +49,16 @@ using SlicedEllpackAlias = Matrices::Legacy::SlicedEllpack< Real, Device, Index // Segments based sparse matrix aliases template< typename Real, typename Device, typename Index > using SparseMatrix_CSR = Matrices::SparseMatrix< Real, Device, Index, Matrices::GeneralMatrix, Algorithms::Segments::CSR >; using SparseMatrix_CSR_Scalar = Matrices::SparseMatrix< Real, Device, Index, Matrices::GeneralMatrix, Algorithms::Segments::CSRScalar >; template< typename Real, typename Device, typename Index > using SparseMatrix_CSR_Vector = Matrices::SparseMatrix< Real, Device, Index, Matrices::GeneralMatrix, Algorithms::Segments::CSRVector >; template< typename Real, typename Device, typename Index > using SparseMatrix_CSR_Hybrid = Matrices::SparseMatrix< Real, Device, Index, Matrices::GeneralMatrix, Algorithms::Segments::CSRHybrid >; template< typename Real, typename Device, typename Index > using SparseMatrix_CSR_Adaptive = Matrices::SparseMatrix< Real, Device, Index, Matrices::GeneralMatrix, Algorithms::Segments::CSRAdaptive >; template< typename Device, typename Index, typename IndexAllocator > using EllpackSegments = Algorithms::Segments::Ellpack< Device, Index, IndexAllocator >; Loading Loading @@ -320,7 +329,10 @@ benchmarkSpmvSynthetic( Benchmark& benchmark, benchmarkSpMV< Real, SparseMatrixLegacy_CSR_Adaptive >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrixLegacy_CSR_MultiVector >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrixLegacy_CSR_LightWithoutAtomic >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrix_CSR >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrix_CSR_Scalar >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrix_CSR_Vector >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrix_CSR_Hybrid >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrix_CSR_Adaptive >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, Matrices::Legacy::Ellpack >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SparseMatrix_Ellpack >( benchmark, hostOutVector, inputFileName, verboseMR ); benchmarkSpMV< Real, SlicedEllpackAlias >( benchmark, hostOutVector, inputFileName, verboseMR ); Loading
src/Benchmarks/SpMV/tnl-benchmark-spmv.h +1 −3 Original line number Diff line number Diff line Loading @@ -63,7 +63,6 @@ std::string getCurrDateTime() timeinfo = localtime( &rawtime ); strftime( buffer, sizeof( buffer ), "%d-%m-%Y--%H:%M:%S", timeinfo ); std::string curr_date_time( buffer ); return curr_date_time; } Loading Loading @@ -134,7 +133,6 @@ main( int argc, char* argv[] ) // prepare global metadata Benchmark::MetadataMap metadata = getHardwareMetadata(); // Initiate setup of benchmarks if( precision == "all" || precision == "float" ) runSpMVBenchmarks< float >( benchmark, metadata, inputFileName, verboseMR ); Loading