Commit 9e8685af authored by Tomáš Oberhuber's avatar Tomáš Oberhuber
Browse files

Merge branch 'TO/matrices' into 'develop' 

To/matrices

See merge request !80
parents 914aff45 5a1c8790

.gitignore

+3 −0
Original line number Diff line number Diff line
@@ -12,3 +12,6 @@ 
/.settings

/.project

/.pydevproject



# VSCode

/.vscode

.gitlab-ci.yml

+27 −15
Original line number Diff line number Diff line
@@ -76,7 +76,7 @@ stages: 
        - cmake ../..

                -G Ninja

                -DCMAKE_BUILD_TYPE=${BUILD_TYPE}

                -DCMAKE_INSTALL_PREFIX=$(pwd)/${BUILD_TYPE}_install_prefix

                -DCMAKE_INSTALL_PREFIX="$(pwd)/${BUILD_TYPE}_install_prefix"

                -DWITH_OPENMP=${WITH_OPENMP}

                -DWITH_MPI=${WITH_MPI}

                -DWITH_CUDA=${WITH_CUDA}
@@ -102,8 +102,6 @@ stages: 
    only:

        changes:

            - src/**/*.{h,hpp,cpp,cu}

            - Documentation/Examples/**/*.{h,hpp,cpp,cu}

            - Documentation/Tutorials/**/*.{h,hpp,cpp,cu}

            - "**/CMakeLists.txt"

            - .gitlab-ci.yml

    interruptible: true
@@ -117,11 +115,12 @@ dummy build job: 
        - merge_requests

    except:

        changes:

            # .build_template

            - src/**/*.{h,hpp,cpp,cu}

            - Documentation/Examples/**/*.{h,hpp,cpp,cu}

            - Documentation/Tutorials/**/*.{h,hpp,cpp,cu}

            - "**/CMakeLists.txt"

            - .gitlab-ci.yml

            # build documentation

            - Documentation/**/*



# Cuda builds are specified first because they take more time than host-only builds,

# which can be allocated on hosts whitout GPUs.
@@ -189,12 +188,6 @@ cuda_examples_Debug: 
        WITH_CUDA: "yes"

        BUILD_TYPE: Debug

        WITH_EXAMPLES: "yes"

        # build output snippets for documentation

        WITH_DOC: "yes"

    # store output snippets for documentation

    artifacts:

        paths:

            - Documentation/output_snippets/



cuda_examples_Release:

    extends: .build_template
@@ -468,6 +461,28 @@ clang_mpi_benchmarks_tools_python_Release: 






documentation examples:

    extends: .build_template

    stage: build:cuda

    tags:

        - docker

        - nvidia

    variables:

        <<: *default_cmake_flags

        WITH_CUDA: "yes"

        BUILD_TYPE: Debug

        # build output snippets for documentation

        WITH_DOC: "yes"

    only:

        changes:

            - Documentation/**/*

            - src/TNL/**/*.{h,hpp}

            - .gitlab-ci.yml

    # store output snippets for documentation

    artifacts:

        paths:

            - Documentation/output_snippets/



build documentation:

    stage: build:doc

    only:
@@ -477,16 +492,13 @@ build documentation: 
            - .gitlab-ci.yml

    # use "needs" instead of "dependencies" to allow out-of-order start of this job

    needs:

        # the job which builds Documentation/output_snippets/

        - job: cuda_examples_Debug

        - job: documentation examples

          artifacts: true

    script:

        - ./Documentation/build

    artifacts:

        paths:

            - ./Documentation/html/

#    tags:

#        - doxygen



deploy documentation:

    stage: deploy

Documentation/Examples/Matrices/DenseMatrix/DenseMatrixExample_setElement.cpp

+3 −3
Original line number Diff line number Diff line
@@ -16,8 +16,8 @@ void setElements() 
   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 );

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

      matrix->addElement( i, j, 5.0 );

   };



   /***
@@ -26,7 +26,7 @@ void setElements() 
    * DenseMatrixView::getRow example for details.

    */

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

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

   TNL::Algorithms::ParallelFor2D< Device >::exec( 0, 0, 5, 5, f );



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

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

Documentation/Examples/Matrices/DenseMatrix/DenseMatrixViewExample_setElement.cpp

+3 −3
Original line number Diff line number Diff line
@@ -14,10 +14,10 @@ void setElements() 
   std::cout << "Matrix set from the host:" << std::endl;

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



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

      matrixView.setElement( i, i, -i );

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

      matrixView.addElement( i, j, 5.0 );

   };

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

   TNL::Algorithms::ParallelFor2D< Device >::exec( 0, 0, 5, 5, f );



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

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

Documentation/Examples/Matrices/LambdaMatrix/CMakeLists.txt

+24 −2
Original line number Diff line number Diff line
@@ -13,8 +13,18 @@ ADD_CUSTOM_COMMAND( COMMAND LambdaMatrixExample_getNonzeroElementsCount > 
                     ${TNL_DOCUMENTATION_OUTPUT_SNIPPETS_PATH}/LambdaMatrixExample_getNonzeroElementsCount.out

                    OUTPUT LambdaMatrixExample_getNonzeroElementsCount.out )





IF( BUILD_CUDA )

   CUDA_ADD_EXECUTABLE( LambdaMatrixExample_Laplace_cuda LambdaMatrixExample_Laplace.cu )

   ADD_CUSTOM_COMMAND( COMMAND LambdaMatrixExample_Laplace_cuda >

                        ${TNL_DOCUMENTATION_OUTPUT_SNIPPETS_PATH}/LambdaMatrixExample_Laplace.out

                     OUTPUT LambdaMatrixExample_Laplace.out )



   CUDA_ADD_EXECUTABLE( LambdaMatrixExample_Laplace_2_cuda LambdaMatrixExample_Laplace_2.cu )

   ADD_CUSTOM_COMMAND( COMMAND LambdaMatrixExample_Laplace_2_cuda >

                       ${TNL_DOCUMENTATION_OUTPUT_SNIPPETS_PATH}/LambdaMatrixExample_Laplace_2.out

                       OUTPUT LambdaMatrixExample_Laplace_2.out )

                  



                     CUDA_ADD_EXECUTABLE( LambdaMatrixExample_rowsReduction_cuda LambdaMatrixExample_rowsReduction.cu )

   ADD_CUSTOM_COMMAND( COMMAND LambdaMatrixExample_rowsReduction_cuda >

                        ${TNL_DOCUMENTATION_OUTPUT_SNIPPETS_PATH}/LambdaMatrixExample_rowsReduction.out
@@ -36,6 +46,16 @@ IF( BUILD_CUDA ) 
                       OUTPUT LambdaMatrixExample_forAllRows.out )



ELSE()

   ADD_EXECUTABLE( LambdaMatrixExample_Laplace LambdaMatrixExample_Laplace.cpp )

   ADD_CUSTOM_COMMAND( COMMAND LambdaMatrixExample_Laplace >

                       ${TNL_DOCUMENTATION_OUTPUT_SNIPPETS_PATH}/LambdaMatrixExample_Laplace.out

                       OUTPUT LambdaMatrixExample_Laplace.out )



   ADD_EXECUTABLE( LambdaMatrixExample_Laplace_2 LambdaMatrixExample_Laplace_2.cpp )

   ADD_CUSTOM_COMMAND( COMMAND LambdaMatrixExample_Laplace_2 >

                        ${TNL_DOCUMENTATION_OUTPUT_SNIPPETS_PATH}/LambdaMatrixExample_Laplace_2.out

                        OUTPUT LambdaMatrixExample_Laplace_2.out )



   ADD_EXECUTABLE( LambdaMatrixExample_rowsReduction LambdaMatrixExample_rowsReduction.cpp )

   ADD_CUSTOM_COMMAND( COMMAND LambdaMatrixExample_rowsReduction >

                        ${TNL_DOCUMENTATION_OUTPUT_SNIPPETS_PATH}/LambdaMatrixExample_rowsReduction.out
@@ -59,6 +79,8 @@ ENDIF() 


ADD_CUSTOM_TARGET( RunLambdaMatricesExamples ALL DEPENDS

   LambdaMatrixExample_Constructor.out

   LambdaMatrixExample_Laplace.out

   LambdaMatrixExample_Laplace_2.out

   LambdaMatrixExample_getCompressedRowLengths.out

   LambdaMatrixExample_getNonzeroElementsCount.out

   LambdaMatrixExample_rowsReduction.out