Loading CMakeLists.txt +1 −0 Original line number Diff line number Diff line Loading @@ -310,6 +310,7 @@ if (DOXYGEN_FOUND) set (DOXYGEN_PROJECT_NAME "Graph Framework") set (DOXYGEN_EXCLUDE ${CMAKE_CURRENT_SOURCE_DIR}/LLVM ${CMAKE_CURRENT_SOURCE_DIR}/build) set (DOXYGEN_GENERATE_TREEVIEW YES) set (DOXYGEN_USE_MATHJAX YES) doxygen_add_docs (doc) endif () Loading graph_c_binding/graph_c_binding.cpp +1 −1 Original line number Diff line number Diff line Loading @@ -1127,7 +1127,7 @@ extern "C" { /// /// @param[in] c The graph C context. /// @param[in] arg The left opperand. /// @returns sin(arg) /// @returns cos(arg) //------------------------------------------------------------------------------ graph_node graph_cos(STRUCT_TAG graph_c_context *c, graph_node arg) { Loading graph_c_binding/graph_c_binding.h +14 −0 Original line number Diff line number Diff line Loading @@ -10,6 +10,16 @@ #include <stdint.h> #include <stdbool.h> //------------------------------------------------------------------------------ /// @def START_GPU /// Starts a Cocoa auto release pool when using the metal backend. No opt /// otherwise. //------------------------------------------------------------------------------ //------------------------------------------------------------------------------ /// @def END_GPU /// Ends a Cocoa auto release pool when using the metal backend. No opt /// otherwise. //------------------------------------------------------------------------------ #ifdef USE_METAL #define START_GPU @autoreleasepool { #define END_GPU } Loading @@ -18,6 +28,10 @@ #define END_GPU #endif //------------------------------------------------------------------------------ /// @def STRUCT_TAG /// C++ mode needs to tag a graph_c_context as a struct. //------------------------------------------------------------------------------ #ifdef __cplusplus extern "C" { #define STRUCT_TAG Loading graph_docs/compiling.dox 0 → 100644 +213 −0 Original line number Diff line number Diff line /*! * @page build_system Build System * @brief Overview of the Cmake based build system for the graph_framework. * @tableofcontents * * @section build_system_introduction Introduction * This page details the <a href="http://www.cmake.org">cmake</a> based build * system. * * Documentation if divided into two parts. * * User documention. * * Developer documentation. * * @section build_system_user User Guide * The following section is for users of framework. * * @subsection build_system_user_dependencies Dependencies * The graph_framwork requires three requires external dependencies and one * optional dependency. <a href="https://llvm.org">LLVM</a> is another * dependency that is used for generating CPU code. However this is * automatically obtained via the build system. The graph_frame is written using * the C++20 standard. The C interface using C17 and the fortran interface using * Fortran 2008. * * @subsubsection build_system_user_dependencies_required Required * * <a href="http://www.cmake.org">cmake</a> version greater than 3.21. * * <a href="https://www.unidata.ucar.edu">NetCDF-C</a> library. * * <a href="https://www.python.org">Python</a> 3 environment. * * @subsubsection build_system_user_dependencies_optional Optional * * <a href="https://www.doxygen.nl/index.html">Doxygen</a> for generating this documentation. * * @subsection build_system_clone Obtaining the code * The framework code itself be obtained from the * <a href="https://github.com/ORNL-Fusion/graph_framework">graph_framework</a> * Github repository. * @code git clone https://github.com/ORNL-Fusion/graph_framework.git @endcode * Source will be downloaded into a <tt>graph_framework</tt> directory * unless a different directory is explicitly used. * * @subsection build_system_gen Generating the build system * After the repository is cloned, create a <tt>build</tt> directory in the * top level source directory and change into that directory. * @code mkdir build cd build @endcode * There are two ways to run cmake. From the command line the build system can * generated by using the <tt>cmake</tt> command with options set using the * the <tt>-D</tt> option. As an exampole. * @code cmake -DOPTION_NAME=OPTION_VALUE ../ @endcode * Where <tt>../</tt> points to the source directory containing the top * level <tt>CMakeLists.txt</tt> file. * * The recommended method is to use the interatice <tt>ccmake</tt> command * instead. * @code ccmake ../ @endcode * Note options can still be set from the command line using the * <tt>-D</tt> option. * * @subsubsection build_system_user_options Build system Options * Initally, there will be no options. Along the botton, there are several * command. Use the 'c' command to start the configuation process. Once * configured several options will apear. During this process cmake is cloning * the LLVM repository. So this step may take some time initally. Mode of the * are various options for configuing LLVM and can be ignored. The important * options are listed below. * * * <tt>CMAKE_BUILD_TYPE</tt> Switch between <tt>Release</tt>, <tt>Debug</tt>, <tt>MinSizeRel</tt>, <tt>RelWithDebInfo</tt> * * <tt>USE_VERBOSE</tt> Show verbose information about compute kernels. * * <tt>BUILD_C_BINDING</tt> Generate the C langauge interface. * * <tt>BUILD_Fortran_BINDING</tt> Generate the Fortran language interface. * * <tt>USE_METAL</tt> Enable the <a href="https://developer.apple.com/metal/">Metal</a> backend (macOS only). * * <tt>USE_CUDA</tt> Enable the <a href="https://developer.nvidia.com/cuda-zone">Cuda</a> backend (Linux only). * * <tt>USE_HIP</tt> Enable the <a href="https://www.amd.com/en/products/software/rocm.html">Hip</a> backend (Linux only, Hip branch). * * <tt>USE_SSH</tt> Use ssh for git instead of html. * * <b>NOTE:</b>macOS uses will need to change the default option for * <tt>CMAKE_CXX_COMPILER</tt> to <tt>clang++</tt>. This is due to the way the * build systems determines default include directories for system libraries. * This can be accomplished using the advacned options using the <tt>t</tt> * command or setting this via the command line. * @code cmake -DCMAKE_CXX_COMPILER=clang++ ../ @endcode * * Every time an option is changed, or a new option is available, you need to * use the configure <tt>c</tt> command for changes to take affect. Once all * options are set, the a generate <tt>g</tt> options will appear. Using this * option will build a make file. * * @subsubsection build_system_trouble_shooting Trouble Shooting. * Some times, cmake will fail to locate the NetCDF library if it is not * installed in a standard path. In these cases you can use the * <tt>CMAKE_PREFIX_PATH</tt> to define the install location of the NetCDF * library. For instance if the netcdf library is installed in * <tt>/foo/bar/lib</tt> the prefix path should be set to <tt>/foo/bar</tt>. * * @subsection build_system_build Building the code. * Once the build system is successfully generated a Makefile will appear. The * code can be build using the * @code make @endcode * command. Note that due build system first starts by pulling the latest * of LLVM. The build system then has to build LLVM first which can take a * while. It is recommended to use a limited parallel build. * @code make -j10 @endcode * The <tt>-j<i>num_processes</i></tt> option determines number of parallel * instances to run. The build products will be found in assocated build * directories in the <tt>build</tt> directory. * * A list of individual components which can be build can be identified using * @code make -h @endcode * * @subsection build_system_test Running unit tests. * Units tests can be run using the command. * @code make test ARGS=-j10 @endcode * Like the parallel build the <tt>-j<i>num_processes</i></tt> option determines * number of parallel instances to run. * * @section build_system_dev Developer Guide * This section covers information for developers of the framework itself. * * @subsection build_system_macros Macro Definitions * The build system defines some macros for defining targets, configuring debug * options, and configuing external dependences. * * @subsubsection build_system_targets Tool targets. * * <hr> * <tt>add_tool_target(target lang)</tt>\n\n * Define a tool target.\n\n * <b>Parameters</b>\n * <tt>[in] <b>target</b></tt> The name of the target.\n * <tt>[in] <b>lang</b></tt> File extention for the target (c, cpp, f90).\n\n * Target assumes there is a source file defined as <tt>target.lang</tt>. For * instance a C++ source file named <tt>foo.cpp</tt> is configures as * @code add_tool_target(foo cpp) @endcode * This will generate a build target called <tt>xfoo</tt>. * * <hr> * <tt>add_test_target(target lang)</tt>\n\n * Define a test target.\n\n * <b>Parameters</b>\n * <tt>[in] <b>target</b></tt> The name of the target.\n * <tt>[in] <b>lang</b></tt> File extention for the target (c, cpp, f90).\n\n * The aguments are the same as <tt>add_tool_target</tt>. This also adds the * target as a unit test. * <hr> * * @subsubsection build_system_sanitizer Sanatizer flags * * <hr> * <tt>register_sanitizer_option(name)</tt>\n\n * Register a sanitizer option.\n\n * <b>Parameters</b>\n * <tt>[in] <b>name</b></tt> The name of the sanitizer flags.\n\n * This add new for using the <tt>SANITIZE_<i>NAME</i></tt> cmake option and * add <tt>-fsanitize=<i>name</i></tt> to the command line arguments. * <hr> * * @subsubsection build_system_project Register an external project * * <hr> * <tt>register_project(reg_name dir url default_tag sub_dir)</tt>\n\n * Register an external project.\n\n * <b>Parameters</b>\n * <tt>[in] <b>reg_name</b></tt> Name for the registered project.\n * <tt>[in] <b>dir</b></tt> Name directory to clone the project to.\n * <tt>[in] <b>url</b></tt> URL for the repository.\n * <tt>[in] <b>default_tag</b></tt> Default tag for to pull from.\n * <tt>[in] <b>sub_dir</b></tt> Subdirectory to locate the project source code.\n\n * This function clones a external project into the directory defined by * <tt>dir</tt>. This also adds a new build option for * <tt>BUILD_TAG_<i>DIR</i></tt>. The URL must have the format of * @code ${URL_PROTO}domain.com${URL_SEP}remining/url @endcode * <hr> * * @subsection build_system_debug Debugging * In addition to the standard build options there are several debugging options * that can be enabled. * * @subsubsection build_system_dev_options Build system Options * * <tt>USE_PCH</tt> Use precomiled headers during computation. Most users should keep this on. * * <tt>SAVE_KERNEL_SOURCE</tt> Option to dump the generated compute kernel source code to disk. * * <tt>USE_INPUT_CACHE</tt> Option to cache registers for the kernel arguments. * * <tt>USE_CONSTANT_CACHE</tt> Option to use registers to cache constant values otherwise constanst are inlined. * * <tt>SHOW_USE_COUNT</tt> Generates information on the number of times a register is used. * * <tt>USE_INDEX_CACHE</tt> Option to use registers to cache array indicies. * * <tt>SANITIZE_ADDRESS</tt> Use address sanitizer debugging option. * * <tt>SANITIZE_LEAK</tt> Use leak sanitizer debugging option. * * <tt>SANITIZE_MEMORY</tt> Use memory sanitizer debugging option. * * <tt>SANITIZE_THREAD</tt> Use thread sanitizer debugging option. * * <tt>SANITIZE_UNDEFINED</tt> Use undefined sanitizer debugging option. * * <tt>SANITIZE_FLOAT-DIVIDE-BY-ZERO</tt> Use float-divide-by-zero sanitizer debugging option. */ graph_docs/general.dox 0 → 100644 +30 −0 Original line number Diff line number Diff line /*! * @page general_concepts General Concepts * @tableofcontents * @section introduction Introduction * This page documents general concepts of the graph_framework. * * @section general_concepts_definitions Definitions. * * <b>leaf_node</b> A leaf on the graph. * * <b>graph</b> A data stucture connecting leaf nodes. * * <b>reduce</b> A tranformation of the graph to remove leaf_nodes. * * <b>auto differentiation</b> A tranformation of the graph build derivatives. * * <b>compiler</b> A tool for translating from one language to another. * * <b>JIT</b> Just-in-time compile. * * <b>kernel</b> A code function that runs on a batch of data. * * <b>pre_item</b> A kernel to run before running the main kernels. * * <b>work_iten</b> A instance of kernel. * * <b>converge_item</b> A kernel that is run until a convergence test is met. * * <b>workflow</b> A series of work items. * * <b>backend</b> The device the kernel is run on. * * @section general_concepts_graph Graph * The graph_framework operates by building tree structure of math operations. * In tree form it is easy to traverse nodes in the graph. Take the example of * equation of line. * @f{equation}{y=mx + b@f} * This equation consists of five leaf nodes. The ends of the tree are clasified * as either variables @f$x@f$ or constants @f$m,b@f$. These leaf_nodes are * connected by leaf nodes for multiply and additon operations. The ouptut * @f$ y@f$ represents the entire graph of operations. */ Loading
CMakeLists.txt +1 −0 Original line number Diff line number Diff line Loading @@ -310,6 +310,7 @@ if (DOXYGEN_FOUND) set (DOXYGEN_PROJECT_NAME "Graph Framework") set (DOXYGEN_EXCLUDE ${CMAKE_CURRENT_SOURCE_DIR}/LLVM ${CMAKE_CURRENT_SOURCE_DIR}/build) set (DOXYGEN_GENERATE_TREEVIEW YES) set (DOXYGEN_USE_MATHJAX YES) doxygen_add_docs (doc) endif () Loading
graph_c_binding/graph_c_binding.cpp +1 −1 Original line number Diff line number Diff line Loading @@ -1127,7 +1127,7 @@ extern "C" { /// /// @param[in] c The graph C context. /// @param[in] arg The left opperand. /// @returns sin(arg) /// @returns cos(arg) //------------------------------------------------------------------------------ graph_node graph_cos(STRUCT_TAG graph_c_context *c, graph_node arg) { Loading
graph_c_binding/graph_c_binding.h +14 −0 Original line number Diff line number Diff line Loading @@ -10,6 +10,16 @@ #include <stdint.h> #include <stdbool.h> //------------------------------------------------------------------------------ /// @def START_GPU /// Starts a Cocoa auto release pool when using the metal backend. No opt /// otherwise. //------------------------------------------------------------------------------ //------------------------------------------------------------------------------ /// @def END_GPU /// Ends a Cocoa auto release pool when using the metal backend. No opt /// otherwise. //------------------------------------------------------------------------------ #ifdef USE_METAL #define START_GPU @autoreleasepool { #define END_GPU } Loading @@ -18,6 +28,10 @@ #define END_GPU #endif //------------------------------------------------------------------------------ /// @def STRUCT_TAG /// C++ mode needs to tag a graph_c_context as a struct. //------------------------------------------------------------------------------ #ifdef __cplusplus extern "C" { #define STRUCT_TAG Loading
graph_docs/compiling.dox 0 → 100644 +213 −0 Original line number Diff line number Diff line /*! * @page build_system Build System * @brief Overview of the Cmake based build system for the graph_framework. * @tableofcontents * * @section build_system_introduction Introduction * This page details the <a href="http://www.cmake.org">cmake</a> based build * system. * * Documentation if divided into two parts. * * User documention. * * Developer documentation. * * @section build_system_user User Guide * The following section is for users of framework. * * @subsection build_system_user_dependencies Dependencies * The graph_framwork requires three requires external dependencies and one * optional dependency. <a href="https://llvm.org">LLVM</a> is another * dependency that is used for generating CPU code. However this is * automatically obtained via the build system. The graph_frame is written using * the C++20 standard. The C interface using C17 and the fortran interface using * Fortran 2008. * * @subsubsection build_system_user_dependencies_required Required * * <a href="http://www.cmake.org">cmake</a> version greater than 3.21. * * <a href="https://www.unidata.ucar.edu">NetCDF-C</a> library. * * <a href="https://www.python.org">Python</a> 3 environment. * * @subsubsection build_system_user_dependencies_optional Optional * * <a href="https://www.doxygen.nl/index.html">Doxygen</a> for generating this documentation. * * @subsection build_system_clone Obtaining the code * The framework code itself be obtained from the * <a href="https://github.com/ORNL-Fusion/graph_framework">graph_framework</a> * Github repository. * @code git clone https://github.com/ORNL-Fusion/graph_framework.git @endcode * Source will be downloaded into a <tt>graph_framework</tt> directory * unless a different directory is explicitly used. * * @subsection build_system_gen Generating the build system * After the repository is cloned, create a <tt>build</tt> directory in the * top level source directory and change into that directory. * @code mkdir build cd build @endcode * There are two ways to run cmake. From the command line the build system can * generated by using the <tt>cmake</tt> command with options set using the * the <tt>-D</tt> option. As an exampole. * @code cmake -DOPTION_NAME=OPTION_VALUE ../ @endcode * Where <tt>../</tt> points to the source directory containing the top * level <tt>CMakeLists.txt</tt> file. * * The recommended method is to use the interatice <tt>ccmake</tt> command * instead. * @code ccmake ../ @endcode * Note options can still be set from the command line using the * <tt>-D</tt> option. * * @subsubsection build_system_user_options Build system Options * Initally, there will be no options. Along the botton, there are several * command. Use the 'c' command to start the configuation process. Once * configured several options will apear. During this process cmake is cloning * the LLVM repository. So this step may take some time initally. Mode of the * are various options for configuing LLVM and can be ignored. The important * options are listed below. * * * <tt>CMAKE_BUILD_TYPE</tt> Switch between <tt>Release</tt>, <tt>Debug</tt>, <tt>MinSizeRel</tt>, <tt>RelWithDebInfo</tt> * * <tt>USE_VERBOSE</tt> Show verbose information about compute kernels. * * <tt>BUILD_C_BINDING</tt> Generate the C langauge interface. * * <tt>BUILD_Fortran_BINDING</tt> Generate the Fortran language interface. * * <tt>USE_METAL</tt> Enable the <a href="https://developer.apple.com/metal/">Metal</a> backend (macOS only). * * <tt>USE_CUDA</tt> Enable the <a href="https://developer.nvidia.com/cuda-zone">Cuda</a> backend (Linux only). * * <tt>USE_HIP</tt> Enable the <a href="https://www.amd.com/en/products/software/rocm.html">Hip</a> backend (Linux only, Hip branch). * * <tt>USE_SSH</tt> Use ssh for git instead of html. * * <b>NOTE:</b>macOS uses will need to change the default option for * <tt>CMAKE_CXX_COMPILER</tt> to <tt>clang++</tt>. This is due to the way the * build systems determines default include directories for system libraries. * This can be accomplished using the advacned options using the <tt>t</tt> * command or setting this via the command line. * @code cmake -DCMAKE_CXX_COMPILER=clang++ ../ @endcode * * Every time an option is changed, or a new option is available, you need to * use the configure <tt>c</tt> command for changes to take affect. Once all * options are set, the a generate <tt>g</tt> options will appear. Using this * option will build a make file. * * @subsubsection build_system_trouble_shooting Trouble Shooting. * Some times, cmake will fail to locate the NetCDF library if it is not * installed in a standard path. In these cases you can use the * <tt>CMAKE_PREFIX_PATH</tt> to define the install location of the NetCDF * library. For instance if the netcdf library is installed in * <tt>/foo/bar/lib</tt> the prefix path should be set to <tt>/foo/bar</tt>. * * @subsection build_system_build Building the code. * Once the build system is successfully generated a Makefile will appear. The * code can be build using the * @code make @endcode * command. Note that due build system first starts by pulling the latest * of LLVM. The build system then has to build LLVM first which can take a * while. It is recommended to use a limited parallel build. * @code make -j10 @endcode * The <tt>-j<i>num_processes</i></tt> option determines number of parallel * instances to run. The build products will be found in assocated build * directories in the <tt>build</tt> directory. * * A list of individual components which can be build can be identified using * @code make -h @endcode * * @subsection build_system_test Running unit tests. * Units tests can be run using the command. * @code make test ARGS=-j10 @endcode * Like the parallel build the <tt>-j<i>num_processes</i></tt> option determines * number of parallel instances to run. * * @section build_system_dev Developer Guide * This section covers information for developers of the framework itself. * * @subsection build_system_macros Macro Definitions * The build system defines some macros for defining targets, configuring debug * options, and configuing external dependences. * * @subsubsection build_system_targets Tool targets. * * <hr> * <tt>add_tool_target(target lang)</tt>\n\n * Define a tool target.\n\n * <b>Parameters</b>\n * <tt>[in] <b>target</b></tt> The name of the target.\n * <tt>[in] <b>lang</b></tt> File extention for the target (c, cpp, f90).\n\n * Target assumes there is a source file defined as <tt>target.lang</tt>. For * instance a C++ source file named <tt>foo.cpp</tt> is configures as * @code add_tool_target(foo cpp) @endcode * This will generate a build target called <tt>xfoo</tt>. * * <hr> * <tt>add_test_target(target lang)</tt>\n\n * Define a test target.\n\n * <b>Parameters</b>\n * <tt>[in] <b>target</b></tt> The name of the target.\n * <tt>[in] <b>lang</b></tt> File extention for the target (c, cpp, f90).\n\n * The aguments are the same as <tt>add_tool_target</tt>. This also adds the * target as a unit test. * <hr> * * @subsubsection build_system_sanitizer Sanatizer flags * * <hr> * <tt>register_sanitizer_option(name)</tt>\n\n * Register a sanitizer option.\n\n * <b>Parameters</b>\n * <tt>[in] <b>name</b></tt> The name of the sanitizer flags.\n\n * This add new for using the <tt>SANITIZE_<i>NAME</i></tt> cmake option and * add <tt>-fsanitize=<i>name</i></tt> to the command line arguments. * <hr> * * @subsubsection build_system_project Register an external project * * <hr> * <tt>register_project(reg_name dir url default_tag sub_dir)</tt>\n\n * Register an external project.\n\n * <b>Parameters</b>\n * <tt>[in] <b>reg_name</b></tt> Name for the registered project.\n * <tt>[in] <b>dir</b></tt> Name directory to clone the project to.\n * <tt>[in] <b>url</b></tt> URL for the repository.\n * <tt>[in] <b>default_tag</b></tt> Default tag for to pull from.\n * <tt>[in] <b>sub_dir</b></tt> Subdirectory to locate the project source code.\n\n * This function clones a external project into the directory defined by * <tt>dir</tt>. This also adds a new build option for * <tt>BUILD_TAG_<i>DIR</i></tt>. The URL must have the format of * @code ${URL_PROTO}domain.com${URL_SEP}remining/url @endcode * <hr> * * @subsection build_system_debug Debugging * In addition to the standard build options there are several debugging options * that can be enabled. * * @subsubsection build_system_dev_options Build system Options * * <tt>USE_PCH</tt> Use precomiled headers during computation. Most users should keep this on. * * <tt>SAVE_KERNEL_SOURCE</tt> Option to dump the generated compute kernel source code to disk. * * <tt>USE_INPUT_CACHE</tt> Option to cache registers for the kernel arguments. * * <tt>USE_CONSTANT_CACHE</tt> Option to use registers to cache constant values otherwise constanst are inlined. * * <tt>SHOW_USE_COUNT</tt> Generates information on the number of times a register is used. * * <tt>USE_INDEX_CACHE</tt> Option to use registers to cache array indicies. * * <tt>SANITIZE_ADDRESS</tt> Use address sanitizer debugging option. * * <tt>SANITIZE_LEAK</tt> Use leak sanitizer debugging option. * * <tt>SANITIZE_MEMORY</tt> Use memory sanitizer debugging option. * * <tt>SANITIZE_THREAD</tt> Use thread sanitizer debugging option. * * <tt>SANITIZE_UNDEFINED</tt> Use undefined sanitizer debugging option. * * <tt>SANITIZE_FLOAT-DIVIDE-BY-ZERO</tt> Use float-divide-by-zero sanitizer debugging option. */
graph_docs/general.dox 0 → 100644 +30 −0 Original line number Diff line number Diff line /*! * @page general_concepts General Concepts * @tableofcontents * @section introduction Introduction * This page documents general concepts of the graph_framework. * * @section general_concepts_definitions Definitions. * * <b>leaf_node</b> A leaf on the graph. * * <b>graph</b> A data stucture connecting leaf nodes. * * <b>reduce</b> A tranformation of the graph to remove leaf_nodes. * * <b>auto differentiation</b> A tranformation of the graph build derivatives. * * <b>compiler</b> A tool for translating from one language to another. * * <b>JIT</b> Just-in-time compile. * * <b>kernel</b> A code function that runs on a batch of data. * * <b>pre_item</b> A kernel to run before running the main kernels. * * <b>work_iten</b> A instance of kernel. * * <b>converge_item</b> A kernel that is run until a convergence test is met. * * <b>workflow</b> A series of work items. * * <b>backend</b> The device the kernel is run on. * * @section general_concepts_graph Graph * The graph_framework operates by building tree structure of math operations. * In tree form it is easy to traverse nodes in the graph. Take the example of * equation of line. * @f{equation}{y=mx + b@f} * This equation consists of five leaf nodes. The ends of the tree are clasified * as either variables @f$x@f$ or constants @f$m,b@f$. These leaf_nodes are * connected by leaf nodes for multiply and additon operations. The ouptut * @f$ y@f$ represents the entire graph of operations. */
