Merge branch 'develop' of https://github.com/ROCmSoftwarePlatform/AMDMIGraphX into develop

CMakeLists.txt

@@ -51,6 +51,7 @@ project(migraphx LANGUAGES C CXX)
 include(CTest)
 
 find_package(ROCM REQUIRED)
+find_package(Threads REQUIRED)
 
 find_path(HALF_INCLUDE_DIR half.hpp PATH_SUFFIXES half)
 if (NOT HALF_INCLUDE_DIR)
@@ -227,14 +228,12 @@ rocm_enable_cppcheck(
         shadowVar
         shadowVariable
         unsafeClassDivZero
+        # Disable because of too many FPs
+        arithOperationsOnVoidPointer
         definePrefix:*test/include/test.hpp
         ctuOneDefinitionRuleViolation:*test/*
         useSmartPointer:*src/api/api.cpp
         useSmartPointer:*make_shared_array.hpp
-        constParameter:*src/targets/gpu/*.cpp
-        constParameter:*src/targets/gpu/*.hpp
-        # Suppress mlir_conv.cpp since this file will be deleted
-        *:*src/targets/gpu/mlir_conv.cpp
     FORCE
     INCONCLUSIVE
     RULE_FILE
@@ -251,10 +250,13 @@ rocm_enable_cppcheck(
         ${CMAKE_CURRENT_SOURCE_DIR}/src/targets/gpu/kernels/include
         ${CMAKE_CURRENT_SOURCE_DIR}/test/include
     DEFINE
+        MIGRAPHX_MLIR=1
         CPPCHECK=1
         __device__=
         __host__=
         __global__=
+    UNDEFINE
+        MIGRAPHX_USE_CLANG_TIDY
 )
 
 enable_testing()

Dockerfile

@@ -114,7 +114,7 @@ RUN git clone --single-branch --branch ${ONNXRUNTIME_BRANCH} --recursive ${ONNXR
 ADD tools/build_and_test_onnxrt.sh /onnxruntime/build_and_test_onnxrt.sh
 
 # Use the /opt/cmake install because LLVM/MLIR need cmake >= 3.20
-RUN env PATH=/opt/cmake/bin:$PATH cget -p /usr/local install ROCmSoftwarePlatform/rocMLIR@1ad9d6df32acc6d29d58e8ed6710e36746d0a4d6 -DBUILD_FAT_LIBROCKCOMPILER=On
+RUN env PATH=/opt/cmake/bin:$PATH cget -p /usr/local install ROCmSoftwarePlatform/rocMLIR@ea15b3597ce55b9088621818228595dd48fb6ec0 -DBUILD_FAT_LIBROCKCOMPILER=On
 
 ENV MIOPEN_FIND_DB_PATH=/tmp/miopen/find-db
 ENV MIOPEN_USER_DB_PATH=/tmp/miopen/user-db

Jenkinsfile

@@ -26,7 +26,7 @@ def rocmtestnode(Map conf) {
             rm -rf build
             mkdir build
             cd build
-            cmake -DCMAKE_C_COMPILER_LAUNCHER=ccache -DCMAKE_CXX_COMPILER_LAUNCHER=ccache -DBUILD_DEV=On ${flags} ..
+            cmake -DCMAKE_C_COMPILER_LAUNCHER=ccache -DCMAKE_CXX_COMPILER_LAUNCHER=ccache -DBUILD_DEV=On -DCMAKE_EXECUTE_PROCESS_COMMAND_ECHO=STDOUT ${flags} ..
             git diff
             git diff-index --quiet HEAD || (echo "Git repo is not clean after running cmake." && exit 1)
             make -j\$(nproc) generate VERBOSE=1
@@ -114,6 +114,10 @@ rocmtest clang_debug: rocmnode('cdna') { cmake_build ->
         cmake_build(flags: "-DCMAKE_BUILD_TYPE=release")
         stash includes: 'build/*.deb', name: 'migraphx-package'
     }
+}, hidden_symbols: rocmnode('cdna') { cmake_build ->
+    stage('Hidden symbols') {
+        cmake_build(flags: "-DMIGRAPHX_ENABLE_PYTHON=Off -DMIGRAPHX_ENABLE_GPU=On -DMIGRAPHX_ENABLE_CPU=On -DCMAKE_CXX_VISIBILITY_PRESET=hidden -DCMAKE_C_VISIBILITY_PRESET=hidden")
+    }
 }, all_targets_debug : rocmnode('cdna') { cmake_build ->
     stage('All targets Release') {
         cmake_build(flags: "-DCMAKE_BUILD_TYPE=release -DMIGRAPHX_ENABLE_GPU=On -DMIGRAPHX_ENABLE_CPU=On -DMIGRAPHX_ENABLE_FPGA=On") 

cmake/Embed.cmake

@@ -24,11 +24,7 @@
 find_program(EMBED_LD ld)
 find_program(EMBED_OBJCOPY objcopy)
 
-if(LINUX)
-    option(EMBED_USE_LD "Use ld to embed data files" ON)
-else()
-    option(EMBED_USE_LD "Use ld to embed data files" OFF)
-endif()
+option(EMBED_USE_LD "Use ld to embed data files" OFF)
 
 function(wrap_string)
     set(options)
@@ -60,8 +56,8 @@ endfunction()
 
 function(generate_embed_source EMBED_NAME)
     set(options)
-    set(oneValueArgs SRC HEADER)
-    set(multiValueArgs OBJECTS SYMBOLS)
+    set(oneValueArgs SRC HEADER RELATIVE)
+    set(multiValueArgs OBJECTS SYMBOLS FILES)
 
     cmake_parse_arguments(PARSE "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
 
@@ -78,6 +74,8 @@ function(generate_embed_source EMBED_NAME)
     foreach(idx RANGE ${LEN})
         list(GET PARSE_SYMBOLS ${idx} SYMBOL)
         list(GET PARSE_OBJECTS ${idx} OBJECT)
+        list(GET PARSE_FILES ${idx} FILE)
+
         set(START_SYMBOL "_binary_${SYMBOL}_start")
         set(END_SYMBOL "_binary_${SYMBOL}_end")
         if(EMBED_USE_LD)
@@ -92,9 +90,11 @@ function(generate_embed_source EMBED_NAME)
             ")
         endif()
 
-        # TODO: Should use NAME_WLE
-        get_filename_component(BASE_NAME "${OBJECT}" NAME)
-        string(REGEX REPLACE ".[A-Za-z0-9_]+$" "" BASE_NAME ${BASE_NAME})
+        if(PARSE_RELATIVE)
+            file(RELATIVE_PATH BASE_NAME ${PARSE_RELATIVE} "${FILE}")
+        else()
+            get_filename_component(BASE_NAME "${FILE}" NAME)
+        endif()
 
         string(APPEND INIT_KERNELS "
             { \"${BASE_NAME}\", { ${START_SYMBOL}, ${END_SYMBOL}} },
@@ -162,6 +162,11 @@ function(embed_file OUTPUT_FILE OUTPUT_SYMBOL FILE)
 endfunction()
 
 function(add_embed_library EMBED_NAME)
+    set(options)
+    set(oneValueArgs RELATIVE)
+    set(multiValueArgs)
+    cmake_parse_arguments(PARSE "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
+
     file(MAKE_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/embed)
     file(MAKE_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}/embed/${EMBED_NAME})
     set(EMBED_DIR ${CMAKE_CURRENT_BINARY_DIR}/embed/${EMBED_NAME})
@@ -171,15 +176,26 @@ function(add_embed_library EMBED_NAME)
     set(OUTPUT_FILES)
     set(SYMBOLS)
     message(STATUS "Embedding files")
-    foreach(FILE ${ARGN})
+    foreach(FILE ${PARSE_UNPARSED_ARGUMENTS})
         embed_file(OUTPUT_FILE OUTPUT_SYMBOL ${FILE})
         list(APPEND OUTPUT_FILES ${OUTPUT_FILE})
         list(APPEND SYMBOLS ${OUTPUT_SYMBOL})
     endforeach()
     message(STATUS "Generating embedding library ${EMBED_NAME}")
-    generate_embed_source(${EMBED_NAME} SRC ${SRC_FILE} HEADER ${HEADER_FILE} OBJECTS ${OUTPUT_FILES} SYMBOLS ${SYMBOLS})
-    add_library(${EMBED_NAME} STATIC ${OUTPUT_FILES} "${SRC_FILE}")
-    target_include_directories(${EMBED_NAME} PUBLIC "${EMBED_DIR}/include")
-    target_compile_options(${EMBED_NAME} PRIVATE -Wno-reserved-identifier -Wno-extern-initializer -Wno-missing-variable-declarations)
-    set_target_properties(${EMBED_NAME} PROPERTIES POSITION_INDEPENDENT_CODE On)
+    generate_embed_source(${EMBED_NAME} SRC ${SRC_FILE} HEADER ${HEADER_FILE} OBJECTS ${OUTPUT_FILES} SYMBOLS ${SYMBOLS} RELATIVE ${PARSE_RELATIVE} FILES ${PARSE_UNPARSED_ARGUMENTS})
+    
+    set(INTERNAL_EMBED_LIB embed_lib_${EMBED_NAME})
+    add_library(${INTERNAL_EMBED_LIB} OBJECT "${SRC_FILE}")
+    target_include_directories(${INTERNAL_EMBED_LIB} PRIVATE "${EMBED_DIR}/include")
+    target_compile_options(${INTERNAL_EMBED_LIB} PRIVATE -Wno-reserved-identifier -Wno-extern-initializer -Wno-missing-variable-declarations)
+    set_target_properties(${INTERNAL_EMBED_LIB} PROPERTIES POSITION_INDEPENDENT_CODE On)
+    
+    add_library(${EMBED_NAME} INTERFACE)
+    if(EMBED_USE_LD)
+        target_sources(${EMBED_NAME} INTERFACE ${OUTPUT_FILES})
+    else()
+        target_sources(${INTERNAL_EMBED_LIB} PRIVATE ${OUTPUT_FILES})
+    endif()
+    target_sources(${EMBED_NAME} INTERFACE $<TARGET_OBJECTS:${INTERNAL_EMBED_LIB}>)
+    target_include_directories(${EMBED_NAME} INTERFACE "${EMBED_DIR}/include")
 endfunction()

cmake/ExportHeader.cmake

@@ -29,7 +29,10 @@ endif()
 include(GenerateExportHeader)
 
 function(migraphx_generate_export_header TARGET)
-    cmake_parse_arguments(PARSE "" "DIRECTORY" "" ${ARGN})
+    set(options)
+    set(oneValueArgs DIRECTORY)
+    set(multiValueArgs)
+    cmake_parse_arguments(PARSE "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
     if(PARSE_DIRECTORY)
         set(__directory ${PARSE_DIRECTORY})
     else()

cmake/PythonModules.cmake

@@ -38,12 +38,22 @@ macro(find_python version)
     find_program(PYTHON_CONFIG_${version} python${version}-config)
     if(EXISTS ${PYTHON_CONFIG_${version}})
         py_exec(COMMAND ${PYTHON_CONFIG_${version}} --includes OUTPUT_VARIABLE _python_include_args)
+        execute_process(COMMAND ${PYTHON_CONFIG_${version}} --ldflags --embed OUTPUT_VARIABLE _python_ldflags_args RESULT_VARIABLE _python_ldflags_result)
+        if(NOT _python_ldflags_result EQUAL 0)
+            py_exec(COMMAND ${PYTHON_CONFIG_${version}} --ldflags OUTPUT_VARIABLE _python_ldflags_args)
+        endif()
         separate_arguments(_python_includes UNIX_COMMAND "${_python_include_args}")
+        separate_arguments(_python_ldflags UNIX_COMMAND "${_python_ldflags_args}")
         string(REPLACE "-I" "" _python_includes "${_python_includes}")
         add_library(python${version}::headers INTERFACE IMPORTED GLOBAL)
         set_target_properties(python${version}::headers PROPERTIES
             INTERFACE_INCLUDE_DIRECTORIES "${_python_includes}"
         )
+        add_library(python${version}::runtime INTERFACE IMPORTED GLOBAL)
+        set_target_properties(python${version}::runtime PROPERTIES
+            INTERFACE_LINK_OPTIONS "${_python_ldflags}"
+            INTERFACE_LINK_LIBRARIES python${version}::headers
+        )
         py_exec(COMMAND ${PYTHON_CONFIG_${version}} --prefix OUTPUT_VARIABLE _python_prefix)
         string(STRIP "${_python_prefix}" _python_prefix)
         set(PYTHON_${version}_EXECUTABLE "${_python_prefix}/bin/python${version}" CACHE PATH "")

cppcheck.rules

@@ -107,6 +107,24 @@
         <summary>Use make_shared or make_unique instead of new</summary>
     </message>
 </rule>
+<rule>
+    <tokenlist>raw</tokenlist>
+    <pattern><![CDATA[ [^\(,;{}:]+ \w+ && (\w+|\(|\+|\-|\*)]]></pattern>
+    <message>
+        <id>UseNamedLogicOperator</id>
+        <severity>style</severity>
+        <summary>Use 'and' instead of &&</summary>
+    </message>
+</rule>
+<rule>
+    <tokenlist>raw</tokenlist>
+    <pattern><![CDATA[ (if|while) \([^\)&]+&& ]]></pattern>
+    <message>
+        <id>UseNamedLogicOperator</id>
+        <severity>style</severity>
+        <summary>Use 'and' instead of &&</summary>
+    </message>
+</rule>
 <rule>
     <tokenlist>raw</tokenlist>
     <pattern><![CDATA[ \|\| ]]></pattern>

dev-requirements.txt

@@ -25,6 +25,6 @@ ROCmSoftwarePlatform/rocm-recipes
 facebook/zstd@v1.4.5 -X subdir -DCMAKE_DIR=build/cmake
 ccache@v4.1 -DENABLE_TESTING=OFF
 pcre,pfultz2/pcre@8.45 -H sha256:d6f7182602a775a7d500a0cedca6449af0400c6493951513046d17615ed0bf11
-danmar/cppcheck@2.9 -DHAVE_RULES=1
+danmar/cppcheck@bb2711c22a0be09efe7f1a8da3030876471026c8 -DHAVE_RULES=1 # 2.11
 RadeonOpenCompute/rocm-cmake@027404a8326da6e7e9338e0b81f9428660190724 --build
 -f requirements.txt

docs/driver/read.rst

@@ -82,6 +82,10 @@ Print out program in text format.
 
 Print out program in binary format.
 
+.. option::  --py
+
+Print out program using python API.
+
 .. option::  --output, -o [std::string]
 
 Output to file.

examples/migraphx/cpp_parse_load_save/parse_load_save.cpp

@@ -114,7 +114,7 @@ int main(int argc, char** argv)
 char* getCmdOption(char** begin, char** end, const std::string& option)
 {
     char** itr = std::find(begin, end, option);
-    if(itr != end && ++itr != end)
+    if(itr != end and ++itr != end)
     {
         return *itr;
     }

examples/vision/cpp_mnist/mnist_inference.cpp

@@ -160,9 +160,9 @@ int main(int argc, char** argv)
     auto lengths = shape.lengths();
     auto num_results =
         std::accumulate(lengths.begin(), lengths.end(), 1, std::multiplies<size_t>());
-    float* results = reinterpret_cast<float*>(outputs[0].data());
-    float* max     = std::max_element(results, results + num_results);
-    int answer     = max - results;
+    float* results   = reinterpret_cast<float*>(outputs[0].data());
+    const float* max = std::max_element(results, results + num_results);
+    int answer       = max - results;
 
     std::cout << std::endl
               << "Randomly chosen digit: " << rand_digit << std::endl
@@ -192,12 +192,12 @@ void read_nth_digit(const int n, std::vector<float>& digit)
         for(int i = 0; i < HEIGHT * WIDTH; ++i)
         {
             unsigned char temp = 0;
-            file.read((char*)&temp, sizeof(temp));
+            file.read(reinterpret_cast<char*>(&temp), sizeof(temp));
             if(d == n)
             {
                 float data = temp / 255.0;
                 digit.push_back(data);
-                std::cout << SYMBOLS[(int)(data * 10) % 11];
+                std::cout << SYMBOLS[static_cast<int>(data * 10) % 11];
                 if((i + 1) % WIDTH == 0)
                     std::cout << std::endl;
             }

src/CMakeLists.txt

@@ -248,8 +248,6 @@ endif()
 target_link_libraries(migraphx PRIVATE -ldl)
 
 target_include_directories(migraphx SYSTEM PUBLIC $<BUILD_INTERFACE:${HALF_INCLUDE_DIR}>)
-
-find_package(Threads)
 target_link_libraries(migraphx PUBLIC Threads::Threads)
 
 find_package(nlohmann_json 3.8.0 REQUIRED)

src/api/CMakeLists.txt

@@ -26,6 +26,7 @@ add_library(migraphx_c
     api.cpp
 )
 set_target_properties(migraphx_c PROPERTIES EXPORT_NAME c)
+migraphx_generate_export_header(migraphx_c DIRECTORY migraphx/api)
 
 # migraphx_c is stable API interface library. SO version of this should be 
 # bumped when binary compatibility is broken. 

src/api/api.cpp

@@ -44,7 +44,7 @@ namespace migraphx {
 
 static thread_local bool disable_exception_catch = false; // NOLINT
 
-extern "C" void migraphx_test_private_disable_exception_catch(bool b)
+extern "C" MIGRAPHX_C_EXPORT void migraphx_test_private_disable_exception_catch(bool b)
 {
     disable_exception_catch = b;
 }
@@ -899,7 +899,7 @@ migraphx_dynamic_dimensions_assign_to(migraphx_dynamic_dimensions_t output,
 
 extern "C" migraphx_status
 migraphx_dynamic_dimensions_create(migraphx_dynamic_dimensions_t* dynamic_dimensions,
-                                   const_migraphx_dynamic_dimension_t* ptr,
+                                   const const_migraphx_dynamic_dimension_t* ptr,
                                    size_t size)
 {
     auto api_error_result = migraphx::try_([&] {
@@ -1432,7 +1432,7 @@ extern "C" migraphx_status migraphx_instructions_assign_to(migraphx_instructions
 }
 
 extern "C" migraphx_status migraphx_instructions_create(migraphx_instructions_t* instructions,
-                                                        const_migraphx_instruction_t* ptr,
+                                                        const const_migraphx_instruction_t* ptr,
                                                         size_t size)
 {
     auto api_error_result = migraphx::try_([&] {

src/api/migraphx.py

@@ -79,7 +79,8 @@ def dynamic_dimension(h):
 def dynamic_dimensions(h):
     h.constructor(
         'create',
-        api.params(ptr='const_migraphx_dynamic_dimension_t*', size='size_t'),
+        api.params(ptr='const const_migraphx_dynamic_dimension_t*',
+                   size='size_t'),
         fname='migraphx::to_obj_vector<const_migraphx_dynamic_dimension_t>')
     h.method('size', returns='size_t')
     h.method('get',
@@ -215,7 +216,7 @@ def instruction(h):
 def instructions(h):
     h.constructor(
         'create',
-        api.params(ptr='const_migraphx_instruction_t*', size='size_t'),
+        api.params(ptr='const const_migraphx_instruction_t*', size='size_t'),
         fname='migraphx::to_obj_vector<const_migraphx_instruction_t>')
 
 

src/driver/CMakeLists.txt

@@ -45,7 +45,7 @@ if(NOT WIN32)
 endif()
 rocm_clang_tidy_check(driver)
 
-target_link_libraries(driver migraphx_all_targets migraphx_onnx migraphx_tf)
+target_link_libraries(driver migraphx_all_targets migraphx_onnx migraphx_tf migraphx_py)
 
 rocm_install_targets(
   TARGETS driver

src/driver/argument_parser.hpp

@@ -338,11 +338,22 @@ struct argument_parser
 
     MIGRAPHX_DRIVER_STATIC auto file_exist()
     {
-        return validate([](auto&, auto&, auto& params) {
+        return validate([](auto&, auto&, const auto& params) {
             if(params.empty())
                 throw std::runtime_error("No argument passed.");
             if(not fs::exists(params.back()))
-                throw std::runtime_error("Path does not exists: " + params.back());
+                throw std::runtime_error("Path does not exist: " + params.back());
+        });
+    }
+
+    MIGRAPHX_DRIVER_STATIC auto matches(const std::unordered_set<std::string>& names)
+    {
+        return validate([=](auto&, auto&, const auto& params) {
+            auto invalid_param = std::find_if(
+                params.begin(), params.end(), [&](const auto& p) { return names.count(p) == 0; });
+            if(invalid_param != params.end())
+                throw std::runtime_error("Invalid argument: " + *invalid_param +
+                                         ". Valid arguments are {" + to_string_range(names) + "}");
         });
     }
 
@@ -570,8 +581,7 @@ struct argument_parser
                         continue;
                     if(flag[0] != '-')
                         continue;
-                    auto d =
-                        levenshtein_distance(flag.begin(), flag.end(), input.begin(), input.end());
+                    std::ptrdiff_t d = levenshtein_distance(flag, input);
                     if(d < result.distance)
                         result = result_t{&arg, flag, input, d};
                 }

src/driver/main.cpp

@@ -32,6 +32,7 @@
 
 #include <migraphx/tf.hpp>
 #include <migraphx/onnx.hpp>
+#include <migraphx/py.hpp>
 #include <migraphx/stringutils.hpp>
 #include <migraphx/convert_to_json.hpp>
 #include <migraphx/load_save.hpp>
@@ -81,6 +82,7 @@ struct loader
            {"--model"},
            ap.help("Load model"),
            ap.type("resnet50|inceptionv3|alexnet"),
+           ap.matches({"resnet50", "inceptionv3", "alexnet"}),
            ap.group("input"));
         ap(file_type, {"--onnx"}, ap.help("Load as onnx"), ap.set_value("onnx"));
         ap(file_type, {"--tf"}, ap.help("Load as tensorflow"), ap.set_value("tf"));
@@ -241,6 +243,20 @@ struct loader
         return options;
     }
 
+    static std::string get_file_type(const std::string& file)
+    {
+        if(ends_with(file, ".onnx"))
+            return "onnx";
+        else if(ends_with(file, ".pb"))
+            return "tf";
+        else if(ends_with(file, ".json"))
+            return "json";
+        else if(ends_with(file, ".py"))
+            return "py";
+        else
+            return "migraphx";
+    }
+
     program load()
     {
         program p;
@@ -248,14 +264,7 @@ struct loader
         {
             if(file_type.empty())
             {
-                if(ends_with(file, ".onnx"))
-                    file_type = "onnx";
-                else if(ends_with(file, ".pb"))
-                    file_type = "tf";
-                else if(ends_with(file, ".json"))
-                    file_type = "json";
-                else
-                    file_type = "migraphx";
+                file_type = get_file_type(file);
             }
             std::cout << "Reading: " << file << std::endl;
             if(file_type == "onnx")
@@ -272,6 +281,10 @@ struct loader
                 options.format = "json";
                 p              = migraphx::load(file, options);
             }
+            else if(file_type == "py")
+            {
+                p = migraphx::load_py(file);
+            }
             else if(file_type == "migraphx")
             {
                 p = migraphx::load(file);
@@ -757,7 +770,7 @@ struct main_command
         {
             std::cout << "'" << color::fg_yellow << wrong_commands.front() << color::reset
                       << "' is not a valid command." << std::endl;
-            std::cout << get_command_help("Available commands:") << std::endl;
+            std::cout << get_command_help("Available commands:");
         }
         else
         {

src/dynamic_loader.cpp

@@ -48,7 +48,7 @@ struct dynamic_loader_impl
 #pragma GCC diagnostic ignored "-Wignored-attributes"
 #endif
     dynamic_loader_impl(const fs::path& p, std::shared_ptr<tmp_dir> t = nullptr)
-        : handle(dlopen(p.string().c_str(), RTLD_LAZY),
+        : handle(dlopen(p.string().c_str(), RTLD_GLOBAL | RTLD_NOW),
                  manage_deleter<decltype(&dlclose), &dlclose>{}),
           temp(std::move(t))
     {
@@ -81,6 +81,18 @@ fs::path dynamic_loader::path(void* address)
     return p;
 }
 
+optional<dynamic_loader> dynamic_loader::try_load(const fs::path& p)
+{
+    try
+    {
+        return dynamic_loader{p};
+    }
+    catch(const std::exception&)
+    {
+        return nullopt;
+    }
+}
+
 dynamic_loader::dynamic_loader(const fs::path& p) : impl(std::make_shared<dynamic_loader_impl>(p))
 {
 }