diff --git a/cmake/CMakeLists.txt b/cmake/CMakeLists.txt index cc63844f46..b03a301976 100644 --- a/cmake/CMakeLists.txt +++ b/cmake/CMakeLists.txt @@ -73,6 +73,11 @@ option(onnxruntime_ENABLE_PYTHON "Enable python buildings" OFF) # Enable it may cause LNK1169 error option(onnxruntime_ENABLE_MEMLEAK_CHECKER "Experimental: Enable memory leak checker in Windows debug build" OFF) option(onnxruntime_USE_CUDA "Build with CUDA support" OFF) +# Enable ONNX Runtime CUDA EP's internal unit tests that directly access the EP's internal functions instead of through +# OpKernels. When the option is ON, we will have two copies of GTest library in the same process. It is not a typical +# use. If you hit any problem with that, please do not report it to GTest. Turn OFF the following build option instead. +cmake_dependent_option(onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS "Build with CUDA unit tests" OFF "onnxruntime_USE_CUDA;onnxruntime_BUILD_UNIT_TESTS;LINUX" OFF) + option(onnxruntime_ENABLE_CUDA_LINE_NUMBER_INFO "When building with CUDA support, generate device code line number information." OFF) option(onnxruntime_USE_OPENVINO "Build with OpenVINO support" OFF) option(onnxruntime_USE_COREML "Build with CoreML support" OFF) @@ -146,10 +151,11 @@ option(onnxruntime_DISABLE_SPARSE_TENSORS "Disable sparse tensors data types" OF option(onnxruntime_DISABLE_OPTIONAL_TYPE "Disable optional type" OFF) option(onnxruntime_DISABLE_FLOAT8_TYPES "Disable float 8 types" OFF) option(onnxruntime_MINIMAL_BUILD "Exclude as much as possible from the build. Support ORT format models. No support for ONNX format models." OFF) -cmake_dependent_option(onnxruntime_DISABLE_RTTI "Disable RTTI" ON "NOT onnxruntime_ENABLE_PYTHON" OFF) +cmake_dependent_option(onnxruntime_DISABLE_RTTI "Disable RTTI" ON "NOT onnxruntime_ENABLE_PYTHON;NOT onnxruntime_USE_CUDA" OFF) # For now onnxruntime_DISABLE_EXCEPTIONS will only work with onnxruntime_MINIMAL_BUILD, more changes (ONNX, non-CPU EP, ...) are required to run this standalone cmake_dependent_option(onnxruntime_DISABLE_EXCEPTIONS "Disable exception handling. Requires onnxruntime_MINIMAL_BUILD currently." ON "onnxruntime_MINIMAL_BUILD;NOT onnxruntime_ENABLE_PYTHON" OFF) -option(onnxruntime_DISABLE_ABSEIL "Do not link to Abseil. Redefine Inlined containers to STD containers." OFF) +# Even when onnxruntime_DISABLE_ABSEIL is ON, ONNX Runtime still needs to link to abseil. +option(onnxruntime_DISABLE_ABSEIL "Do not use Abseil data structures in ONNX Runtime source code. Redefine Inlined containers to STD containers." OFF) option(onnxruntime_EXTENDED_MINIMAL_BUILD "onnxruntime_MINIMAL_BUILD with support for execution providers that compile kernels." OFF) option(onnxruntime_MINIMAL_BUILD_CUSTOM_OPS "Add custom operator kernels support to a minimal build." OFF) @@ -269,10 +275,6 @@ if (onnxruntime_ENABLE_TRAINING_APIS) endif() endif() -if (onnxruntime_USE_CUDA) - set(onnxruntime_DISABLE_RTTI OFF) -endif() - if (onnxruntime_USE_ROCM) if (WIN32) message(FATAL_ERROR "ROCM does not support build in Windows!") diff --git a/cmake/external/onnxruntime_external_deps.cmake b/cmake/external/onnxruntime_external_deps.cmake index e1671bcf43..019c6341d2 100644 --- a/cmake/external/onnxruntime_external_deps.cmake +++ b/cmake/external/onnxruntime_external_deps.cmake @@ -37,8 +37,12 @@ if (onnxruntime_BUILD_UNIT_TESTS) set(gtest_disable_pthreads ON) endif() set(INSTALL_GTEST OFF CACHE BOOL "" FORCE) - # Set it to ON will cause crashes in onnxruntime_test_all when onnxruntime_USE_CUDA is ON - set(GTEST_HAS_ABSL OFF CACHE BOOL "" FORCE) + if (CMAKE_SYSTEM_NAME STREQUAL "iOS") + # Needs to update onnxruntime/test/xctest/xcgtest.mm + set(GTEST_HAS_ABSL OFF CACHE BOOL "" FORCE) + else() + set(GTEST_HAS_ABSL ON CACHE BOOL "" FORCE) + endif() # gtest and gmock FetchContent_Declare( googletest diff --git a/cmake/onnxruntime_providers.cmake b/cmake/onnxruntime_providers.cmake index 4861643832..96c05e5282 100644 --- a/cmake/onnxruntime_providers.cmake +++ b/cmake/onnxruntime_providers.cmake @@ -460,13 +460,17 @@ if (onnxruntime_USE_CUDA) if (onnxruntime_REDUCED_OPS_BUILD) substitute_op_reduction_srcs(onnxruntime_providers_cuda_src) endif() - # cuda_provider_interface.cc is removed from the object target: onnxruntime_providers_cuda_obj and - # add to the lib onnxruntime_providers_cuda separatedly. - # onnxruntime_providers_cuda_ut can share all the object files with onnxruntime_providers_cuda except cuda_provider_interface.cc. - set(cuda_provider_interface_src ${ONNXRUNTIME_ROOT}/core/providers/cuda/cuda_provider_interface.cc) - list(REMOVE_ITEM onnxruntime_providers_cuda_src ${cuda_provider_interface_src}) - onnxruntime_add_object_library(onnxruntime_providers_cuda_obj ${onnxruntime_providers_cuda_src}) - onnxruntime_add_shared_library_module(onnxruntime_providers_cuda ${cuda_provider_interface_src} $) + if(onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS) + # cuda_provider_interface.cc is removed from the object target: onnxruntime_providers_cuda_obj and + # add to the lib onnxruntime_providers_cuda separatedly. + # onnxruntime_providers_cuda_ut can share all the object files with onnxruntime_providers_cuda except cuda_provider_interface.cc. + set(cuda_provider_interface_src ${ONNXRUNTIME_ROOT}/core/providers/cuda/cuda_provider_interface.cc) + list(REMOVE_ITEM onnxruntime_providers_cuda_src ${cuda_provider_interface_src}) + onnxruntime_add_object_library(onnxruntime_providers_cuda_obj ${onnxruntime_providers_cuda_src}) + onnxruntime_add_shared_library_module(onnxruntime_providers_cuda ${cuda_provider_interface_src} $) + else() + onnxruntime_add_shared_library_module(onnxruntime_providers_cuda ${onnxruntime_providers_cuda_src}) + endif() # config_cuda_provider_shared_module can be used to config onnxruntime_providers_cuda_obj, onnxruntime_providers_cuda & onnxruntime_providers_cuda_ut. # This function guarantees that all 3 targets have the same configurations. function(config_cuda_provider_shared_module target) @@ -600,7 +604,9 @@ if (onnxruntime_USE_CUDA) target_compile_definitions(${target} PRIVATE ENABLE_ATEN) endif() endfunction() - config_cuda_provider_shared_module(onnxruntime_providers_cuda_obj) + if(onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS) + config_cuda_provider_shared_module(onnxruntime_providers_cuda_obj) + endif() config_cuda_provider_shared_module(onnxruntime_providers_cuda) install(TARGETS onnxruntime_providers_cuda diff --git a/cmake/onnxruntime_unittests.cmake b/cmake/onnxruntime_unittests.cmake index ec83eb2095..0e642c5a7e 100644 --- a/cmake/onnxruntime_unittests.cmake +++ b/cmake/onnxruntime_unittests.cmake @@ -35,13 +35,17 @@ function(AddTest) if (MSVC AND NOT CMAKE_SIZEOF_VOID_P EQUAL 8) #TODO: fix the warnings, they are dangerous - target_compile_options(${_UT_TARGET} PRIVATE "/wd4244") + target_compile_options(${_UT_TARGET} PRIVATE "$<$:SHELL:--compiler-options /wd4244>" + "$<$>:/wd4244>") endif() if (MSVC) - target_compile_options(${_UT_TARGET} PRIVATE "/wd6330") + target_compile_options(${_UT_TARGET} PRIVATE "$<$:SHELL:--compiler-options /wd6330>" + "$<$>:/wd6330>") #Abseil has a lot of C4127/C4324 warnings. - target_compile_options(${_UT_TARGET} PRIVATE "/wd4127") - target_compile_options(${_UT_TARGET} PRIVATE "/wd4324") + target_compile_options(${_UT_TARGET} PRIVATE "$<$:SHELL:--compiler-options /wd4127>" + "$<$>:/wd4127>") + target_compile_options(${_UT_TARGET} PRIVATE "$<$:SHELL:--compiler-options /wd4324>" + "$<$>:/wd4324>") endif() set_target_properties(${_UT_TARGET} PROPERTIES FOLDER "ONNXRuntimeTest") @@ -60,6 +64,11 @@ function(AddTest) Threads::Threads) target_compile_definitions(${_UT_TARGET} PRIVATE -DUSE_ONNXRUNTIME_DLL) else() + if(onnxruntime_USE_CUDA) + #XXX: we should not need to do this. onnxruntime_test_all.exe should not have direct dependency on CUDA DLLs, + # otherwise it will impact when CUDA DLLs can be unloaded. + target_link_libraries(${_UT_TARGET} PRIVATE cudart) + endif() target_link_libraries(${_UT_TARGET} PRIVATE ${_UT_LIBS} GTest::gtest GTest::gmock ${onnxruntime_EXTERNAL_LIBRARIES}) endif() @@ -85,29 +94,22 @@ function(AddTest) # include dbghelp in case tests throw an ORT exception, as that exception includes a stacktrace, which requires dbghelp. target_link_libraries(${_UT_TARGET} PRIVATE debug dbghelp) - if (onnxruntime_USE_CUDA) - # disable a warning from the CUDA headers about unreferenced local functions - if (MSVC) - target_compile_options(${_UT_TARGET} PRIVATE "$<$:-Xcompiler /wd4505>" - "$<$>:/wd4505>") - endif() - endif() if (MSVC) # warning C6326: Potential comparison of a constant with another constant. # Lot of such things came from gtest - target_compile_options(${_UT_TARGET} PRIVATE "$<$:-Xcompiler /wd6326>" + target_compile_options(${_UT_TARGET} PRIVATE "$<$:SHELL:--compiler-options /wd6326>" "$<$>:/wd6326>") # Raw new and delete. A lot of such things came from googletest. - target_compile_options(${_UT_TARGET} PRIVATE "$<$:-Xcompiler /wd26409>" + target_compile_options(${_UT_TARGET} PRIVATE "$<$:SHELL:--compiler-options /wd26409>" "$<$>:/wd26409>") # "Global initializer calls a non-constexpr function." - target_compile_options(${_UT_TARGET} PRIVATE "$<$:-Xcompiler /wd26426>" + target_compile_options(${_UT_TARGET} PRIVATE "$<$:SHELL:--compiler-options /wd26426>" "$<$>:/wd26426>") endif() target_compile_options(${_UT_TARGET} PRIVATE ${disabled_warnings}) else() target_compile_options(${_UT_TARGET} PRIVATE ${DISABLED_WARNINGS_FOR_TVM}) - target_compile_options(${_UT_TARGET} PRIVATE "$<$:SHELL:-Xcompiler -Wno-error=sign-compare>" + target_compile_options(${_UT_TARGET} PRIVATE "$<$:SHELL:--compiler-options -Wno-error=sign-compare>" "$<$>:-Wno-error=sign-compare>") target_compile_options(${_UT_TARGET} PRIVATE "-Wno-error=uninitialized") endif() @@ -698,7 +700,7 @@ onnxruntime_add_static_library(onnxruntime_test_utils ${onnxruntime_test_utils_s if(MSVC) target_compile_options(onnxruntime_test_utils PRIVATE "$<$:SHELL:--compiler-options /utf-8>" "$<$>:/utf-8>") - target_compile_options(onnxruntime_test_utils PRIVATE "$<$:-Xcompiler /wd6326>" + target_compile_options(onnxruntime_test_utils PRIVATE "$<$:SHELL:--compiler-options /wd6326>" "$<$>:/wd6326>") else() target_compile_definitions(onnxruntime_test_utils PUBLIC -DNSYNC_ATOMIC_CPP11) @@ -755,13 +757,8 @@ set_target_properties(onnx_test_runner_common PROPERTIES FOLDER "ONNXRuntimeTest set(all_tests ${onnxruntime_test_common_src} ${onnxruntime_test_ir_src} ${onnxruntime_test_optimizer_src} ${onnxruntime_test_framework_src} ${onnxruntime_test_providers_src} ${onnxruntime_test_quantiztion_src}) -if(NOT TARGET onnxruntime AND NOT CMAKE_SYSTEM_NAME STREQUAL "Emscripten") - list(APPEND all_tests ${onnxruntime_shared_lib_test_SRC}) -endif() -if (onnxruntime_USE_CUDA) - onnxruntime_add_static_library(onnxruntime_test_cuda_ops_lib ${ONNXRUNTIME_SHARED_LIB_TEST_SRC_DIR}/cuda_ops.cu) - list(APPEND onnxruntime_test_common_libs onnxruntime_test_cuda_ops_lib) +if (onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS) file(GLOB onnxruntime_test_providers_cuda_ut_src CONFIGURE_DEPENDS "${TEST_SRC_DIR}/providers/cuda/test_cases/*" ) @@ -822,7 +819,9 @@ if (onnxruntime_USE_TENSORRT) # made test name contain the "ep" and "model path" information, so we can easily filter the tests using cuda ep or other ep with *cpu_* or *xxx_*. list(APPEND test_all_args "--gtest_filter=-*cpu_*:*cuda_*" ) endif () - +if(NOT onnxruntime_ENABLE_CUDA_EP_INTERNAL_TESTS) + list(REMOVE_ITEM all_tests ${TEST_SRC_DIR}/providers/cuda/cuda_provider_test.cc) +endif() AddTest( TARGET onnxruntime_test_all SOURCES ${all_tests} ${onnxruntime_unittest_main_src} @@ -832,11 +831,15 @@ AddTest( DEPENDS ${all_dependencies} TEST_ARGS ${test_all_args} ) + if (MSVC) # The warning means the type of two integral values around a binary operator is narrow than their result. # If we promote the two input values first, it could be more tolerant to integer overflow. # However, this is test code. We are less concerned. - target_compile_options(onnxruntime_test_all PRIVATE "/wd26451" "/wd4244") + target_compile_options(onnxruntime_test_all PRIVATE "$<$:SHELL:--compiler-options /wd26451>" + "$<$>:/wd26451>") + target_compile_options(onnxruntime_test_all PRIVATE "$<$:SHELL:--compiler-options /wd4244>" + "$<$>:/wd4244>") else() target_compile_options(onnxruntime_test_all PRIVATE "-Wno-parentheses") endif() @@ -1092,18 +1095,18 @@ if (NOT onnxruntime_ENABLE_TRAINING_TORCH_INTEROP) target_include_directories(onnxruntime_benchmark PRIVATE ${ONNXRUNTIME_ROOT} ${onnxruntime_graph_header} ${ONNXRUNTIME_ROOT}/core/mlas/inc) target_compile_definitions(onnxruntime_benchmark PRIVATE BENCHMARK_STATIC_DEFINE) if(WIN32) - target_compile_options(onnxruntime_benchmark PRIVATE "$<$:-Xcompiler /wd4141>" + target_compile_options(onnxruntime_benchmark PRIVATE "$<$:SHELL:--compiler-options /wd4141>" "$<$>:/wd4141>") # Avoid using new and delete. But this is a benchmark program, it's ok if it has a chance to leak. - target_compile_options(onnxruntime_benchmark PRIVATE "$<$:-Xcompiler /wd26409>" + target_compile_options(onnxruntime_benchmark PRIVATE "$<$:SHELL:--compiler-options /wd26409>" "$<$>:/wd26409>") - target_compile_options(onnxruntime_benchmark PRIVATE "$<$:-Xcompiler /wd26400>" + target_compile_options(onnxruntime_benchmark PRIVATE "$<$:SHELL:--compiler-options /wd26400>" "$<$>:/wd26400>") - target_compile_options(onnxruntime_benchmark PRIVATE "$<$:-Xcompiler /wd26814>" + target_compile_options(onnxruntime_benchmark PRIVATE "$<$:SHELL:--compiler-options /wd26814>" "$<$>:/wd26814>") - target_compile_options(onnxruntime_benchmark PRIVATE "$<$:-Xcompiler /wd26814>" + target_compile_options(onnxruntime_benchmark PRIVATE "$<$:SHELL:--compiler-options /wd26814>" "$<$>:/wd26497>") - target_compile_options(onnxruntime_benchmark PRIVATE "$<$:-Xcompiler /wd26426>" + target_compile_options(onnxruntime_benchmark PRIVATE "$<$:SHELL:--compiler-options /wd26426>" "$<$>:/wd26426>") target_compile_options(onnxruntime_benchmark PRIVATE "$<$:SHELL:--compiler-options /utf-8>" "$<$>:/utf-8>") @@ -1255,7 +1258,7 @@ if (NOT onnxruntime_ENABLE_TRAINING_TORCH_INTEROP) list(APPEND onnxruntime_shared_lib_test_LIBS cpuinfo) endif() if (onnxruntime_USE_CUDA) - list(APPEND onnxruntime_shared_lib_test_LIBS onnxruntime_test_cuda_ops_lib cudart) + list(APPEND onnxruntime_shared_lib_test_LIBS cudart) endif() if (onnxruntime_USE_TENSORRT) list(APPEND onnxruntime_shared_lib_test_LIBS ${TENSORRT_LIBRARY_INFER}) @@ -1270,7 +1273,10 @@ if (NOT onnxruntime_ENABLE_TRAINING_TORCH_INTEROP) LIBS ${onnxruntime_shared_lib_test_LIBS} DEPENDS ${all_dependencies} ) - + if (onnxruntime_USE_CUDA) + target_include_directories(onnxruntime_shared_lib_test PRIVATE ${CMAKE_CUDA_TOOLKIT_INCLUDE_DIRECTORIES}) + target_sources(onnxruntime_shared_lib_test PRIVATE ${ONNXRUNTIME_SHARED_LIB_TEST_SRC_DIR}/cuda_ops.cu) + endif() if (CMAKE_SYSTEM_NAME STREQUAL "Android") target_sources(onnxruntime_shared_lib_test PRIVATE "${ONNXRUNTIME_ROOT}/core/platform/android/cxa_demangle.cc" @@ -1356,13 +1362,13 @@ if (NOT onnxruntime_ENABLE_TRAINING_TORCH_INTEROP) ) onnxruntime_add_executable(onnxruntime_mlas_test ${onnxruntime_mlas_test_src}) if(MSVC) - target_compile_options(onnxruntime_mlas_test PRIVATE "$<$:-Xcompiler /wd26409>" + target_compile_options(onnxruntime_mlas_test PRIVATE "$<$:SHELL:--compiler-options /wd26409>" "$<$>:/wd26409>") target_compile_options(onnxruntime_mlas_test PRIVATE "$<$:SHELL:--compiler-options /utf-8>" "$<$>:/utf-8>") - target_compile_options(onnxruntime_mlas_test PRIVATE "$<$:-Xcompiler /wd6326>" + target_compile_options(onnxruntime_mlas_test PRIVATE "$<$:SHELL:--compiler-options /wd6326>" "$<$>:/wd6326>") - target_compile_options(onnxruntime_mlas_test PRIVATE "$<$:-Xcompiler /wd26426>" + target_compile_options(onnxruntime_mlas_test PRIVATE "$<$:SHELL:--compiler-options /wd26426>" "$<$>:/wd26426>") endif() if(${CMAKE_SYSTEM_NAME} STREQUAL "iOS") @@ -1476,7 +1482,7 @@ if (NOT CMAKE_SYSTEM_NAME STREQUAL "Emscripten") "${TEST_SRC_DIR}/testdata/custom_op_library/cuda/cuda_ops.*") list(APPEND custom_op_lib_include ${CMAKE_CUDA_TOOLKIT_INCLUDE_DIRECTORIES} ${onnxruntime_CUDNN_HOME}/include) if (HAS_QSPECTRE) - list(APPEND custom_op_lib_option "$<$:SHELL:-Xcompiler /Qspectre>") + list(APPEND custom_op_lib_option "$<$:SHELL:--compiler-options /Qspectre>") endif() endif() @@ -1503,7 +1509,7 @@ if (NOT CMAKE_SYSTEM_NAME STREQUAL "Emscripten") else() set(ONNXRUNTIME_CUSTOM_OP_LIB_LINK_FLAG "-DEF:${TEST_SRC_DIR}/testdata/custom_op_library/custom_op_library.def") if (NOT onnxruntime_USE_CUDA) - target_compile_options(custom_op_library PRIVATE "$<$:-Xcompiler /wd26409>" + target_compile_options(custom_op_library PRIVATE "$<$:SHELL:--compiler-options /wd26409>" "$<$>:/wd26409>") endif() endif() diff --git a/onnxruntime/core/platform/windows/debug_alloc.cc b/onnxruntime/core/platform/windows/debug_alloc.cc index b08d189f79..2b612a4303 100644 --- a/onnxruntime/core/platform/windows/debug_alloc.cc +++ b/onnxruntime/core/platform/windows/debug_alloc.cc @@ -12,7 +12,7 @@ // #ifndef NDEBUG #ifdef ONNXRUNTIME_ENABLE_MEMLEAK_CHECK -constexpr int c_callstack_limit = 16; // Maximum depth of callstack in leak trace +constexpr int c_callstack_limit = 32; // Maximum depth of callstack in leak trace #define VALIDATE_HEAP_EVERY_ALLOC 0 // Call HeapValidate on every new/delete #pragma warning(disable : 4073) // initializers put in library initialization area (this is intentional) @@ -223,6 +223,10 @@ Memory_LeakCheck::~Memory_LeakCheck() { // empty_group_names = new std::map; }); if (string.find("RtlRunOnceExecuteOnce") == std::string::npos && string.find("re2::RE2::Init") == std::string::npos && + string.find("dynamic initializer for 'FLAGS_") == std::string::npos && + string.find("AbslFlagDefaultGenForgtest_") == std::string::npos && + string.find("::SetProgramUsageMessage") == std::string::npos && + string.find("testing::internal::ParseGoogleTestFlagsOnly") == std::string::npos && string.find("testing::internal::Mutex::ThreadSafeLazyInit") == std::string::npos && string.find("testing::internal::ThreadLocalRegistryImpl::GetThreadLocalsMapLocked") == std::string::npos && string.find("testing::internal::ThreadLocalRegistryImpl::GetValueOnCurrentThread") == std::string::npos && diff --git a/onnxruntime/test/contrib_ops/qordered_attention_test.cc b/onnxruntime/test/contrib_ops/qordered_attention_test.cc index 24e4bff528..1dd0162ad7 100644 --- a/onnxruntime/test/contrib_ops/qordered_attention_test.cc +++ b/onnxruntime/test/contrib_ops/qordered_attention_test.cc @@ -240,12 +240,6 @@ static std::vector transpose(const T& src, size_t h, size_t w) { } TEST(QOrderedTest, Attention_WithData_ROW_ORDER) { - int cuda_runtime_version = 0; - // Need 11.4 or higher cuda runtime - if ((cudaRuntimeGetVersion(&cuda_runtime_version) != cudaSuccess) || (cuda_runtime_version < 11040)) { - return; - } - // Needs Turing architecture if (NeedSkipIfCudaArchLowerThan(750) || NeedSkipIfCudaArchGreaterEqualThan(800)) { return; diff --git a/onnxruntime/test/contrib_ops/qordered_longformer_attention_op_test.cc b/onnxruntime/test/contrib_ops/qordered_longformer_attention_op_test.cc index 55209d9422..06fe42ca98 100644 --- a/onnxruntime/test/contrib_ops/qordered_longformer_attention_op_test.cc +++ b/onnxruntime/test/contrib_ops/qordered_longformer_attention_op_test.cc @@ -34,12 +34,6 @@ static void run_qordered_longformer_attention_op_test( const int64_t head_size, const int64_t window, int64_t input_hidden_size = 0) { - int cuda_runtime_version = 0; - // Need 11.4 or higher cuda runtime - if ((cudaRuntimeGetVersion(&cuda_runtime_version) != cudaSuccess) || (cuda_runtime_version < 11040)) { - return; - } - // Needs Turing architecture if (NeedSkipIfCudaArchLowerThan(750) || NeedSkipIfCudaArchGreaterEqualThan(800)) { return; diff --git a/onnxruntime/test/contrib_ops/qordered_matmul_op_test.cc b/onnxruntime/test/contrib_ops/qordered_matmul_op_test.cc index e5b3d59ef8..e3905db635 100644 --- a/onnxruntime/test/contrib_ops/qordered_matmul_op_test.cc +++ b/onnxruntime/test/contrib_ops/qordered_matmul_op_test.cc @@ -21,12 +21,6 @@ static void RunQOrdered_MatMul_Test( OrderCublasLt weight_order, float scale_A, float scale_B, float scale_C, float scale_Y, bool add_bias = false, bool broadcast_c_batch = false, bool per_channel = false) { - int cuda_runtime_version = 0; - // Need 11.4 or higher cuda runtime - if ((cudaRuntimeGetVersion(&cuda_runtime_version) != cudaSuccess) || (cuda_runtime_version < 11040)) { - return; - } - // Needs Turing architecture if (NeedSkipIfCudaArchLowerThan(750) || NeedSkipIfCudaArchGreaterEqualThan(800)) { return; diff --git a/onnxruntime/test/contrib_ops/qordered_qdq_op_test.cc b/onnxruntime/test/contrib_ops/qordered_qdq_op_test.cc index 15e97751ac..0f3f702695 100644 --- a/onnxruntime/test/contrib_ops/qordered_qdq_op_test.cc +++ b/onnxruntime/test/contrib_ops/qordered_qdq_op_test.cc @@ -73,12 +73,6 @@ static void RunQOrdered_Quantize_Test( std::vector const& shape, OrderCublasLt order_q, float scale) { - int cuda_runtime_version = 0; - // Need 11.4 or higher cuda runtime - if ((cudaRuntimeGetVersion(&cuda_runtime_version) != cudaSuccess) || (cuda_runtime_version < 11040)) { - return; - } - auto qvec = QuantizeTransform(shape, scale, fvec, order_q); std::vector> execution_providers; @@ -153,12 +147,6 @@ static void RunQOrdered_Dequantize_Test( std::vector const& shape, OrderCublasLt order_q, float scale) { - int cuda_runtime_version = 0; - // Need 11.4 or higher cuda runtime - if ((cudaRuntimeGetVersion(&cuda_runtime_version) != cudaSuccess) || (cuda_runtime_version < 11040)) { - return; - } - auto fvec = DequantizeTransform(shape, scale, qvec, order_q); std::vector> execution_providers; diff --git a/onnxruntime/test/providers/cpu/model_tests.cc b/onnxruntime/test/providers/cpu/model_tests.cc index 9b41ba8c0d..da906ebf76 100644 --- a/onnxruntime/test/providers/cpu/model_tests.cc +++ b/onnxruntime/test/providers/cpu/model_tests.cc @@ -1172,6 +1172,12 @@ static ORT_STRING_VIEW provider_name_dml = ORT_TSTR("dml"); ORT_TSTR("bvlc_alexnet"), ORT_TSTR("bvlc_reference_caffenet"), ORT_TSTR("coreml_VGG16_ImageNet"), + ORT_TSTR("VGG 16-fp32"), + ORT_TSTR("VGG 19-caffe2"), + ORT_TSTR("VGG 19-bn"), + ORT_TSTR("VGG 16-bn"), + ORT_TSTR("VGG 19"), + ORT_TSTR("VGG 16"), ORT_TSTR("faster_rcnn"), ORT_TSTR("GPT2"), ORT_TSTR("GPT2_LM_HEAD"), diff --git a/onnxruntime/test/xctest/xcgtest.mm b/onnxruntime/test/xctest/xcgtest.mm index 5367f3e89c..c02f18d906 100644 --- a/onnxruntime/test/xctest/xcgtest.mm +++ b/onnxruntime/test/xctest/xcgtest.mm @@ -201,7 +201,7 @@ static void RunTest(id self, SEL _cmd) { delete listeners.Release(listeners.default_result_printer()); free(argv); - BOOL runDisabledTests = testing::GTEST_FLAG(also_run_disabled_tests); + BOOL runDisabledTests = GTEST_FLAG_GET(also_run_disabled_tests); NSMutableDictionary* testFilterMap = [NSMutableDictionary dictionary]; NSCharacterSet* decimalDigitCharacterSet = [NSCharacterSet decimalDigitCharacterSet];