Commit c568fa66 authored by Cianciosa, Mark's avatar Cianciosa, Mark
Browse files

Fix problems with LLVM JIT on Linux. Adjust unit test tolarance to account for... 

Fix problems with LLVM JIT on Linux. Adjust unit test tolarance to account for errors on the order of machine precision.
parent 0bc8657b

graph_framework/CMakeLists.txt

+2 −2
Original line number Diff line number Diff line
@@ -13,8 +13,8 @@ target_compile_definitions (rays 
                            INTERFACE

                            $<$<PLATFORM_ID:Darwin>:CXX="${CMAKE_CXX_COMPILER} -isysroot ${CMAKE_OSX_SYSROOT} -I${CMAKE_CURRENT_SOURCE_DIR} -std=gnu++2a">

                            $<$<PLATFORM_ID:Darwin>:CXX_FLAGS="-g -fsanitize=undefined -fsanitize=float-divide-by-zero -fsanitize-trap=all">

                            $<$<PLATFORM_ID:Linux>:CXX="${CMAKE_CXX_COMPILER} -I${CMAKE_CURRENT_SOURCE_DIR} -std=gnu++2a">

                            $<$<PLATFORM_ID:Linux>:CXX_FLAGS="-g -fsanitize=undefined -fsanitize=float-divide-by-zero">

                            $<$<PLATFORM_ID:Linux>:CXX="${CMAKE_CXX_COMPILER} -I${CMAKE_CURRENT_SOURCE_DIR} -std=gnu++2a -fPIC -fno-use-cxa-atexit">

                            $<$<PLATFORM_ID:Linux>:CXX_FLAGS="-g"> #-fsanitize=undefined -fsanitize=float-divide-by-zero">

                            EFIT_FILE="${CMAKE_CURRENT_SOURCE_DIR}/../graph_tests/efit.nc"

                            VMEC_FILE="${CMAKE_CURRENT_SOURCE_DIR}/../graph_tests/vmec.nc"

                            $<$<BOOL:${USE_CUDA}>:HEADER_DIR="$<BUILD_INTERFACE:${CMAKE_CURRENT_SOURCE_DIR}>">

graph_tests/backend_test.cpp

+21 −11
Original line number Diff line number Diff line
@@ -404,8 +404,10 @@ template<jit::float_scalar T> void test_backend() { 
    exp_scalar.set(static_cast<T> (-0.23));

    const backend::buffer<T> pow_gen = backend::pow(base_scalar, exp_scalar);

    assert(pow_gen.size() == 1 && "Expected a size of 1");

    assert(pow_gen.at(0) == std::pow(static_cast<T> (4.0),

                                     static_cast<T> (-0.23)) &&

    assert(std::abs(pow_gen.at(0) -

                    std::pow(static_cast<T> (4.0),

                             static_cast<T> (-0.23))) <=

           std::abs(static_cast<T> (1.1102230246251565e-16)) &&

           "Expected 4^-0.23.");



    backend::buffer<T> base_vec(std::vector<T> ({
@@ -452,8 +454,10 @@ template<jit::float_scalar T> void test_backend() { 
    exp_scalar.set(static_cast<T> (-0.23));

    const backend::buffer<T> x_over_cubed_vec = backend::pow(base_vec, exp_scalar);

    assert(x_over_cubed_vec.size() == 2 && "Expected a size of 2");

    assert(x_over_cubed_vec.at(0) == std::pow(static_cast<T> (4.0),

                                              static_cast<T> (-0.23)) &&

    assert(std::abs(x_over_cubed_vec.at(0) -

                    std::pow(static_cast<T> (4.0),

                             static_cast<T> (-0.23))) <=

           std::abs(static_cast<T> (1.1102230246251565e-16)) &&

           "Expected 4^-0.23.");

    assert(x_over_cubed_vec.at(1) == std::pow(static_cast<T> (2.0),

                                              static_cast<T> (-0.23)) &&
@@ -466,11 +470,15 @@ template<jit::float_scalar T> void test_backend() { 
    }));

    backend::buffer<T> scale_base = backend::pow(base_scalar, exp_vec);

    assert(scale_base.size() == 2 && "Expected a size of 2");

    assert(scale_base.at(0) == std::pow(static_cast<T> (4.0),

                                        static_cast<T> (4.0)) &&

    assert(std::abs(scale_base.at(0) -

                    std::pow(static_cast<T> (4.0),

                             static_cast<T> (4.0))) <=

           std::abs(static_cast<T> (5.6843418860808015e-14)) &&

           "Expected 4^4.");

    assert(scale_base.at(1) == std::pow(static_cast<T> (4.0),

                                        static_cast<T> (2.0)) &&

    assert(std::abs(scale_base.at(1) -

                    std::pow(static_cast<T> (4.0),

                             static_cast<T> (2.0))) <=

           std::abs(static_cast<T> (1.7763568394002505e-15)) &&

           "Expected 4^2.");



    const auto non_int = static_cast<T> (0.438763);
@@ -499,8 +507,10 @@ template<jit::float_scalar T> void test_backend() { 
    }));

    const backend::buffer<T> vec_vec = backend::pow(base_vec, exp_vec);

    assert(vec_vec.size() == 2 && "Expected a size of 2");

    assert(vec_vec.at(0) == std::pow(static_cast<T> (4.0),

                                     static_cast<T> (-4.0)) &&

    assert(std::abs(vec_vec.at(0) -

                    std::pow(static_cast<T> (4.0),

                             static_cast<T> (-4.0))) <=

           std::abs(static_cast<T> (8.6736173798840355e-19)) &&

           "Expected 4^-4.");

    assert(vec_vec.at(1) == std::pow(static_cast<T> (2.0),

                                     static_cast<T> (0.30)) &&

graph_tests/solver_test.cpp

+1 −1
Original line number Diff line number Diff line
@@ -109,7 +109,7 @@ int main(int argc, const char * argv[]) { 
    (void)argv;

    run_tests<float> (4.0E-15);

    run_tests<double> (1.0E-30);

    run_tests<std::complex<float>> (3.0E-15);

    run_tests<std::complex<float>> (4.0E-15);

    run_tests<std::complex<double>> (1.0E-30);

    END_GPU

}