Move parallel primitives lib: moderngpu to CUB

cuda_macro.h

src/CMakeLists.txt

src/memory/CudaMemory.cu

src/util/cudify/alpaka/cudify_alpaka.hpp

src/util/cudify/alpaka/cudify_hardware_alpaka.hpp

src/util/cudify/cuda/cudify_cuda.hpp

src/util/cudify/cuda/operators.hpp

src/util/cudify/cudify.hpp

src/util/cudify/cudify_hardware_cpu.hpp

src/util/cudify/cudify_vars.cpp

src/util/cudify/hip/cudify_hip.hpp

src/util/cudify/openmp/cudify_openmp.hpp

src/util/cudify/sequential/cudify_sequencial.hpp

src/util/gpu_context.hpp

src/util/gpu_types.hpp

src/util/ofp_context.hpp