AMD ROCm Performance Primitives (RPP) library is a comprehensive high-performance computer vision library for AMD processors with HIP/OpenCL/CPU back-ends.
If the user has passed in the compiler to use for building, use it. Otherwise set it to clang++
On Fedora there is no /opt/rocm/llvm/bin/clang++, the compiler used is a system compiler wrapped by hipcc. So set COMPILER_FOR_HIP to llvm/bin/clang++ only if that file exists, otherwise use the CMAKE_CXX_COMPILER.
If the user has passed in the compiler to use for building, use it. Otherwise set it to clang++
On Fedora there is no /opt/rocm/llvm/bin/clang++, the compiler used is a system compiler wrapped by hipcc. So set COMPILER_FOR_HIP to llvm/bin/clang++ only if that file exists, otherwise use the CMAKE_CXX_COMPILER.