CUDA/HIP support in VkFFT

-VkFFT now supports three backends - Vulkan, CUDA and HIP to cover a wider range of devices (VKFFT_BACKEND in CMakeLists). All use runtime shader/kernel generation by the means of glslang compiler/NVRTC/HIPRTC
- CUDA/HIP support is in beta, some systems are not producing correct result (while Vulkan works), some extra features are not implemented yet (multiple buffer input/output, half-precision, input/output automatic type conversion, 4x register overutilization, cache shuffling, big convolutions also not tested)
-QoL changes: simplified startup configuration of VkFFT - secondary parameters are default initialized to optimized values for easier use of the library (can be changed manually). Forward/inverse direction can be selected at kernel launch. Automatic temporary buffer allocation in Four-step algorithm (can be done manually)
-Performance improvements: LUT reuse, improved general kernel stability and speed

CMakeLists.txt

@@ -1,7 +1,8 @@
-cmake_minimum_required(VERSION 3.11)
+cmake_minimum_required(VERSION 3.10)
 project(Vulkan_FFT)
 set(CMAKE_CONFIGURATION_TYPES "Release" CACHE STRING "" FORCE)
 set(CMAKE_BUILD_TYPE "Release" CACHE STRING "" FORCE)
+set(VKFFT_BACKEND 0)#0 - Vulkan, 1 - CUDA, 2 - HIP
 option(build_VkFFT_cuFFT_benchmark "Build VkFFT cuFFT benchmark" OFF)
 option(build_VkFFT_FFTW_precision "Build VkFFT FFTW precision comparison" OFF)
 if (MSVC)
@@ -10,29 +11,64 @@ if (MSVC)
 endif()
 
 add_executable(${PROJECT_NAME} Vulkan_FFT.cpp)
-
 target_compile_features(${PROJECT_NAME} PUBLIC cxx_std_11)
-
-find_package(Vulkan REQUIRED)
-target_compile_definitions(${PROJECT_NAME} PUBLIC -DVK_API_VERSION=11)#10 - Vulkan 1.0, 11 - Vulkan 1.1, 12 - Vulkan 1.2 
-
-target_include_directories(${PROJECT_NAME} PUBLIC ${CMAKE_CURRENT_SOURCE_DIR}/glslang-master/glslang/Include/)
-add_subdirectory(${CMAKE_CURRENT_SOURCE_DIR}/glslang-master)
+add_definitions(-DVKFFT_BACKEND=${VKFFT_BACKEND})
+if(${VKFFT_BACKEND} EQUAL 0)
+	find_package(Vulkan REQUIRED)
+elseif(${VKFFT_BACKEND} EQUAL 1)
+	find_package(CUDA 11.0 REQUIRED)
+	enable_language(CUDA)
+	#set_source_files_properties(Vulkan_FFT.cpp PROPERTIES LANGUAGE CUDA)
+	set_property(TARGET ${PROJECT_NAME} PROPERTY CUDA_ARCHITECTURES 35 60 70 75 80 86)
+	target_compile_options(${PROJECT_NAME} PUBLIC "$<$<COMPILE_LANGUAGE:CUDA>:SHELL:
+	-DVKFFT_BACKEND=${VKFFT_BACKEND}
+	-gencode arch=compute_35,code=compute_35
+	-gencode arch=compute_60,code=compute_60
+	-gencode arch=compute_70,code=compute_70 
+	-gencode arch=compute_75,code=compute_75 
+	-gencode arch=compute_80,code=compute_80 
+	-gencode arch=compute_86,code=compute_86>")
+	set_target_properties(${PROJECT_NAME} PROPERTIES CUDA_SEPARABLE_COMPILATION ON)
+	set_target_properties(${PROJECT_NAME} PROPERTIES CUDA_RESOLVE_DEVICE_SYMBOLS ON)
+elseif(${VKFFT_BACKEND} EQUAL 2)
+	list(APPEND CMAKE_PREFIX_PATH /opt/rocm/hip /opt/rocm)
+	find_package(hip)
+    	find_library(HIP_hiprtc_LIBRARY hiprtc
+      		PATHS
+      		ENV ROCM_PATH
+       		ENV HIP_PATH
+      		/opt/rocm
+      		PATH_SUFFIXES lib
+       		NO_DEFAULT_PATH)
+	add_library(hiprt UNKNOWN IMPORTED)
+	set_target_properties(hiprt PROPERTIES
+		INTERFACE_LINK_LIBRARIES ${HIP_hiprtc_LIBRARY}
+	)
+endif()
+target_compile_definitions(${PROJECT_NAME} PUBLIC -DVK_API_VERSION=10)#10 - Vulkan 1.0, 11 - Vulkan 1.1, 12 - Vulkan 1.2 
+if(${VKFFT_BACKEND} EQUAL 0)
+	target_include_directories(${PROJECT_NAME} PUBLIC ${CMAKE_CURRENT_SOURCE_DIR}/glslang-master/glslang/Include/)
+	add_subdirectory(${CMAKE_CURRENT_SOURCE_DIR}/glslang-master)
+endif()
+target_include_directories(${PROJECT_NAME} PUBLIC ${CMAKE_CURRENT_SOURCE_DIR}/vkFFT/)
+target_include_directories(${PROJECT_NAME} PUBLIC ${CMAKE_CURRENT_SOURCE_DIR}/half_lib/)
 
 add_library(VkFFT INTERFACE)
-target_include_directories(VkFFT INTERFACE ${CMAKE_CURRENT_SOURCE_DIR}/vkFFT/)
-target_include_directories(VkFFT INTERFACE ${CMAKE_CURRENT_SOURCE_DIR}/glslang-master/glslang/Include/)
-
+target_compile_definitions(VkFFT INTERFACE -DVKFFT_BACKEND=${VKFFT_BACKEND})
 add_library(half INTERFACE)
-target_include_directories(${PROJECT_NAME} PUBLIC ${CMAKE_CURRENT_SOURCE_DIR}/half_lib/)
-
-target_link_libraries(${PROJECT_NAME} PUBLIC SPIRV glslang Vulkan::Vulkan VkFFT half)
-
+if(${VKFFT_BACKEND} EQUAL 0)
+	target_link_libraries(${PROJECT_NAME} PUBLIC SPIRV glslang Vulkan::Vulkan VkFFT half)
+elseif(${VKFFT_BACKEND} EQUAL 1)
+	find_library(CUDA_NVRTC_LIB libnvrtc nvrtc HINTS "${CUDA_TOOLKIT_ROOT_DIR}/lib64" "${LIBNVRTC_LIBRARY_DIR}" "${CUDA_TOOLKIT_ROOT_DIR}/lib/x64" /usr/lib64 /usr/local/cuda/lib64)
+	target_link_libraries(${PROJECT_NAME} PUBLIC cuda ${CUDA_NVRTC_LIB} VkFFT half)
+elseif(${VKFFT_BACKEND} EQUAL 2)
+	target_link_libraries(${PROJECT_NAME} PUBLIC hip::host VkFFT half)
+endif()
 
 if(build_VkFFT_FFTW_precision)
 	add_definitions(-DUSE_FFTW)
-	set(FFTW3_LIB_DIR "")
-	set(FFTW3_INCLUDE_DIR "")
+	set(FFTW3_LIB_DIR "/usr/lib/x86_64-linux-gnu/")
+	set(FFTW3_INCLUDE_DIR "/usr/include/")
 	find_library(
 		FFTW_LIB
 		NAMES "libfftw3-3" "fftw3"
@@ -55,7 +91,7 @@ endif()
 if(build_VkFFT_cuFFT_benchmark)
 	add_definitions(-DUSE_cuFFT)
 	find_package(CUDA 11.0 REQUIRED)
-	enable_language(CUDA)
+	enable_language(CUDA)	
 	if(build_VkFFT_FFTW_precision)
 		add_library(cufft_scripts STATIC
 		cufft_scripts/benchmark_cuFFT.cu