modern-cmake/chapters/packages/CUDA.md

CUDA (in progress)

CUDA support is available in two flavors. The new method, introduced in CMake 3.8 (3.9 for Windows), will be what I focus on first. The old method will be covered afterwards, but as you'll see, it's uglier and harder to get right. I'd stick to requiring CMake 3.8 or 3.9 for CUDA.

A good resource for CUDA and Modern CMake is this talk by CMake developer Robert Maynard at GTC 2017.

Method 1: CUDA as a First Class Language

This method is quite new, and doesn't seem to have much documentation. There are several issues you need to watch out for when using it, but overall is should be a much nicer and cleaner way to use CUDA.

Adding the CUDA Language

There are two ways to enable CUDA support. If CUDA is not optional:

project(MY_PROJECT LANGUAGES CUDA CXX)

You'll probably want CXX listed here also. And, if CUDA is optional, you'll want to put this in somewhere conditionally:

enable_language(CUDA)

Variables for CUDA

Many variables with CXX in the name have a CUDA version with CUDA instead. For example, to set the C++ standard required for CUDA,

if(NOT DEFINED CMAKE_CUDA_STANDARD)
    set(CMAKE_CUDA_STANDARD 11)
    set(CMAKE_CUDA_STANDARD_REQUIRED ON)
endif()

Adding a library

This is the easy part; as long as you use .cu for CUDA files, you can just add libraries like you normally would.

You can also use separable compilation:

set_target_properties(mylib PROPERTIES
                            CUDA_SEPERABLE_COMPILATION ON)

You can also direcly make a PTX file with the CUDA_PTX_COMPILATION property.

Working with targets

Using targets should work similarly to CXX, but there's a problem. If you include a target that includes compiler options (flags), most of the time, the options will not be protected by the correct includes (and the chances of them having the correct CUDA wrapper is even smaller). Here's what a correct compiler options line should look like:

However, if you using almost any find_package, and using the Modern CMake methods of targets and inheritence, everything will break. I've learned that the hard way.

For now, here's a pretty reasonable solution, as long as you know the un-aliased target name. It's a function that will fix a C++ only target by wrapping the flags if using a CUDA compiler:

function(CUDA_CONVERT_FLAGS EXISTING_TARGET)
    get_property(old_flags TARGET ${EXISTING_TARGET} PROPERTY INTERFACE_COMPILE_OPTIONS)
    if(NOT "${old_flags}" STREQUAL "")
        string(REPLACE ";" "," CUDA_flags "${old_flags}")
        set_property(TARGET ${EXISTING_TARGET} PROPERTY INTERFACE_COMPILE_OPTIONS
            "$<$<BUILD_INTERFACE:$<COMPILE_LANGUAGE:CXX>>:${old_flags}>$<$<BUILD_INTERFACE:$<COMPILE_LANGUAGE:CUDA>>:-Xcompiler=${CUDA_flags}>"
            )
    endif()
endfunction()

Note that FindCUDA is deprecated, but for now, the following functions require FindCUDA:

• CUDA version checks / picking a version
• Architecture detection
• Linking to CUDA libraries from non-.cu files

Method 2: FindCUDA