Getting Cuda and CMake to run together

From NAMIC Wiki
Jump to: navigation, search
Home < Getting Cuda and CMake to run together

Cuda

o Download and install Cuda SDK, Cuda Toolkit from Cuda Downloads o Download and install CMake

         o Execute CMake
               + Windows: start->CMake2.4->CMake)
                     # Set the Source Directory
                     # Set the Binary Directory
                     # Push Configure
                           * Accept all warnings and read them FOUND_CUT and FOUND_FUT_INCLUDE should not be found
                     # Edit the values. They should be found by default in
                           * FOUND_CUT: C:\[Programdir]\NVidia Corporation\NVidia Cuda SDK\bin\win32\*\cutil32*.dll
                           * FOUND_CUT_INCLUDE: c:\[Programdir]\NVidia Corporation\NVidia Cuda SDK\common\inc
                     # Push Configure again
                     # Push Ok (when all the errors are resolved
                     # Open the produced snl file in the binary directory using Visual Studio
               + Linux ccmake .      [In the directory you are working in]
                     # hit 'c' for configure
                     # set the following folders:
                           * FOUND_CUT: [CudaSDK-Install-Dir]/bin/libcutil.a
                           * FOUND_CUT_INCLUDE: [CudaSDK-Install-Dir]/common/inc
                     # hit 'c' again
                     # hit 'g' for Generate
                     # make and execute.