How to .... Using CMAKE

Accelerated Computing CUDA CUDA Programming and Performance
1

Ok I want to fix some parameters to be used when Release mode is selected and others when Debug mode is selected. But I also want that the user could modify this parameters using ccmake o gui-make or …

I’m supposing that I should fill the variables ‘CUDA_NVCC_FLAGS_DEBUG’ and ‘CUDA_NVCC_FLAGS_RELEASE’ and cmake of cuda will used the appropiate variable (is this right???)

Second question… How to modify. To be more clear I’am appending elements to the variables in this way:

LIST(APPEND CUDA_NVCC_FLAGS_RELEASE -DNDEBUG )
LIST(APPEND CUDA_NVCC_FLAGS_RELEASE -O3 )
LIST(APPEND CUDA_NVCC_FLAGS_RELEASE --use_fast_math )

But later when I execute the ccmake the variable is empty. why?? and more important… how could show the value???

2

See build - Building CUDA object files using cmake - Stack Overflow