Run CUDA code without a capable CUDA card

Hi,

it seems that device emulation is no longer supported, so I wonder what happens if I run CUDA C code without a capable CUDA graphics card? For instance, in a matrix multiplication operation what values would be stored in the result matrix?

Thanks,

JP

If you try to run device code without a device present? I would guess the context will not be created at all. So you would not be able to malloc the result array at all.

If you try to run device code without a device present? I would guess the context will not be created at all. So you would not be able to malloc the result array at all.