Passing C++ Objects to CUDA kernels Is the accessing of the Object members same as usual Cpp code??

I have a set of cpp object and would like to perform computation on that.

Can someone tell if the accessing procedure is exactly same as normal cpp or what is it…!!?
Can I access the member function and other member variable like any other code…?? if its different , how can this be accomplished??

Awaiting reply…!!

i tried to do that aswell but came up with some errors…
i know that passing STRCUTS is legal , for example the float3 float4 for example…

by the way , how did you manage to split your code into CPP class and still make the CUDA work properly?
i’ve tried to implement the STRATEGY design pattern and came up with some few problems…

should all cuda kernels be in one CU file?