How to allocate double pointer in CUDA programming?

Hello.

I want to know double point allocation.

for example,

int *first <- size 100 this is separated to range 50, 30, and 20.
int **second <- size 3 this pointer has 50’s start point, 30’s start point, and 20’s start point.

if this code is worked in C or C++, it is easy.

but CUDA allocation system is a little bit difficult for me, so I can’t imagine…

please help me.

add)
I found the guide!
https://stackoverflow.com/questions/26111794/how-to-use-pointer-to-pointer-in-cuda