How to call C functions within CUDA kernel?

Accelerated Computing CUDA CUDA Programming and Performance
1

Hello, I’m trying to port C codes to CUDA environment. I see that we have to allocate memory from the device to the GPU. How do you allocate memory for several functions to the GPU? What is easy way to port C codes to CUDA?

Also, what is value “1” means in this code? Is it the “Dimension”?

2

Nevermind,
I got it:

device
global
host