examples for using multiple GPUs and one CPU?

Accelerated Computing CUDA CUDA Programming and Performance
1

Where can one find examples for using multiple GPUs
and one CPU?
That is, using either a single thread or multiple threads on
a multi-core CPU to access multiple GPUs?

Can you give specific pointers?

Thank you!

2

NVIDIA SDK,

1)simpleMultiGPU: pthread

2)cudaOpenmp: OpenMP

3

Are these the only examples?

What are the advantages of pthreads vs. OpenMP?

Which one is easier to use with CUDA?