Arcitecture of NVIDIA GPU

I want to know about the Architecture of NVIDIA GPU. Can we say that it is shared memory parallel architecture?

Thanks

There is a list of documents here on the Nvidia developer site that you may find useful.