Playing in the cloud
Nvidia provides computational resources hosted in the cloud to execute the programs of our tutorial on external GPUs located on Amazon AWS services. We have enabled a lab within the qwik platform maintained by Nvidia on the Web to centralize the access to these resources. The following video illustrates the process to get the corresponding token which provides you access to a GPU (IP address for the machine hosting the GPU, login and password). From that on, you have to login via a SSH (Secure Shell) client installed by default on every Linux and MacOs distribution (ssh command). Windows users may use the Putty utility instead (free download).
This video shows how to access a GPU in the cloud, and once connected, how to carry out the basic development cycle for this tutorial, that is, compile + execute + modify-the-code + compile + execute + ... All files containing CUDA source codes are already allocated in the file system, so you do not have to worry about it. Just go there and learn how to compile and execute.