If you don't have the necessary hardware or you'd like to run your workloads for your GPU-enabled code in the cloud, you may decide that you'd rather use CUDA on a shared hosting environment. In the following example, we'll show you how to get set up using GPUs on GCP.
There are many other hosted GPU providers (you can see all of them listed in the GPU-accelerated computing – utilizing the hardware section of this chapter)—we are going to use GCP's GPU instances as an example here.
You can learn more about GCP's GPU offerings at https://cloud.google.com/gpu.