can find my GPU p3.2xlarge

You might need to install the NVIDIA drivers to enable the operating system to recognize and use the GPU's capabilities.

When you launch a p3.2xlarge instance on Windows Server, the physical GPU hardware is already present and connected to the instance, but the operating system doesn't have the necessary software to interact with it yet.

After installed driver, you can use nvidia-smi to view its status and utilization.

You can find some information about NVIDIA dirver here: https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/install-nvidia-driver.html

You will need to install NVIDIA driver. Refer to Install NVIDIA drivers on Windows instances, Public NVIDIA Driver for instructions.

Alternatively, the blog Building a high-performance Windows workstation on AWS for graphics intensive applications may help. It provides step-by-step instructions for launching Windows GPU instances, and includes a CloudFormation template that install NICE DCV server for remote access, and GPU drivers.

With the CloudFormation template, specify p3.2xlarge for instanceType, and NVIDIA-Tesla for driverType. Screenshot below