Easy-GPU-PV is an open-source project that automates the relatively complex steps required to set up GPU Paravirtualization (GPU-PV) on Windows Hyper-V, letting you partition a physical GPU and assign slices to virtual machines for low-latency remote use. It creates and installs a Windows VM, copies the necessary GPU-PV driver files into the VM, and helps install Parsec (an ultra-low-latency remote-desktop client) so you can connect to the VM at high framerates. The repo is practical and script-driven: the author provides PowerShell scripts (PreChecks.ps1, CopyFilesToVM.ps1, Update-VMGpuPartitionDriver.ps1, etc.) plus configuration values and an unattended install template to minimize manual steps. The README documents hardware and OS prerequisites (Windows 10 20H1+/Windows 11 Pro/Enterprise/Education, matched host/VM versions, virtualization and Hyper-V enabled, recent GPU drivers, and supported GPUs.
Features
- Automates creating a Hyper-V VM and performing an unattended Windows install
- Copies and configures GPU-PV driver files into the VM for GPU partitioning
- PowerShell toolset with PreChecks, CopyFilesToVM, and Update-VMGpuPartitionDriver scripts
- Integrated Parsec installation guidance to enable low-latency remote access to the VM
- Explicit hardware and OS prerequisites plus recommended driver/update workflow
- Configurable parameters for VM size, disk, CPU cores, GPU allocation percentage, and Parsec team keys
Project Samples
