What Is GPU Passthrough?
A technique that gives a virtual machine or container direct access to a physical GPU for hardware-accelerated computing.
Definition
A technique that gives a virtual machine or container direct access to a physical GPU for hardware-accelerated computing.
Essential for AI/ML development in CDEs, enabling model training and inference on cloud GPUs (NVIDIA H200, A100, AMD MI300X) from remote workspaces.
Where This Fits
This term is covered in depth on GPU and Accelerated Computing in CDEs.
GPU-accelerated CDEs for AI model training, 3D rendering, and scientific computing. NVIDIA A100, H100, H200 workspace provisioning.