Ampere Architecture vGPU Types

NVIDIA Ampere data center and workstation GPUs use second-generation RT Cores, third-generation Tensor Cores, and CUDA cores for training, inference, HPC, and analytics. Many SKUs support Multi-Instance GPU (MIG) hardware partitioning; others are time-sliced only, as shown per board in the family pages below.

MIG-backed vGPU types reference named GPU instance profiles (slice geometry and memory). Time-sliced types divide the framebuffer across guests on one physical GPU. NVIDIA A16 is the Ampere exception with more than one physical GPU per board.

See also

• MIG-Backed vGPU - MIG-backed vGPU concepts and constraints

• Scheduling Policies - vGPU scheduling policies for time-sliced types

• Configuration - Host MIG provisioning and guest reconfiguration steps

Browse Ampere vGPU types by SKU family:

Table 120 Ampere SKU families

Family	Boards
Ampere Workstation and General Compute vGPU Types	A40, A16, A10, RTX A6000, RTX A5500, RTX A5000
Ampere A800 and AX800 vGPU Types	A800 PCIe / HGX variants and AX800
Ampere A100 vGPU Types	A100 PCIe / HGX variants and A100X
Ampere A30 vGPU Types	A30, A30X, A30 liquid-cooled