Termsslug: multi-instance-gpu

Multi-Instance GPU

多实例 GPU

An NVIDIA hardware partitioning technology that splits one physical GPU into multiple isolated instances, each with dedicated co

Curated Intelligence

1 items · Page 1/1
AWS Machine Learning BlogT1·Infra
62 pts
SageMaker Inference Adds Multi-Instance GPU Partitioning

Amazon SageMaker Inference now supports GPU partitioning via NVIDIA Multi-Instance GPU (MIG), splitting a physical GPU into hardware-isolated partitions each with a fixed share of compute and dedicated memory. This removes the whole-accelerator minimum, letting teams host more models per instance or request a fraction of a newer GPU instead of over-provisioning.