NVIDIA Multi-Instance GPU

 

Seven independent instances in a single GPU.

Multi-Instance GPU (MIG) expands the performance and value of NVIDIA Blackwell and Hopper™ generation GPUs. MIG can partition the GPU into as many as seven instances, each fully isolated with its own high-bandwidth memory, cache, and compute cores. This gives administrators the ability to support every workload, from the smallest to the largest, with guaranteed quality of service (QoS) and extending the reach of accelerated computing resources to every user.

Benefits Overview

Expand GPU Access

With MIG, you can achieve up to 7X more GPU resources on a single GPU. MIG gives researchers and developers more resources and flexibility than ever before.

Optimize GPU Utilization

MIG provides the flexibility to choose many different instance sizes, which allows provisioning of the right-sized GPU instance for each workload, ultimately optimizing utilization and maximizing data center investment.

Run Simultaneous Workloads

MIG enables inference, training, and high-performance computing (HPC) workloads to run at the same time on a single GPU with deterministic latency and throughput. Unlike time slicing, each workload runs in parallel, delivering higher performance.

How the Technology Works

Without MIG, different jobs running on the same GPU, such as different AI inference requests, compete for the same resources. A job consuming larger memory bandwidth starves others, resulting in several jobs missing their latency targets. With MIG, jobs run simultaneously on different instances, each with dedicated resources for compute, memory, and memory bandwidth, resulting in predictable performance with QoS and maximum GPU utilization.

Provision and Configure Instances as Needed

A GPU can be partitioned into different-sized MIG instances. For example, on an NVIDIA GB200, an administrator could create two instances with 95GB of memory each, four instances with 45GB each, or seven instances with 23GB each.

MIG instances can also be dynamically reconfigured, enabling administrators to shift GPU resources in response to changing user and business demands. For example, seven MIG instances can be used during the day for low-throughput inference and reconfigured to one large MIG instance at night for deep learning training.

Run Workloads in Parallel, Securely

With a dedicated set of hardware resources for compute, memory, and cache, each MIG instance delivers guaranteed QoS and fault isolation. That means that a failure in an application running on one instance doesn’t impact applications running on other instances.

It also means that different instances can run different types of workloads—interactive model development, deep learning training, AI inference, or HPC applications. Since the instances run in parallel, the workloads also run in parallel—but separate and isolated—on the same physical GPU.

MIG in Blackwell GPUs

Blackwell and Hopper GPUs support MIG with multi-tenant, multi-user configurations in virtualized environments across up to seven GPU instances, securely isolating each instance with confidential computing at the hardware and hypervisor level. Dedicated video decoders for each MIG instance deliver secure, high-throughput intelligent video analytics (IVA) on shared infrastructure. With concurrent MIG profiling, administrators can monitor right-sized GPU acceleration and allocate resources for multiple users.

For researchers with smaller workloads, rather than renting a full cloud instance, they can use MIG to isolate a portion of a GPU securely while being assured that their data is secure at rest, in transit, and in use. This improves flexibility for cloud service providers to price and address smaller customer opportunities.

Built for IT and DevOps

MIG enables fine-grained GPU provisioning by IT and DevOps teams. Each MIG instance behaves like a standalone GPU to applications, so there’s no change to the CUDA® platform. MIG can be used in all major enterprise computing environments​.

Deploy from Data Center to Edge

Deploy from Data Center to Edge Use MIG on premises, in the cloud, and at the edge.

Leverage Containers

Run containerized applications on MIG instances​.

Support Kubernetes

Schedule Kubernetes pods on MIG instances​.

Virtualize Applications

Run applications on MIG instances inside a virtual machine​.

MIG Specifications

 
 GB200/B200/B100H100
Confidential computingYesYes
Instance typesUp to 7x 23GB
Up to 4x 45GB
Up to 2x 95GB
Up to 1x 192GB
7x 10GB
4x 20GB
2x 40GB
1x 80GB
GPU profiling and monitoringConcurrently on all instancesConcurrently on all instances
Secure Tenants7x7x
Media decodersDedicated NVJPEG and NVDEC per instanceDedicated NVJPEG and NVDEC per instance

Preliminary specifications, may be subject to change