# MortalGPU

MortalGPU is a Kubernetes device plugin that enables multiple workloads to share a single NVIDIA GPU, giving operators a way to run multiple GPU-accelerated containers on one accelerator. It is open source and self-hosted, installable on Kubernetes (Helm).

Canonical page: https://ourlocalhost.com/app/4892/mortalgpu

## Description

MortalGPU is a Kubernetes device plugin that enables multiple workloads to share a single NVIDIA GPU, giving operators a way to run multiple GPU-accelerated containers on one accelerator.

## Platform availability

Available on 1 of 10 platforms: Helm.

## Links

- Kubernetes (Helm): https://artifacthub.io/packages/helm/mortalgpu/mortalgpu

## Common questions

### Can MortalGPU be installed from a Helm chart?

Yes. The Helm package is listed as version 1.3.7.

### Can MortalGPU expose a MIG partition as a separate GPU device?

Yes. It looks for available GPUs and represents each GPU device, physical or MIG partition, with a configurable amount of meta-devices.

### Does MortalGPU share Nvidia GPUs between Kubernetes workloads?

Yes. It is a Kubernetes device plugin implementing the sharing of Nvidia GPUs between workloads.

### Is MortalGPU suitable for batch workloads such as AI/ML or scientific software?

Yes. It provides generic GPU support for batch workloads such as AI/ML or scientific software.

---
Source: Our Local Host (https://ourlocalhost.com/app/4892/mortalgpu), the self-hosted and open-source app directory.
