HAMi Project

Heterogeneous GPU Sharing on Kubernetes

HAMi website, join the community https://discord.gg/Amhy7XmbNq

Fast migration from Device Plugin to DRA for HAMi users.

HAMi GPU virtualization demo on GKE: vLLM + YOLOv8 multi-workload inference with fractional A100 sharing, Terraform IaC

Device-plugin for volcano vgpu which support hard resource isolation

GPU memory isolation for KAI-Scheduler GPU-sharing workloads, powered by HAMi-core.

Kubernetes device plugin for Ascend NPUs, supporting hard and soft NPU slicing with HAMi and Volcano.

Kubernetes device plugin for AMD GPUs, supporting soft vGPU with HAMi.

A transparent, in-container GPU resource controller that enforces memory and compute limits by intercepting CUDA calls without application or driver changes.

Open-source GPU resource management and observability platform for Kubernetes, built on HAMi.