Skip to main content
riseunion
ZH | EN
HAMi · CNCF Sandbox

Open Source Tech

HAMi (Heterogeneous AI Computing Virtualization Middleware) is a CNCF Sandbox project that brings GPU sharing and scheduling to Kubernetes. It supports NVIDIA, Ascend, Cambricon, Hygon DCU, and more — production-proven at hundreds of enterprises worldwide.

CNCF
Sandbox
10+
Chip Support
100s
Production Deployments

Open Source Tech

HAMi v2.8.0 Released: Dual Evolution in Standardization and Ecosystem Completeness

HAMi v2.8.0 Released: Dual Evolution in Standardization and Ecosystem Completeness

2/4/2026 Read more →

From 'Compute Available' to 'Compute Efficient': RiseUnion Deep Participation in HAMi Meetup Beijing

From 'Compute Available' to 'Compute Efficient': RiseUnion Deep Participation in HAMi Meetup Beijing

12/29/2025 Read more →

HAMi Beijing Meetup: Optimize Efficiency, Not Scale

HAMi Beijing Meetup: Optimize Efficiency, Not Scale

12/24/2025 Read more →

HAMi 2.4.0 Major Update: Making AI Computing Management More Efficient

HAMi 2.4.0 Major Update: Making AI Computing Management More Efficient

1/1/2025 Read more →

HAMi v2.6.0: Stronger Support for Chinese GPUs, Enhanced Stability and Observability

HAMi v2.6.0: Stronger Support for Chinese GPUs, Enhanced Stability and Observability

1/1/2025 Read more →

HAMi v2.7.0: Comprehensive Heterogeneous Hardware Support

HAMi v2.7.0: Comprehensive Heterogeneous Hardware Support

1/1/2025 Read more →

HAMi v2.5.0 Released: Dynamic MIG Support and Enhanced Stability

HAMi v2.5.0 Released: Dynamic MIG Support and Enhanced Stability

1/1/2025 Read more →

HAMi Dynamic MIG Design Principles

HAMi Dynamic MIG Design Principles

1/1/2025 Read more →

Open Source vGPU Solution HAMi: Core & Memory Isolation Test

Open Source vGPU Solution HAMi: Core & Memory Isolation Test

1/1/2025 Read more →

[Q&A] HAMi Frequently Asked Questions - Series 1

[Q&A] HAMi Frequently Asked Questions - Series 1

1/1/2025 Read more →

HAMi-core Compute Partitioning Mechanism: Feedback-Based Time-Slicing

HAMi-core Compute Partitioning Mechanism: Feedback-Based Time-Slicing

1/1/2025 Read more →

HAMi vGPU code Analysis Part 1: hami-device-plugin-nvidia

HAMi vGPU code Analysis Part 1: hami-device-plugin-nvidia

1/1/2025 Read more →

HAMi Community First Offline Salon Successfully Held

HAMi Community First Offline Salon Successfully Held

1/1/2025 Read more →

Ascend NPU Virtualization Guide: 910 Series & 310P Best Practices

Ascend NPU Virtualization Guide: 910 Series & 310P Best Practices

1/1/2025 Read more →

HAMi-core CUDA Compatibility: Cross-Version Stability

HAMi-core CUDA Compatibility: Cross-Version Stability

1/1/2025 Read more →

HAMi vGPU code Analysis Part2: hami-webhook

HAMi vGPU code Analysis Part2: hami-webhook

1/1/2025 Read more →

HAMi Source Code Analysis: Device Management and Scheduling

HAMi Source Code Analysis: Device Management and Scheduling

1/1/2025 Read more →

HAMi Configuration Guide: GPU Resource Pool Management

HAMi Configuration Guide: GPU Resource Pool Management

1/1/2025 Read more →

Iluvatar GPU Virtualization Guide: MR-V100/BI-V150 Best Practices

Iluvatar GPU Virtualization Guide: MR-V100/BI-V150 Best Practices

1/1/2025 Read more →

Kunlunxin P800 Virtualization Guide

Kunlunxin P800 Virtualization Guide

1/1/2025 Read more →

Enflame VGCU Virtualization Guide

Enflame VGCU Virtualization Guide

1/1/2025 Read more →

How HAMi(GPU Virtualization Technology) Can Save You Money!

How HAMi(GPU Virtualization Technology) Can Save You Money!

1/1/2025 Read more →

Complete Guide: Deploy HAMi-Core DRA Driver in kind

Complete Guide: Deploy HAMi-Core DRA Driver in kind

1/1/2025 Read more →

HAMi Introduces Dynamic MIG Support for NVIDIA GPUs

HAMi Introduces Dynamic MIG Support for NVIDIA GPUs

1/1/2025 Read more →

HAMi: Open Source GPU Virtualization for AI Computing

HAMi: Open Source GPU Virtualization for AI Computing

1/1/2025 Read more →