A high-throughput and memory-efficient inference and serving engine for LLMs

Common recipes to run vLLM

Google Cloud Karpenter Provider

OpenAIOS vGPU scheduler for Kubernetes is originated from the OpenAIOS project to virtualize GPU device memory.

HAMi-core compiles libvgpu.so, which ensures hard limit on GPU in container

A shim driver allows in-docker nvidia-smi showing correct process list without modify anything

Karmada APIs

Hooked CUDA-related dynamic libraries by using automated code generation tools.

SCUDA is a GPU over IP bridge allowing GPUs on remote machines to be attached to CPU-only machines.

eBPF-based Networking, Security, and Observability

Open, Multi-Cloud, Multi-Cluster Kubernetes Orchestration

很多镜像都在国外。比如 gcr 。国内下载很慢，需要加速。

helm repo add daocloud https://daocloud.github.io/dce-charts-repackage/

Home for Cluster API, a subproject of sig-cluster-lifecycle

Kubernetes-native Job Queueing

Connect, secure, control, and observe services.

Here you can find community content around Cilium

DaoCloud Enterprise Open Documents

Ingress-NGINX Controller for Kubernetes

Defines the shared interfaces which Karmada cloud providers implement. These interfaces allow various controllers to integrate with any cloud provider in a pluggable fashion.

Load Balancer Implementation for Kubernetes in Bare-Metal, Edge, and Virtualization

The Encyclopedia of Kubernetes clusters

Karmada website and documentation repo