volcano-sh/volcano
📦 Open Source Projectvolcano-sh
Cloud-native batch system for AI, ML, and HPC workloads on Kubernetes.
Volcano is an open-source batch system built on Kubernetes, providing advanced scheduling capabilities for AI, machine learning, big data, and HPC workloads. It introduces scheduling mechanisms beyond default Kubernetes, such as fair-share, priority, and resource reservation. Volcano handles Job orchestration, resource management, and lifecycle management, enabling efficient execution of distributed training, data processing, and simulation tasks. Key features include: heterogeneous resource support, GPU scheduling, gang scheduling, and pluggable scheduling policies. It integrates with popular frameworks like TensorFlow, PyTorch, Spark, and MPI. As a CNCF project, it is production-grade and widely adopted.
💡Highlights
- ├─CNCF-graduated batch system
- ├─Advanced gang & fair-share scheduling
- └─Integrates with Spark, TF, PyTorch
🎯For
- ├─DevOps professionals
- ├─Data scientists
- └─HPC engineers