Complete all prerequisites.
| Deployment Option | Description | Guide |
|---|---|---|
| Single Node | Quick start for testing or lightweight workloads on a single Intel® Xeon® server | Single Node Guide |
| Single Master, Multiple Workers | For higher throughput workloads with one control-plane and N worker nodes | Multi-Node Guide |
| Docker Compose | Simplified single-node deployment for development, demos, and resource-constrained environments | Docker Guide |
| Feature | Description | Guide |
|---|---|---|
| Hybrid Router | Complexity-aware, cost/latency-optimizing inference routing for the GenAI Gateway | Hybrid Router Guide |
| Intel® Arc™ (XPU) GPU | Deploy vLLM models on Intel discrete GPUs (Arc/Flex/Max, incl. Battlemage) via compute_platform=xpu |
Intel Arc XPU Guide |
| Stack Validation | Post-deploy health checks, preflight node readiness, burn tests, and traffic-light reporting | Validate Stack Guide |