Launching Large Language Models on PowerEdge servers