Physics-based LLM capacity planner - size a deployment before provisioning hardware. Roofline model → replica count, cost envelope, and confidence band. Benchmark harness validates estimates against real endpoints and compares configs on latency, cost, and quality.
benchmark nextjs capacity-planning mlops roofline-model fastapi openai-api llm llm-serving vllm llm-inference sglang gpu-sizing
-
Updated
Jun 22, 2026 - Python