NVIDIA's new PAIR beta is a free, open source tool that finds compatible PCs on a local network and routes independent inference requests to whichever system has capacity. NVIDIA says it works with Ollama and LM Studio and supports Windows, macOS, and Linux, plus RTX 20-series GPUs and newer, RTX PRO workstation GPUs, DGX Spark, and Apple M4 or newer.
The tool is aimed at local agent workflows that split a task into smaller jobs. NVIDIA also says its IFA updates bring simpler local model setup to Hermes Agent, OpenClaw, and Perplexity Portable Computer, and up to 1.9x higher throughput for llama.cpp on a GeForce RTX 5090. Those are vendor-reported figures. The useful test is whether multi-PC routing improves real workflows without making setup and privacy harder to manage.
Source: https://blogs.nvidia.com/blog/local-ai-ifa-next-gen-agents-nv-pair-rtx-spark/
[link] [comments]