Using a remote NVIDIA GPU from a Mac: llama.cpp RPC, exo, served endpoints, CUDA forwarding
Parent: Apple Mac x NVIDIA eGPU (TinyGPU, CUDA userspace, Thunderbolt, legacy drivers) · Topic entry · 1 branch
A published reference is not available for this topic yet.
Also known as: llama.cpp RPC Mac CUDA, RDMA over Thunderbolt