Apple Silicon and macOS VMs: 11–16× Faster LLM Inference with Llama.cpp

Re:Apple Silicon and macOS VMs: 11–16× Faster LLM Inference ... by thehamkercat on Tuesday August 11, 2026 @11:11AM (#49259643)

> 11.08× faster and generated tokens 16.36× faster than the same workload in the same stock VM.

So this was the comparison, for me the title was a bit confusing

[ Parent | on HN ]

Re:Apple Silicon and macOS VMs: 11–16× Faster LLM Inference ... by frabonacci on Tuesday August 11, 2026 @11:27AM (#49259853)

yeah fair point. it's always tricky to get the whole idea across within HN's title limit. tldr: we ran the same workload in the same Lume macOS VM on the same Apple Silicon host, first with stock Metal capability reporting and then with our process-scoped dynamic library. The 11.08x figure is prompt processing, while 16.36x is token generation. the mechanism technically extends to graphics workloads too but these figures are specifically from llama.cpp

[ Parent | on HN ]