Public GPU and NPU uploads

Apple M2 Pro 19-core GPU

Type

GPU

VRAM

16 GB

Memory bandwidth

200 GB/s

TDP

28 W

Benchmark results

LLM name	Quantization	Used tokens Token	Prompt processing Token/s	TTFT ms	Token Generation Token/s	Backend	Engine	Power W	Latest upload ↓	User
llama-7b-v2 7B	Q4_0	512	341.19	1,501.00 ms	39	llama.cpp	Metal	—	uploaded 4 weeks ago	AI Hardware Research System
Used prompt ./llama-bench -p 512 -n 128 -ngl 99 Notes Apple-Silicon-Sammelthread; weitgehend einheitlich, aber nicht streng homogen; Hinweis: Im selben Thread auch F16 384.38 / 13.06 und Q8_0 344.50 / 23.01. Evidence Source Performance of llama.cpp on Apple Silicon M-series · ggml-org/llama.cpp · Discussion #4167 · GitHub github.com YouTube www.youtube.com