Running Llama Locally: ~2-8 tokens/s on CPU, Boosted by GPU
thisdudelikesAI · x · 2026-07-07
Running quantized LLMs purely on a CPU yields about 2-8 tokens/second—functional but relatively slow. Performance sees a significant boost when hardware acceleration is added via GPUs (NVIDIA, AMD, or Apple Silicon's unified memory architecture).
More from Infra
- Revenue per GPU can rise as token prices fall, argues Gavin Baker — GavinSBaker · 2026-10-27
- Anthropic S-1: only $19B total compute spend since inception, overbuild warning — bookwormengr · 2026-10-27
- CoreWeave, You.com and NVIDIA lift BrowseComp 22.9% in 3-week RL post-training sprint — RichardSocher · 2026-10-27
- Grid queues make power architecture an AI-factory decision; Exowatt and Redwood offer fixes — ingliguori · 2026-10-27
- Two CVSS 9.9 Wasmtime flaws allow sandbox escape, one ARM64-only since 2020 — basedjensen · 2026-10-27
- AI Infra Meme: While Countries Need TW-Scale Power for Datacenters, People Argue About Lightbulbs — mgostIH · 2026-10-27