Don't Sell Your Old GPU: Run Local LLMs Across Both Cards
Techniboy · reddit · 2026-11-09
A Redditor reminds GPU upgraders not to let old cards sit idle: install both GPUs in one rig and split LLM inference across them with llama.cpp. The author combines an RTX 5080 and 3080, pooling their VRAM to run larger local models.
More from Infra
- Data center logic flips: Argentina's Patagonia courts compute with stranded energy — matt_slotnick · 2026-11-09
- Cathie Wood pushes back on data center fears: electricity prices are real, the rest is overblown — CathieDWood · 2026-11-09
- Run free local LLMs by RAM: Qwen 4B on 8GB, 35B MoE on 32GB, fully offline — alex_verem · 2026-11-09
- Bipartisan Senate bill could fast-track US AI data centers, vote as early as next week — Polymarket · 2026-11-09
- Larry Page predicts 99.999% of compute will end up in orbit, citing US data center regulation — elonmusk · 2026-11-09
- xAI's turbines had no exhaust cleanup, but Anthropic's gas engines likely do NOx control — austinc3301 · 2026-11-09