Home Lab: 8x RTX 3090 Rig Runs 10 Parallel Agents at 370 tok/s
dazzou5ouh · reddit · 2026-11-15
A Redditor shows off a stable 8x RTX 3090 home inference rig (GPUs bought at £600-700 before prices spiked). Key details: Romed8-2T motherboard with full PCIe 4.0 x16 (two cards share a slot at x8), dual PSUs, 270W power limits, modified NVIDIA drivers enabling P2P at a measured 22GB/s. Running quantized Qwen on vLLM + opencode: 60 tok/s single agent, 370 tok/s aggregate with 10 parallel agents. Idle draw is 300W, but memory bandwidth makes it a beast for local inference and small-model finetuning.
More from Infra
- Alchemy adds Rivet, Effect Cluster and 8 sandbox platforms with swap-one-layer portability — samgoodwin89 · 2026-11-15
- Claim: Gemini 4 Trained Entirely on Google's Own TPUs, No Nvidia Involved — gaganghotra_ · 2026-11-15
- Chamath explains why CoreWeave credit default swap spreads are jumping — rohanpaul_ai · 2026-11-15
- Solar-powered local AI: old MacBook + Meshtastic mesh keeps inference running off-grid — natesiggard · 2026-11-15
- Do developers want a simpler middle ground between AWS AgentCore and Cloudflare Agents? — line9779 · 2026-11-15
- Thousands protest AI data centres in Edinburgh amid European backlash — emax · 2026-11-15