Reddit user builds $3360 Threadripper box with 256GB DDR4 to run large LLMs locally
sitytitan · reddit · 2026-10-26
A Reddit user shared a $3360 local inference build: 2x 16GB 4060 Ti plus 1x 12GB 4070 (44GB VRAM total) on a TRX40 Threadripper 3960X with 256GB DDR4 quad-channel 3200MHz (100GB/s bandwidth, the main bottleneck). After being pleased with Qwen 27B at 256K context, they're asking what token/s to expect running larger models like DeepSeek and GLM.
More from Infra
- $1.5M local inference cluster running 240+ kWh/day is the new 'lottery winner' flex — TheZachMueller · 2026-10-26
- Gartner: 80% of neoclouds to cease independent ops by 2030, $4T at write-down risk — rohanpaul_ai · 2026-10-26
- Engineer builds telecom control-plane anomaly detector with local-model RCA on tens of GB/day — l4rz · 2026-10-26
- SoftBank plans $50B AI and data center outlay in Malaysia — VectorWireAI · 2026-10-26
- Chunking latency cut from 37.1s to 585ms in another weekly perf improvement — DanielLockyer · 2026-10-26
- Smart TVs spy on you: Samsung TV found with always-on mic and ACR data shared with ~400 partners — FinanceYF5 · 2026-10-26