Redditor builds dual AI workstation with 2x RTX 3090 and 4x Tesla P100, asks how to optimize
FearFactory2904 · reddit · 2026-09-14
Reddit user FearFactory2904 assembled a two-system local AI workstation on aging Xeon x99 boards and asked the community for optimization advice.
System 1 (running): Xeon E5-2630V4 with 128GB DDR4 and two RTX 3090s (no NVLink, PCIe 3.0 x16):
- One GPU runs Qwen 3.8 27B Q4 quant served via Open WebUI
- The other runs ComfyUI for image generation/editing, called through Open WebUI's API
- CPU handles speech-to-text and text-to-speech
- A separate Docker container runs SearXNG to add web search to Open WebUI
System 2 (being rebuilt): Same Xeon with 256GB DDR4 and four Tesla P100 16GB cards. Plans include a Q8 quant of Qwen 3.8 27B with larger context, or trying Qwen 3.8 Flash, plus experimenting with agentic workflows.
The goal is a local "Swiss army knife" of AI features; the author asks how others would reallocate the same hardware.
More from Infra
- Claymation body horror on 8GB: three wrong settings, not the GPU, made local renders slow — roro9- · 2026-10-30
- durable-object-migrator: zero-downtime exports for Cloudflare Durable Objects — samgoodwin89 · 2026-10-30
- An unnamed firm already runs a fairly scaled dedicated vehicle for compute lending, says analyst — matt_slotnick · 2026-10-30
- GPU break-even throughput vs open-weight token prices up 25-83% since May — bookwormengr · 2026-10-30
- Sam Altman says the chip supply chain's invisible hand is failing — and DRAM is why — SnoozeDoggyDog · 2026-10-30
- MoEShift runs a 125B MoE on a 24GB Mac: 7.5GB peak memory at ~10 tok/s — dave-tro · 2026-10-30