Uploaded January 2026 | Updated September 2026, 1 minute ago

---

AKA "Can you run LLMs/do AI stuff on a €9k computer from early 2014?"

Answer: YES! That 384 bit bus on the Titan Black and X79 being a quad channel platform really helps with inference speed. Couldn't get dual GPU to work with KoboldCPP though.

Recorded on my Better Retro Rig: E5-1680 V2 @ 4.75 GHz|R4BE|32GB|Titan Black SLI|970EP|X-Fi Titanium|AX1600i|OBT BC1|7/10 dual boot|MoRA3 420|CW-5000 Chiller