Poor man’s guide to servicing a used RTX 3090 for local LLM inference
Wrote up the whole process with disassembly photos and HWiNFO before/after data. Hope it saves someone some headaches. https://preview.redd.it/glwl9427hgyg1.png?width=1678&format=png&auto=webp&s=8e0755ccd9be456e2e65a96613edb0bce573991…