Comment Comment by myrmidon Loosely related question:What prevents manufacturers from taking some existing mid/toprange consumer GPU design, and just slapping like 256GB VRAM onto it? (enabling consumers to run big-LLM inference locally).Would that be useless for some reason? What am I missing? parent myrmidon 1 year ago
What prevents manufacturers from taking some existing mid/toprange consumer GPU design, and just slapping like 256GB VRAM onto it? (enabling consumers to run big-LLM inference locally).
Would that be useless for some reason? What am I missing?