A user exploring options to enhance their local Large Language Model (LLM) performance is contemplating a CPU upgrade from an AMD Ryzen 3700x to an 8700g. The primary motivation behind this potential upgrade is to increase the size of LLMs they can load and run locally, while preserving their current AMD Radeon RX 6700XT GPU for gaming. Currently, the user is restricted to models around 9-10GB in Q4 format and aims to maintain their existing BackyardAI setup. This question was raised on the r/artificialintelligence subreddit, sparking a discussion about the advantages and disadvantages of such a CPU replacement for local LLM tasks. The original Reddit thread can be found here: [https://old.reddit.com/r/artificial/comments/1nssxqo/considering_upgrading_from_3700x_to_8700g/]
CPU Upgrade Under Consideration to Boost Local LLM Loading Capacity
