Comment by acidburnNSA

How many tokens/second you getting? I have the same CPU but a 5070 and only 64 GB of ram. I just got llama.cpp built and am now hitting a whopping 5-6 tokens/s.