Is this some kind of trick question? Your core needs to be fully programmable so you can do things like kernel fusion. The simplest form is to load quantized w

Is this some kind of trick question? Your core needs to be fully programmable so you can do things like kernel fusion. The simplest form is to load quantized weights and dequantize them to bfloat16 as you go. Llama.cpp and it's gguf files support various types of quantization and most of them require programmability to efficiently support them.