How can LLM inference be done better on non Linux? I can run models on my 96GB RAM MacBook Pro incredibly well. As soon as someone tells me how this can be do

How can LLM inference be done better on non Linux? I can run models on my 96GB RAM MacBook Pro incredibly well. As soon as someone tells me how this can be done in Linux, I'm ready to switch.