I'm attempting to do a simulation comparing resutls from a random effects model with a random intercept to some other regression models, but the mixed command slows down my simulation too much to be useful. In my experience running mixed effects commands in SAS is relatively quick, and the LEMR command in R is relatively quick. Why is mixed so slow, and can it be sped up?

For context, I'm using stata 14. I have a very simple model that I run with "mixed y T || cluster_id:"

Thanks.