Technical problem: A regression model runs slowly

[Shir Caller]

Hello,
I am doing multilevel logistic regression models And I run into a problem when I add two slopes.
This is a model for example:

xi: xtmelogit racist i.men i.academic_education || cntry_numeric: i.men i.academic_education , var mle or cov(un)

on my laptop (I have STATA 13 MP) it takes about 50 minutes.
on my teacher's computer (STATA 15 MP) it takes about 15-20 minutes.
just to clarify a few issues:
1. without the slopes or with 1 slope only it's work ok and quickly.
2. I still didn't include interaction and more variables to the model but I'll need to.

therefor i wanted to ask, do you know why the slopes make everything work so slow?
and to be more specific:
What technical specifications does the computer need to have so that the commands with two or three slopes work faster?
and what version of STATA we can buy that will solve this situation?

Thank you very much,
Shir

[Carlo Lazzaro]

Shir:
it is not unusual that models like the own you describe take many minutes (and even hours) to give back an output,
That said, and possibly off (the main) target, -xi:- is redundant if you have Stata 13 on your laptop.