But that is a really subjective approach to compare performance between operating systems by how the UI response feels, isn't it? Wouldn't it be better to measure the average cpu usage over a certain time frame? Afterall, with trading automation we don't expect to click the UI all day and night.