 So as you can probably imagine, a performance number on its own isn't really worth a whole lot. If you've ever seen the result of a synthetic benchmark, they give you a number. What does that number mean? Is it good? Is it bad? Well, we need to be able to compare it to another computer before we have some idea about whether or not this is actually a good result. Is our computer fast or not? Well, really we want to be able to compare it to another machine to see, is this better or works? Would I rather have machine A or machine B? So we can have a relative performance equation which puts both results into a ratio and gives us a way to compare the two. So we can put that in terms of either the performance numbers or the execution times. Either one will be equivalent because as we saw before, performance is inversely related to the execution time. So we can use whichever one makes more sense for a given task. If we know benchmark scores, we'd probably be interested in using the performance numbers. If we have execution times, we'd want to use the execution times. But this will give us a way to compare two machines and see which one is doing better than the other. And we can also start looking at some even more complex workloads, see how things work, what sorts of things are going to affect different things. And we can start getting into some really interesting details here.