 Okay, so next we're going to be looking at relative performance on a workload. And as you can imagine, it's not too different from what we've already seen. We've already seen relative performance and we've seen performance on a workload. So really we're just going to be putting these two ideas together. This time though, we will be able to look at workloads with continuous terms. Problems that are just expected to run continuously for days, weeks, months, even years. We'll still be using the same relative performance equation that we used before. And we will still be focusing on calculating the execution time from both machines. So we saw how a discrete workload worked before, and we're going to do the same thing here. When we start looking at continuous workload, we're going to be focusing on proportion. We know that we'll spend part of our time on one task, part of our time on another task. And together that will take up the entire time that we've got. But these proportions are likely to change when we move from one machine to the other. So we're going to be looking at comparing two machines on the individual tasks, and how those come together to create an execution time that we can actually use.