(Performance Analysis)   In this chapter, we used %timeit to compare the average execution times of generating a list of 6,000,000 random die rolls vs. generating an array of 6,000,000 random die rolls. Though we saw approximately two orders of magnitude performance improvement with array, we generated the list and the array using two different random-number generators and different techniques for building each collection. If you use the same techniques we showed to generate a one-element list and a one-element array, creating the list is slightly faster. Repeat the %timeit operations for one-element collections. Then do it again for 10, 100, 1000, 10,000, 100,000, and 1,000,000 elements and compare the results on your system. The table below shows the results on our system, with measurements in nanoseconds (ns), microseconds (µs), milliseconds (ms) and seconds (s). This analysis shows why %timeit is convenient for quick performance studies. However, you also need to develop performance-analysis wisdom. Many factors can affect performance—the underlying hardware, the operating system, the interpreter or compiler you’re using, the other applications running on your computer at the same time, and many more. The way we thought about performance over the years is changing rapidly now with big data, data analytics and artificial intelligence. As we head into the AI portion of the book, you’ll place enormous performance demands on your system, so it’s always good to be thinking about performance issues.