In this chapter, you explored the performance of different configurations when you perform compute-intensive linear algebra operations.
Benchmarking is a serious business, and you at least have the basic skills now to run benchmarks. The material you have studied in this chapter is nowhere near complete, but it gave you an idea where to start, and you can definitely improve on many things.
One thing you can look at is how performance metrics behave when you increase the size of vectors and matrices gradually. Ideally, you'll need more powerful hardware, but t2.micro instances are free in most cases or very cheap to provision.
As you will need to handle more compute-intensive workloads, it's important to understand what your options are, and which one will give you the best performance. You can run these kinds of simple experiment to at least have an idea about...