Welcome to Part 3

Transcript

0:11 - We’ve mainly talked about hardware in the first two weeks. Now we’re going to focus on software. We’ll take a couple of simple examples and consider how you could split up each calculation to take advantage of a shared or distributed memory parallel computer. We’ll cover some of the key issues of parallel computing this week at a conceptual level. Again, the analogy of different people working on whiteboards is very useful to illustrate the core concepts. One of the examples we’ll use which is a very simple way of simulating the way the traffic flows on a road is something I hope you’ll find interesting to illustrate how we can use computer simulation to make predictions about the real world.

0:48 - Now the way the traffic simulation is parallelised will allow us to look at what overheads are introduced by parallelisation and enable us to start to quantify when running on a parallel computer is worthwhile and when it isn’t.

In this video David will give a brief description of what awaits you in Part 3.

In the last section we talked about supercomputers mostly from a hardware perspective, in this part we will focus more on the software side i.e. how supercomputers are programmed.

The first activity is centred around a traffic simulation example, which hopefully will help you understand the concepts of parallel computing better.

The second activity will teach you about different programming models and how they relate to machine architectures.

Finally, in the last activity we will talk about performance - you will learn how to determine if a parallel program uses computing resources efficiently.