5.4 Motherboard and CPU
Activity 13 (exploratory)
Watch the third section of the ‘Inside the box’ video (entitled ‘Deep inside – the motherboard and CPU’), which continues the exploration of a desktop PC by giving you an overview of the motherboard and the CPU. In this closing section of the video you will see how the transistor bridges the computer science and engineering views.
Download this video clip.Video player: Inside the box: section 3
Transcript: Inside the box: section 3
[Graphic title: Deep inside – the motherboard and CPU]
[Music – duration 00:10]
Spencer Kelly
There’s not much left to examine now. We’re left with a large printed circuit board, lots of chips and sockets all over the place. Now, the CPU and the memory are both here – before we get to that, Darien, what else is on this board?
Darien Graham-Smith
Well on this board, the motherboard, there are lots of chips that control the various connectors, such as the keyboard and mouse connectors, or here we have the drive connectors for DVD drives or hard disks.
Spencer
And then there’s wiring and there’s sensors, I know, and there’s other stuff as well. There’s a battery here, now what’s that about?
Darien
Yes, on the motherboard there’s a clock and the computer can also remember a couple of basic settings, even when it’s disconnected from the mains.
Spencer
So the battery keeps that information alive?
Darien
Yes.
Spencer
And Parisa, this is still just engineering for you?
Parisa
It is indeed. For me it doesn’t hold much interest. For me it’s all hardware engineering.
Spencer
Well hang on, ’cos things are about to get more interesting. Darien, unleash the memory, where’s that?
Darien
Yes, these here are the memory modules and each one has a big row of memory chips on it, all the same. Thank you for that. They connect via a standard interface as you can see.
Spencer
You can have 1, 2, 3 or 4 on this one, can’t you?
Darien
Yes.
Spencer
Yeah. OK, thanks for the memory. Parisa, this is going to fit into your model, isn’t it?
Parisa
Indeed, so it can go to the memory box where it stores the symbolic data and instruction which will be used by the central processing unit.
Spencer
Talking of which, that is all that’s left, so if I just flip this up here we get to the heart – indeed, the brain – of the computer. This is the central processing unit. Darien, this is a really complicated thing, isn’t it?
Darien
That’s right. The CPU, central processing unit, is what makes this a computer. The CPU is able to actually perform operations, to work out calculations and to send signals back and forth between itself and the various components of the PC. Those little pins there that go into the socket, the electrical signals travel on those pins.
Spencer
And there are lots more than any other device that I’ve seen. There’s hundreds of them.
Darien
There are hundreds indeed, and newer processors have even more. And inside there, there are literally millions of tiny transistors. So it’s an extremely complicated piece of technology.
Spencer
This is what it’s all about, isn’t it?
Parisa
It is indeed an excellent piece of engineering which can go to the model, in the model, the processor box. But in my view it’s still a physical component which is capable of carrying out really complex computation and processing algorithms.
But it’s not the only way of doing the computation and processing. In old days researchers built mechanical computers by using gears, levers and cogs. And in recent days researchers are building computers by using molecular, biological, hydraulic and quantum technologies. So in my view this old central processing unit can be modelled the same way as I modelled before.
But the winning piece of all this central processing unit and in general-purpose computers are switches called transistors. The processor which carries out symbolic manipulation on the symbolic data relies heavily on the transistor logic-based circuits.
Spencer
I see. OK, well the CPU is – as we’ve established – an incredibly complex thing, so let’s hear more about processor development from someone involved in making them.
Steve Cutler
Intel’s primary business is really the design, development and manufacture of microprocessors, the central processing units that go into typical computer applications – desktop PCs, laptop PCs, servers, those kind of devices.
These days people want to do other things. They want to communicate, they want to collaborate, they want to share information – videos, pictures, things like that, perhaps. And so the use you make of a computing device is quite different from considering how you’d actually design one of those devices. But of course the two are closely linked in many ways.
So the internal instruction sets are focused onto the end application primarily, which we identify by looking at the market trends to try and see, you know, how we can take advantage of future trends and build them in.
The types of processors you’d get in traditional computing devices will look and feel fairly similar. Devices for laptops will probably be the smallest package we can put a device into. The desktop processors will be somewhere in between. Server processors will be bigger still, probably, because typically they are slightly higher power, means they generate more heat and the bigger package means it’s easier to remove the heat from the device and do it. So, you know, there’s many different physical flavours of these devices to allow you to build exactly the device you want for its end purpose.
So it does take quite some time to go from the theoretical to the practical device itself, and there’s a lot of stress in between those two as you work out what is possible and what is not.
If we look back in time, the first microprocessor that Intel developed – the first microprocessor anybody developed that was commercially available, the 4004: 2300 transistors. If we look at today’s processors, the most common devices on the market for desktop and laptop consumption were up to 560 million transistors with the Intel Core family of processors, and for server processors even more so. So the Itanium family of processors, up to two billion transistors these days. So to try and do an analogy to help you understand this, you know, let’s suppose we say transistors are homes for people. So 2300 homes, a small town perhaps. Two billion homes, we’re looking at pretty much every home in the world, I would say. So we’ve gone from interconnecting a small town to interconnecting the entire world in terms of interconnecting the transistors on our processors.
So trying to design a device of that complexity really does require you to break it down into smaller subsections and into modular pieces. And then just ensure that the individual modules interface to each other correctly and communicate correctly so you end up with a fully functional final design.
But at the lowest level of this design it still comes down to the transistors on the silicon. And that’s, you know, where Intel puts in a huge amount of research and development effort. So yeah, ultimately from our point of view everything comes down to how many transistors can we put on a piece of silicon.
Spencer
Now, I noticed that both of you mentioned the transistor when you were talking about the CPU. We’ve also heard that the transistor level is the lowest level of processing, so could this possibly be where your two different views coincide?
Parisa
Yes, it probably is. I can still model or represent an on and off state of the transistor using binary. I can model the binary processing using the Boolean algebra, so I can develop the representation of the data and instruction which can be embedded in my models.
Darien
Yes, we can build devices which realise your models. We can construct circuits which directly reflect your symbolic representations.
Spencer
See, friends really. And without the underlying model of computation, the hardware engineer will have nothing to build – and without the hardware, the computer scientist just has theoretical tools for computation. But because these two distinct views can come together to turn theory into hardware, we have the rich variety of computing systems that have found their place in so many aspects of our daily lives.
Inside the box: section 3
Interactive feature not available in single page view (see it in standard view).