Could we build a Computer Processor from Mechanical Parts?

Let's assume for a moment that the transistor had never been invented and humankind were still using electromechanical relays, like in the early 1900's. We can still apply boolean logic to them, just as we can apply it to transistors, but what about the engineering side of things? Let's take a look:

Mechanical parts are larger, consume more electricity and wear more quickly than semiconductors.

Modern processors consist of billions of transistors. Let’s assume we had a 1 billion transistor chip and we wanted to build it with relays. If each transistor were to be replaced with one relay, then we would require 1 billion relays. Let’s further assume that 1 relay would require 1cm^3 (the size of a sugar cube) of space and that we need another 1cm^3 of space around each relay for wiring, cooling, etc. So we'd need 2cm^3 of space per relay. That would be 2 billion cm^3. for all our 1 billion relays. That’s 2,000,000,000 cm^3 = 2,000 m^3, a cube with a side length of 12.6m, equivalent to a 4 storey building. Ok, that's BIG.

What about powering these 1 billion relays?

If each relay required 50mA of current at 5V, then we’d need 50mA*1,000,000,000=50,000,000A.

50,000,000A*5V=250,000,000W, which is 250 Mega Watt. A smaller coal fired power plant produces 500 megawatt of electricity and burns 1.4 million tons of coal each year. We’d need half of this.

In summary: If we could build such a relays computer, it would be the size of a 4 storey house, require half a coal-fired power plant and consume 700,000 tons of coal each year. This would be a tad too big to carry around with us. Not to mention the heat that the 700,000 tons of coal generate.

I'm glad the transistor got invented :-)

Tutorial 11: The Role of Timing

Timing plays an important role in the proper function of a computer’s internal and external communication. When we press the Enter button on the Program Counter (which triggers a clock signal), a sequence of events takes place that we have documented in the following little video. This, and more, happens a billion times in our computers and smartphones every single second. So let's see what is going on inside the machine when we compute 5+4-2.

The Art of Algorithms

The memory is a bit blurry, but I think I started with computer science at the age of 13 or 14 when some of my friends got their first computers. In the afternoons, we would meet at their places to play games, write code and explore these cool machines. I purchased my first programmable calculator when I was 15 years old (It had 420 bytes of memory) and programmed my first neural network when I was 19 (on an Atari 1040). I have since seen many different aspects of computer science. Only recently, when I designed the Blueberry4 computer, did I fully appreciate that algorithms are, most fundamentally, the art of translating really complex problems into simple additions of 1's and 0's.

This is true whether we program an autonomous vehicle control system, an artificial intelligence, a chess program, ... anything. We normally don't see this because modern computers are black boxes and because we re-use lower-level algorithms which have been written by others. We call these drivers, libraries and frameworks. Often we don't even realise that we stand on the shoulders of giants. Ultimately, the computer resolves all algorithms, layer by layer, into simple additions of 1's and 0's (helped by storing and retrieving data).

This is the foundation of computer science. Writing widgets, apps, games, animations, databases, etc, often seem like disconnected activities to us, the teachers, and especially to the students. The underlying algorithms sometimes appear useless, without purpose and artificial. But if we can convey the bigger picture that is at the heart of all our computer science efforts, namely the great reductionism of complex real-world stuff into simple additions of 1's and 0's, then we have found a common theme that spans all computer science education.

If we are prepared, then this realisation can influence our very understanding of the world and the wonders of mathematics, biology, physics, chemistry and philosophy. It can encourage students to contemplate this very universe we live in and think about the great questions of life. And it all starts with 1+1.