More computing power than Apollo

It is often said these days that any given piece of technology has more computing power than it took to go to the Moon. Let’s do a little quantitative analysis to see if this claim holds true, comparing the Apollo Guidance Computer (AGC) to some modern tech.

Apollo Guidance Computer interface (Source: Wikipedia)

As a measure of computing power, I’ll use transistor count which isn’t a perfect way of going about it (since you can accomplish some task with different numbers of transistors depending on your design) but since we’re only doing an order of magnitude analysis it should be sufficient. The AGC is a good starting point for comparison, since it was one of the first major deployments of integrated circuits (ICs) in a computer.

The Block II version of the AGC which was used in the actual manned flights contained some 2,800 ICs, each of which contained two NOR3 gates (that is, a three-input logic gate implementing the Not OR function) using resistor-transistor logic (an early type of digital circuit that was phased out due to high power consumption).

RTL NOR3 gate (Source: Wikipedia)

An RTL NOR3 gate, as seen in the above picture, requires three transistors. That means 6 transistors per IC or about 10,000 transistors total to make up the Apollo Guidance Computer.

Now let’s compare it to something from the present time, for example my graphics processing unit (GPU) which I mostly use these days for catching butterflies and shouting at things, which has a comparatively staggering 2.64 billion transistors. That’s a few hundred-thousand fold difference in power, in just one IC. Nevermind including memory or storage in the form of solid state memory when adding up the components from the entire computer, in which case I’d hazard a guess of about 2 trillion transistors total (mostly from a 240 GB SSD). Which means, taking a decent 2011 computer as a point of comparison, the technical specs of the AGC are exceeded by about a billion-fold. Even something that nowadays we would consider fairly simple like an iPod nano probably has at least tens of millions of transistors for logic (it does have an animated, colour touchscreen after all) and billions of transistors for storage.

It seems like the common refrain not only has a solid quantitative basis, but is probably more true than most people guess, and the difference will only get more spectacular from here on.

For the truly dedicated: Schematics

This entry was posted in Electronics, Technology. Bookmark the permalink.