Columns

This month, I'll be introducing you to a practical application of Amdahl's law. No, really! Unlike Gordon Moore's law, which is nothing more than a rule of thumb, this is actually a proper law. Wikipedia can furnish you with the mathematical definition, but, briefly, Amdahl's law is all about the most efficient way of optimising a program for best performance. Say you have an application comprising two distinct sections of code that need to execute one after the other. Before optimisation, section A takes 90% of the total application runtime to complete, and section B the remaining 10%. Amdahl's law basically says you should put your efforts into optimising section A, because even if you manage to optimise the hell out of section B and make it run 90% faster, your program as a whole will run only 9% faster overall. If, on the other hand, you focus on section A, even a small speed increase will have a big positive impact.

This sprang to mind after a trip to a restaurant on Camden High Street. There were two little tables outside, and two outdoor electric patio heaters. On one of them I could see a sticker showing its power rating: 1,300W. That's 2.6kW of power for the pair, all devoted to the pointless pursuit of trying to make the universe warmer for the couple of people standing outside having a crafty fag. On the train home, I picked up a newspaper and started reading a piece about the evils of mobile phone chargers: apparently, if you leave yours plugged in when it isn't actually charging the phone, you're an evil person and will destroy the planet.

Here's a simple test. Plug your mobile phone

ADVERTISEMENT

charger into the wall and switch on the mains. Wait an hour or so and then put your hand on it. Is it warm? No, because it isn't wasting a significant amount of energy. Now wave your hand around the back of your PC. Feel that warm air being pumped out? That is wasted energy. Simple, no?

PC Pro is all about testing and hard results, soI sallied forth to the Labs and grabbed one of our power-consumption meters. I plugged a mobile phone charger (a Samsung TAD137UBE to be precise) into a four-way mains extender attached to the meter. The meter read zero Watts. Then I plugged in a Nokia one as well (an ACP-8X). The meter continued to read zero Watts. Then I plugged in two more: another Nokia and a Sony Ericsson. Guess what? Zero Watts.

That doesn't mean a mobile phone charger left plugged in consumes zero power, but it does mean it consumes an amount of power below the measurable threshold of our meter, which is less than a Watt. And since I had four of them plugged in, we can safely say that on average each consumed no more than 250mW (a quarter of a Watt). This means a charger left plugged in with no phone attached uses - at the very most - one four-hundredth of the power of the 100W lightbulb you probably have in your hallway. Put another way, leaving the hall light on for an hour uses at least as much energy as leaving your mobile charger plugged in for almost four working weeks.

Now consider that patio-heater use is predicted to double over the next year, and that those two electric heaters at the restaurant - which were achieving just a little more than bugger all - were chewing through well over 10,000MCs. An MC is a unit of power I've just made up, equivalent to the power consumed by one mobile charger.

Back to Amdahl's law. The Chillblast Fusion Trident PC on p38 consumes 400W of power even when it's idle: 1,600MCs. Even a frugal PC such as the Very PC Treeton (web ID: 145179) consumes 60W, or about 240MCs. Hopefully, you can see where I'm going with this. Switching off your mobile phone charger is almost as close to being a pointless energy optimisation as it's possible to get.