Your phone is supposed to be a phone, not a hand-warmer. So why does it, at times, feel like it could barbeque your palm?
The components in your phone or tablet – the screen (which it’s easy to forget is actually powered by a tiny light bulb), the GPS, the camera, the processor – draw electricity from the battery. When they do so, they aren’t completely efficient: They always waste some energy, which is released in the form of heat. By measuring how hot your phone gets, you could figure out how inefficient its electronics are.
For now, let’s ignore all the other parts of your phone that require power and focus on the tiny computer inside. Every time you search for a contact, shuffle your tiles in Words With Friends, or ask Siri a question, your phone or tablet’s internal computer crunches data. At a fundamental level, it’s pushing around bits of information – 0s and 1s – which takes energy.
Just how much energy does it take to crunch data?
Let’s take one common action: Instagram. Let’s apply some filters to a photo, and do a mini-experiment to see how much energy the computer in a mobile device expends to do this. I’m going to use my Nexus 7 tablet, but you could just as easily do this with a phone or laptop. First, I took a photo:
Jordan Wirfs-Brock
My colleagues at Rocky Mountain PBS I-News and Inside Energy, Anna Boiko-Weyrauch and Dan Boyce, helped me with this experiment by posing for a photo. They were trying to figure out the best way to fight afternoon doldrums.
Next, I opened it in Instagram, sharpened it across the diagonal, brightened the colors and applied the “Rise” filter. To do that, my tablet did a whole lot of complex computing. The photo itself is composed of 160 million 0s and 1s. So applying filters takes, at minimum, billions of calculations. It’s enough computing that you can actually see it – the image processing took about two seconds:
I timed how long it took my tablet to apply some filters in Instagram.
How much energy did the tablet’s internal computer use in those two seconds? To figure that out, we need to know how much power the tablet’s processor draws when it is working full-bore. This took a bit of digging (if you enjoy rabbit holes, you can see the reasoning here), but it’s a nice round one watt. So, assuming that my tablet’s processor was working at capacity when it was filtering the photo, two seconds at one watt gives us two watt-seconds, or 0.00000056 kilowatt hours. Compare that to the U.S. average monthly household electricity consumption, 903 kilowatt hours. You could filter 1.6 billion images in Instagram with that much electricity.
The final product, which cost me roughly a millionth of a kilowatt hour in terms of processing power:
Why does this matter?
The tiny computer in my tablet can perform 16 billion computational instructions a second using a mere one watt. Contrast that with the processor in the first Mac (known as a Macintosh back then) introduced in 1984, which could perform only 700,000 instructions a second using 60 watts. As our processors have gotten smaller and faster, they’ve also gotten more energy efficient: In terms of information processed per watt, my tablet is on the order of a million times more efficient than the 1984 Mac.
But here’s the thing: Those giant leaps in computer energy efficiency have slowed to incremental improvements in recent years. There’s simply a limit to how many 0s and 1s we can process per watt, and we’re approaching that limit. You can feel that limit every time your phone gets hot. When your phone is “working hard,” its internal computer is running at full capacity, drawing somewhere around one watt but wasting part of it, and heating up. Which, in turn, means it has to slow down to avoid damaging itself.
This problem is way, way bigger than the annoyance you feel when your battery drains too quickly or you burn the skin on your hand. While the kilowatt hours that my tablet’s internal computer uses to filter a photo in Instagram is miniscule, when we multiply that by all the computers in all the devices out there in the world and the massive amounts of data they are crunching, the energy cost is huge. As the amount of data we’re crunching is exploding, we’re bumping up against the energy use of electronics as a limiting factor that could contribute to the end of Moore’s Law and the pace of digital innovation as we know it.
Stay tuned. Inside Energy will be exploring the mind-blowing relationship between energy and digital information throughout 2015.
What’s your energy question? Send it to us at Ask@insideenergy.org and we’ll explore the answer.
