Slashdot Mirror
Gaming Computers Offer Huge, Untapped Energy Savings Potential
Required Snark writes: According to Phys.org, a study by Evan Mills at Berkeley Lab shows that "gamers can achieve energy savings of more than 75 percent by changing some settings and swapping out some components, while also improving reliability and performance" because "your average gaming computer is like three refrigerators." Gaming computers represent only 2.5 percent of the global installed personal computer (PC) base but account for 20 percent of the energy use. Mills estimated that gaming computers consumed 75 TWh of electricity globally in 2012, or $10 billion, and projects that will double by 2020 given current sales rates and without efficiency improvements. Potential estimated savings of $18 billion per year globally by 2020, or 120 terawatt hours (TWh) are possible. Mills started the site GreeningtheBeast.org.
You can read the full paper as a PDF.
So I should swap out my video card to save a little power, drop a few frames and die a virtual death? I think they have their priorities backwards.
I will not mourn that which I never had to lose. - Unknown
Puritanism: The haunting fear that someone, somewhere, may be happy.
H. L. Mencken
Prove anything by multiplying Huge Number times Tiny Number
Wouldn't it be easier to TURN OFF the gaming computer when you're not using it?
If you read or even browse the paper, all he really says is if you use newer components, they are more energy efficient. Which is like well, pretty much everything else on the damn planet.
Power saving settings on computers are super annoying. I work for a company whose software is ran in the cloud and the dang power saving settings on network cards make our program freeze in about 2 minutes of inactivity to save power annoying as heck to walk non-technical people through changing their windows power settings.
Hopefully this guy doesn't find out how much power is being used for bitcoin mining.
file:
How is it possible to take 20 pages to say so little that is actually meaningful? It basically boils down to "newer and/or lower performance components draw less power". No shit, Sherlock.
It's also rather misinformed when it comes to the availability of power-consumption information for gaming PC components. My current PC is a self-built gaming PC and I can assure you that when I was putting it together, power consumption information was absolutely something I looked at, because it affected my choice of PSU. And if you go to the manufacturers' websites, power consumption information is usually available upfront. If it's not, or if you want to know how it varies depending on loads, then there are any number of testing, benchmarking and review sites just a google search away.
There is probably an interesting article that could be written about minimising power consumption in a gaming PC, but it's not this one. In reality, power consumption is one aspect of a sensitive series of trade-offs. On graphics cards, for instance, you get get the same brute-force performance from AMD cards as you can from Nvidia cards at (usually) a much lower price - but the trade-off comes in heat and power consumption. So you can base your decision on a balance between how much you care about the up-front purchase costs of the card, vs ongoing power costs, potentially the cost of a new PSU and the noise/discomfort factor of having something that burns with the heat of a billion fiery suns in your PC. Most people building gaming PCs are not blind to this stuff.
The article reads like a lightweight piece of political advocacy for more regulation, trying to solve a problem which increasingly doesn't exist (the general trend over time is towards more power-efficient components and electricity prices act as a further restraint). So the author can, to be blunt, fuck right off.
No games require an i7. Not even wolfenstein, which runs fine on both my 4690K and 4790K pcs.
Ummmm....the 4790K IS a Core I7. So obviously you wouldn't have any problems running it.....
You're messin' with my Zen Thing, man.....
Although it's not perfect, money is a decent proxy for environmental harm.* So, if a $100 upgrade will save you $200 in electricity over its lifetime, then the upgrade will probably do more environmental good than harm. However, if a $500 upgrade will save you $100 in electricity, then you're probably doing more harm than good.
* At least for normal consumer goods, the price _roughly_ reflects the amount of energy and resources to manufacture the good, which roughly reflects the environmental harm. It's by no means a perfect metric, but it's a start. Some goods clearly do not fit this model. For example, a painting costs almost no resources to produce but can sell for a high price. Some computer parts are similar. For example, sometimes identical graphics cards are deliberately crippled (lower clock speed, parts of the processor disabled, etc.) just to create different price points. Both cards have the same environmental cost to produce but can have very different sale prices. However, that means the environmental cost is best represented by the cost of the _cheapest_ version. So maybe the aforementioned $500 upgrade really costs $50 to produce and thus has a positive environmental impact.
(Totally off topic: I wonder about the environmental impact of moving to cities. Say you move to a city, sell your car, etc. but your income remains constant; you instead spend money on a new TV, more beef for dinner, etc. Then it's not obvious to me that you're having a significant, positive environmental impact.)
His numbers are way off. First a gaming computer is not "three refrigerators." A fridge/freezer combo uses like 400-800 watts when spun up depending on size and if it is frostless or not. Your typical reasonably high end gaming computer (high end quad core processor, single high end GPU) uses in the 300-400 watt range when fully spun up. There are, of course, higher end systems but they are not common as they cost a lot, for not a ton of gain.
Well the idea that there are tons of components or settings that'll just tank energy use is stupid. In terms of settings, ya those are default. By default a system will put its processor and GPU in to an idle state when not heavily loaded, and indeed most systems draw 90 watts or less when idle. In terms of componentry, there really isn't a ton of room for gain.
Like with PSUs. Any reasonable quality PSU that you might see in a gamer build is at least 80% efficient, and usually more like 85%. Go all the way to the high end, which many gamers already do, and you are only pushing 90-92% max. A gain, sure, but not much. If a system draws 300 watts DC going from an 85% (bronze) PSU to a 92% (platinum) PSU is the difference between 350 and 326 watts at the wall.
Then there's things like GPUs and CPUs. Well guess what? A give one is as efficient as it can be at a given performance level. There aren't the better and worse ones. You can't buy the efficient model GTX 980 and the inefficient model. They are the same. You can swap one kind of component for another and maybe gain efficiency. Like you can swap an AMD 390X for an nVidia GTX 980Ti and that'll use less power, but what if you want the AMD card?
Also there's the issue that usually the new ones are more efficient than older ones. Fair enough but in addition to the cost of upgrading that ignores the energy cost of producing the cards. Suggesting that everyone buy the newest shit all the time is not realistic, or energy efficient (a lot of our energy use goes in to making things).
This guy just doesn't know anything about computers. He's convinced that there's these vast optimizations that could happen, if only people wanted it. Not really the case.
Power consumption is really only a problem when my gaming computer is awake and that's not that often. The power consumption when sleeping is minimal and not worth worrying about. Even if I were to run my gaming PC 24x7 and never have it sleep, it only costs about $5 / month and I don't really care about that. If we were talking about a major household appliance then sure, I care about its power consumption but not when it comes to my gaming computer.
It's the classical "iron triangle" (aka constraint triangle): you have three sides to a triangle; performance, power consumption, and cost. Pick any two that you want to be favorable, and the more favorable you make them, the more unfavorable the third will be.
Nvidia and AMD discrete cards for desktops are designed to tug on the performance and cost sides as much as possible, leaving only a passing thought for power savings. Granted, it's reasonably efficient when in standby and not that bad at idle, but it's horrendous under load.
If you want a GPU that's less horrendous under load, without compromising on performance, expect to pay a LOT more for it. And in most cases, because of market forces, you're also going to take at least a slight performance hit.
For instance, the GTX 980M is way more power efficient than the GTX 980 desktop card, because of the design constraints of laptops. It's effectively power efficient by design because of the form factor. But the performance is notably worse, and buying a laptop that has one is way more expensive than buying a desktop with "good enough" components (some kind of recent i7 and 8GB or more memory) and a GTX 980 or even 980 Ti.
We're not going to see chips that maintain the desired performance level (the one constraint that most people are unwilling to compromise on if they're owning a "real" gaming PC) while saving on energy, unless the cost goes so high that only the most elite can afford it.
We see similar problems in the car market, too. The Prius and Prius C are relatively inexpensive (the C is very inexpensive), great fuel efficiency - though not as good as an EV or PHEV - but the performance is terrible. The Tesla Model S achieves amazing performance, range, and efficiency in a pure EV package, but no "commoner" can afford it, only the upper crust. Then there are loads and loads of cars that are not particularly efficient, very cheap, and have serviceable performance.
Who's paying for the more expensive chips that give us the same performance we're already getting but with less energy consumption? The gamer? Why would they do that?
P.S. - Before you accuse me of not thinking of "the greater good", I *drive* a Prius C. I bet the same scientists who wrote this paper drive conventional SUVs.
Although this isn't high-end gaming, I noticed a couple of interesting power measurements at work:
My 10-year old Dell desktop, running Windows XP and no anti-virus (AV) boots ~25% faster and draws ~30% less power overall (idle & taxed with performance software) than the new guy's new Dell laptop that runs Windows 7 & Norton AV. We can both run all the same software effectively (albeit all my versions are several years older), except that I have a different brand of PCB design software than he; I'm guessing if I ran his PCB program my computer might drag.
Also, my ball mouse draws 1/5 the power that his optical mouse does! That was a surprise, too. I wonder whether a ball mouse is a better idea for laptop users when it comes to extending battery life.
My single CRT drew 1/3 more power than an LCD. My new, micromanaging, pointy-haired boss forced me to toss it(?), and replace it with two LCD widescreens. So now I draw 50% MORE than my single CRT did. My productivity is unchanged. Sigh...