Choose an Energy Efficient Computer - an article by Mike Chin in the Aug/Sept 2006 issue of Home Power magazine. (Written many months ago...)

MikeC - Fantastic! I've been reading Home Power longer than I've been visiting SPCR, and had always thought that someone should write an article about efficient computing. You were just the right person to do it. Lots of good information without being over the typical Home Power reader's head. Great work; I'm sure many readers will appreciate it like I did!

Glad you like it. Claire Anderson, the associate editor of Home Power, noticed my articles about PC component power efficiency, 80 Plus and Energy Star... and called on me, originally to do a "top-10 most efficient computers" type article -- which, of course, is just not doable or realistic. This is what resulted in the end.

Excellent stuff - well done on spreading the word!

The UK government has just finished a review of power use and predicted trends in the UK and concluded that the majority of the rise in power demand is due to increasing no.s (and complexity) of appliances, including PCs. Given that power generation is stretched in most developing countries and renewables are unlikely to take up the slack (for various economic and political reasons), this inevitably will result in more gas/coal/nuclear powered stations with their commensurate environmental issues. Wow, silent computing saves the world!

Excellent article, Mike. It has just the sort of advice that non-techies can understand and act on. I only hope that lots of people do, as the figures on global power consumption due to PC/electronics use are frightening!

Anyone else having problems downloading the PDF? It seems that www.homepower.com is misconfigured.

I think you're right. Several people mentioned this; I didn't notice because I routinely save them to my own PC before opening PDFs. It's so much quicker that way, and easier to navigate through the PDF. That worked fine for me. Later when I tried opening the file in my browser, I had the same problem that others were having. I did email the editor, but perhaps I should resend that email.

Do you have the PDF saved? If so, would you mind emailing it to me?

Thanks!

All you have to do is right click on the link and choose "save link as..." (saved to SPCR's server.)

Got it, thanks!

That looks like a really cool magazine. I just may subscribe. Now all I need to find is an outdoor survial magazine and my "interested, but haven't gotten around to it" perodical needs will be taken care of lol.

Job well done on the article MikeC!

I haven't dropped into Silent PC as often as I'd like lately (not in the middle of building a computer), but I recognized your name when I read the Home Power article!

Great work as always, though a bit of a retread for us silentpcreview folks!

Good article.

Hope this is the right place to discuss this - I still regularly use for webbowsing and email an IBM thinkpad 560x (233mhz mmx) with 96mb ram. Measured with a Killawatt power meter, when plugged into the mains with the screen on full brightness (admittedly not as bright as modern screens) it draws between 11 and 20 watts in total depending on what the processor is doing (it is speed adjustable and is set to auto). I also have a modern 40 gb hard drive in it.

I mean, of course word processing is still fine on it, and with web browsing most streaming video is watchable if not perfect, and I still find the major limitation is the wireless reception rather than the compurter power. The biggest drawback is the 800x600 resolution probably.

Well, considering the imac in the article is only 33w at low power idle, and it may stay at this (can it do anything in this?) for whatever the ibm laptop can handle, then you may say well the huge difference in performance is well worth the roughly 1/3 extra power, or even roughly double the power if the imac still needs just idle (46w).

But still the ibm does have less power consumption in absolute terms and would be interested to hear what the lowest power consumption others have managed to achieve in a computer that they still regulaarly use.

On my main desktop pc I use a duron 700 mhz (think it maybe overclocked ot 800 - can't quite rmemeber) with 512 mb and it just seems fast for everything until i come to play video (and yes I am not only comparing it to my ibm laptop but to other much more modern computers I've used). Here though I'm not sure of the total power consumption and guess the imac would beat it.

It would have been nice to slightly develop that historical perspective that you referred to at the beginning of the article with the 386 and 486s. I guess a brief statement in the article to the effect that if you have an old computer one of your options is to try and see if it is already power efficient and look into whether you really need a new one (sorry in advance if I missed this!) - i.e. you might watch video on a tv, play games on a console (not a 360 though!), and trim down your OS or "always running" applications on your existing PC (even for non-techies you could have them ask someone more technically literate about unnecessary programs always running on their pcs), might have been good too. Of course, if you have a "recent old" pc, i.e. pentium 4, then you are best getting rid of it, but if you go further back then of course, as the SPCR processor recommendations show, then it is not so obvious.

Actually, in relation to the above paragraph, and certainly in relation to the most modern batch of processors that are intelligent and scale up or down depending on what they are doing, stressing the importance of keeping your computer in a "clean state" in terms of constantly running software processes is an important factor in its own right (and of course in this way increased power consumption is a potential effect of viruses - although on the other hand Symantec AV and internet utilities etc is a real processor hog so what one suggests to people I don't really know).

Oh, even with modern processors HDTV looks like being a big resource hog and so a big power drain potentially.

I'm glad I was able to download your article. I tried from your blog - it downloads something, but I couldn't open it. I was able to open the one linked above in this forum.

Having just downloaded the article, I've not had time to read it yet, but I don have a question:

How quickly does this information change? How often does one have to re-evaluate the relative consumption of various classes of computer (laptop vs. desktop etc.)?

I'm starting a new website (tiny as of this writing, but growing) which has information to help consumers and organizations become more efficient in how they consume the world's resources. Energy is a big part of that.

I am researching computers just now, which is how I found your blog article.

Thanks
Doug

The broken link is because homepower.com removed the pdf from their downloadables; it's been changed to point to the copy on the SPCR server.

As to how often... that's not easy to answer. It depends how uptodate you want to be -- and how many different types of computers you want to keep on top of.

The article I wrote was already somewhat outdated by the time it was published, that's due to the slow process of print magazine publishing and the speed of CPU development, for the latter, especially over the past year or so.

In general, there is greater focus on energy efficiency in the industry, so you might expect continuing improvements. On the other hand, mobility is driven by the need to conserve battery power and maximize run time, needs that don't exist for desktops in general, so I'd expect a gap to remain between the two general classes -- until/unless the 2 classes merge fully. That process is happening, but I doubt it will be complete for some years yet, if ever. There may always be a need for big PCs that have higher capacity for components & connections.

The other thing is that the energy consumption of a PC during its operational life is only 20-25% of its total. Around 75% of the energy it uses occurs during the production stage. This is because the "energy density" of CPUs, IC chips and other electronics and IT gear (hard disk, memory, etc) is very high.

A typical desktop PC (with monitor) requires 240kg of fossil fuels to produce; its total weight might be 20kg, so each kg in a computer requires 12kg of fossil fuels. A typical car requires 2000kg; typical weight is around 1000kg, so each kg in a car requires only 2kg of fossil fuels.

Even if you decrease its electricity consumption by half, the total energy consumption represented by a computer would only drop by maybe 12~15%. Not insignificant, but a 50% reduction in average power consumption would be VERY difficult to achieve with tweaks and adjustments to an existing computer; you'd just have to go to a newer, more minimalist computer with higher efficiency parts to start with -- but then you're financing another computer whose manufacture requires 7-8 times the energy you'll save.

Hopefully, that addresses your question somewhat?

Hi Mike,

Thanks for the detailed reply. It does help.

There are other factors to consider - the computer's power consumption and energy requirements can't be considered in isolation.

For example, in a large office building, the energy required to cool the building becomes important. If there are dozens or hundreds of computers per floor, the difference in load on the air-conditioning system is significant.

Also, a watt saved by the consumer is not a watt saved at the power plant. I can't put my finger on the article I was reading a while back, but I seem to remember numbers like 30% efficiency from a pile of coal to a watt in the home (or office). Even at 50% efficiency, the power saved at the generating plant is double the energy saved by the consumer! I think that might be a conservative number, if I remember correctly.

The biggest issue is peak power (rush hour on the electron super-highway!). Marginal power is really expensive to produce. It requires building power plants that are not needed during average and off-peak consumption. So how does building the extra power generation plants needed to power inefficient computers compare to the energy and material cost of producing laptops?

Laptops may fail more often, but what is saved in the aggregate, when you compare 1,000,000 desktops vs. 1,000,000 laptops? They both have failure rates.

Has anyone tried to put a laptop motherboard into a desktop case (in any kind of sensible, potentially commercial way)?

My conclusion? Needs more analysis!

Doug

Apple (iMac / Mac Mini)
Shuttle (X100)
Puget Systems
Several other boutique system builders who use AOpen's MoDT parts as a platform.

Yes. AOpen even coined a name for it; MoTD or mobile on the desktop. However they tend to be a lot more expensive than budget desktop solutions and the energy savings usually aren't enough to offset the higher initial costs for the average consumer; however some SPCRers have gone down this route, for low noise.