Slashdot Mirror

← Back to Stories (view on slashdot.org)

Standby TVs Waste Electricity, How About ACPI?

Posted by Cliff on from the do-computers-waste-more-less-or-about-the-same dept.

twitter asks: "There's power management and there's standby, do you know the difference? The BBC is running story on how much electricity is wasted by TV standby mode. Thanks to the very useful EnergyStar program, I'd be the one in seven who thought they were saving electricity, with the standby button. I've been very happy with APM and hibernation on laptops, and want to do something similar with the desktops I use. What's the state of APM / ACPI Wake-on-LAN for Linux these days?" Slashdot touched on this issue, earlier in the week, but that article was more on TVs, not on computer power saving technologies.

31 of 166 comments (clear)

  1. Convenience by kestasjk · · Score: 3, Interesting

    If it's more convenient we'll keep on wasting energy. The worst part is the standby circuits use practically no power compared to the transformers, which waste far more energy as heat than the standby circuitry uses. There should be a seperate battery power source powering the suspend-mode circuitry, which lets current into the transformer to provide the power needed for normal operation.. But of course this would cost extra, and consumers wouldn't pay extra for it even if it saved money on power bills in the long run.

    --
    // MD_Update(&m,buf,j);
    1. Re: Convenience by XanC · · Score: 4, Insightful

      Okay... Suppose it costs an extra $10 for the battery, smart circuitry to run it, design costs, etc etc. Suppose disabling the transformer for standby saves you 2W. Suppose it's on standby year-round. That's 8,760 hours, or 17.52 kWh. Say 8 cents per kWh, you're now saving $1.40 per year. It would take over seven years for you to make up the initial cost.

    2. Re: Convenience by Anonymous Coward · · Score: 2, Insightful

      And say there are 100 million machines with your reduced power consumption. How many power stations does that save, with all the polution not being belched into the air?

      Now lets say we look at other devices and repeat the exercise. Oh look, poisonous emissions are going down, respitory health proplems are down, medical bills are down, medical insurance (for those countries with retarted health care like ours) are down (nah, maybe not this one).

    3. Re: Convenience by XanC · · Score: 4, Insightful

      Meanwhile we're filling the landfills and oceans with dinky little transformer-saving batteries.

    4. Re: Convenience by David+Horn · · Score: 2, Interesting

      Ignoring that fact adding a battery would probably use more energy in the manufacturing that it could possibly save, the point of the parent author was energy saving in general, not saving you money.

      I don't know how many TVs there are in the US, and I also suspect that shutting down the transformers in a large set will save more than 2W. I'm going to guess at 5W saved, over 500 million TVs. That's 44kWh per TV per year saved.

      In every TV, that's 2x10^10kWh of energy saved across the whole of the US. Those are wild guesses, but they're probably in the ballpark. I don't want to come across as a troll, but it highlights the fact that you think solely about yourself rather than the bigger picture.

      I'm not a tree-hugger, but you only have to look out of the window to see all the energy being wasted (PCs / TVs left on, cars sitting in traffic jams etc. etc.). Like it or not, unless we go nuclear we're going to have a serious energy problem in 50-100 years time. I think we need to start changing attitudes now.

      --
      PocketGamer.org - For the gamer on the go!
    5. Re: Convenience by alienw · · Score: 4, Insightful

      The standby circuitry in most devices could probably run for days on the charge in a $1.50 capacitor.

      I'm an electrical engineer, and no, it can't. That's why there is a transformer. The real solution would be to get off your lazy ass and hit the power switch when you are done watching instead of turning the TV off with the remote. The other solution is to put in a very high-efficiency switching power supply, but those are very expensive.

    6. Re: Convenience by alienw · · Score: 2, Informative

      It's a fallacy to point out the total energy used by such TVs.

      It's not a fallacy. There isn't just one of those TVs, there are hundreds of millions of them. They all use energy. Legislative mandates for more efficient electronics would go a long way. Right now, efficiency is simply not a criterion the manufacturer even attempts to optimize when developing a power supply; cost is a much, much bigger factor. This obviously needs to change.

      Are you basing that on anything but intuition?

      Don't be childish. If you take money out of your bank account at an exponentially growing rate and never put any in, it will run out in only a few decades. Regardless of how much is in there initially. Considering that energy use is growing exponentially, 50 years would be the upper bound for a shortage. I'm willing to bet on 10-20 before energy prices go up a LOT. Not like going to $2.20 per gallon for gas instead of $1.20, more like $20.00 per gallon. If this is to be avoided, the exponential growth has to be stopped.

    7. Re: Convenience by icepick72 · · Score: 2, Funny

      So our planet can be seen from space as a million pinpoints of light. It will camouflage our earth against the stars from the eyes of onlookers and would-be invaders, allowing us enough time to build up our technology and arsenal against them. C'mon, doesn't anybody see the big picture anymore? The media grabs ahold of small environmental issue and look what happens ... Think people. THINK!

    8. Re: Convenience by alienw · · Score: 2, Interesting

      Actually, that's already how it's done, and that's exactly why a lot of power is wasted. A small, cheap transformer is usually a lot less efficient than a large one. A large transformer has thicker wire for the windings, so less resistive losses. It also has a higher-quality core, so there are fewer core losses. Of course, this is largely irrelevant, since the last time large transformers were used in consumer electronics was in the '70s. These days, almost all power supplies are switching types, whose efficiency is determined by their design and the quality of parts used. Since nobody really tries to optimize their efficiency, they can be extremely inefficient, and often very poorly designed. Most of them are made by Chinese companies who usually copy other companies' designs without even understanding how they work.

      I think a law to mandate maximum standby power consumption (or, even better, incentives to minimize it as much as possible) can do quite a bit. Right now, it is simply not a factor in the design of equipment.

    9. Re: Convenience by alienw · · Score: 2, Interesting

      Actually, I don't think most modern TVs keep the tube warmed up. My TV takes the same time to turn on whether it's in standby mode or not. Nobody uses separate IR decoder chips, it's usually merged into the main ASIC and it's actually a microcontroller. If it's carefully optimized, it will draw about 0.3mA in sleep mode and a few milliamps when it runs. The real problem is the IR receiver (the ones I looked at draw 1-2mA), and more importantly the large relay which is supposed to switch on the main supply. It might be possible to solve this problem with something exotic, like a photo-triac, but it's definitely not simply a matter of putting in one capacitor.

  2. Software suspend: ACPI/APM independent by TeknoHog · · Score: 3, Informative
    Linux has software suspend in the kernel. It's the same as hibernation in Windows. Memory contents are saved to swap, and when you boot the same kernel again, it picks up where it left off. It's independent of APM/ACPI and you can use it on any Linux machine.

    While the vanilla version works basically, Suspend2 is a more complete implementation. I use it on my laptop regularly.

    --
    Escher was the first MC and Giger invented the HR department.
    1. Re:Software suspend: ACPI/APM independent by TeknoHog · · Score: 2, Informative
      My big question though is whether it will bring my KDE desktop up faster than a plain vanilla boot? Does software suspend get round all that time-consuming service startup (and all the rest of the stuff that goes on while those messages are scrolling up the console)?

      Yes. When you boot again, only the kernel is booted, which takes about two seconds or so. Then instead of init (all those startup services), it picks up the RAM image from the disk (takes a few seconds on my 512MB machine).

      --
      Escher was the first MC and Giger invented the HR department.
    2. Re:Software suspend: ACPI/APM independent by jonored · · Score: 2, Informative

      The simple answer: Yes. The system never shuts down, it just saves the current state of the machine to disk and stops - when it comes back, it comes back with every process that it stopped with still running, almost like nothing happened. The long answer: stuff that was talking to the outside world will probably have it's open connections dropped, and the network card might need to be restarted or something. Nothing big, nothing that should take much time. Not sure how long it takes to bring the system back, though.

    3. Re:Software suspend: ACPI/APM independent by TheRaven64 · · Score: 2, Informative
      takes a few seconds on my 512MB machine

      Most laptop hard drives are hard pressed to give a 25MB/s sustained read rate. I am not sure exactly how the Linux implementation exists, but you have basically two choices:

      1. You flush everything from user-space RAM to swap. When you are back up, you page things in as required. This is fast to sleep (most of the data is on disk already, just flush dirty pages), fast to wake up (just demand-page everything), but will slow the machine down a lot as all of your applications are now generating large numbers of page faults.
      2. You write the entire contents of RAM to a disk file. This wastes as much disk space as you have RAM (the contents of RAM is going to be in your swap space as well as in your suspend-to-disk file), and can take a long time to sleep and wake. Worst, the amount of time taken is proportional to the amount of RAM you have. Most laptop users try to have as much RAM as possible, since spare RAM acts as disk cache and reducing disk accesses is a really good way of prolonging battery life.
      It sounds like Linux uses the second one, which is far from ideal (oh, it also eliminates encrypted swap as an effective security measure - another big drawback for mobile users). It is particularly bad for laptops, since the slow speed of laptop hard disks makes the suspend / resume time a lot longer than on a comparable desktop. Now, when Flash prices drop to the point that allows putting a suspend file in Flash, then this might suddenly become interesting. Particularly if it were done above the OS layer so it could take the kernel and machine state with it.
      --
      I am TheRaven on Soylent News
  3. WOL by SillyNickName4me · · Score: 3, Informative

    WakeOnLan is basicly a matter of sending a 'magic packet' to the MAC address of the comuter you want to wake up. There is no need ofr the OS on that machine to actually support that functionality (except for there being a few cards around that require WOL to be re-enabled after each boot).

    Sending the 'magic packet' is not difficult, and there is a variety of tools that can do this, including a ready made perl script, on a gentoo system, type 'emerge wakeonlan'. I bet it is available with most other distributions as well.

  4. Compact Flourescent by the+eric+conspiracy · · Score: 3, Insightful

    You will save far more energy by investing in some compact flourescent light bulbs.

    1. Re:Compact Flourescent by brunes69 · · Score: 2, Informative

      I haven't checked, but I suspect he also leaves his TV and amp on standby.

      I don't know what model TV you have, but if I unplug mine to get it out of standby, I lose all the programmed in channels and settings. Next time I plug it back in I have to reprogram it all. Same with my reciever.

      No thanks, it's worth the $1 a month.

  5. ACPI ? What ACPI ? by Hymer · · Score: 2, Insightful

    ...servers run 24x365.25 no need for ACPI here.
    Save power somwhere else...

    1. Re:ACPI ? What ACPI ? by arivanov · · Score: 2, Informative

      Same here. But they have both ACPI and CPU frequency scaling (aka centrino, powernow, longrun, longhaul) enabled.

      The reasons are fairly simple.

      • Keeping the servers cooler decreases the electricity bill by up to 20% from ACPI and up to 75% from cpufreq for an average Pentium/Xeon based corporate install when idle. Athlon/Opteron numbers are comparable. This also decreases airconditioner wear and tear.
      • If your servers are well designed and have temperature controlled fans (all IBMs and Compaqs fall in this category) this increases fan MTBF by up to 2 times. Most importantly if a fan fails the server usually is cold enough to send out a clean alert and even work until you get around to replace it.
      • Your capacitor deterioration rate (they are the first thing to blow in a server nowdays) is considerably less which means that your overall server MTBF goes up and reliability increases considerably

      There is only one case where power management is unnecessary. It is if you are running a computational load which requires your servers to work 365x24x7 flat out. Running without ACPI and cpufreq (if supported) for an average corporate or ISP load is plain stupid. It results in a less reliable server installation.

      By the way, do not worry, 99.99 of the server sysadmins out there join you in this stupidity. In fact I did 7 years ago when I knew less about server administration.

      --
      Baker's Law: Misery no longer loves company. Nowadays it insists on it
      http://www.sigsegv.cx/
  6. Re:Layer 2 Access Required [Security?!?] by alienw · · Score: 2, Insightful

    There is a good reason why Java never became particularly popular, and this is it. It has an absolutely idiotic security model, in that it prohibits you from doing "dangerous" or "nonportable" things. Where "dangerous" or "nonportable" actually means "useful".

    With the appropriate privileges, Perl (or any other native application) can send any type of packets, layer 2 or above (possibly even layer 1). A programming language is NOT the place to impose security constraints. That's what the operating system is for.

  7. Re:Layer 2 Access Required [Security?!?] by SillyNickName4me · · Score: 2, Insightful

    Since perl scripts running on a Linux machine have access to the commandline, yes, they can send raw ethernet packages, given they run with the correct privileges and availability of correct tools. No native capability in perl for doing this is required at all. When running natively compiled JAVA on the same platform, and giving it access to the commandline, you'd have the exact same situation.

  8. Stop Blaming Environmentalists (was: Convenience) by nathanh · · Score: 3, Insightful
    Fusion is up in the air at the moment, nuclear will work if you ever get round the environmentalists,

    Yeah, because those damn environmentalists wield so much power and have so much money, why they're practically running the US government!

    and wind and tidal power can provide about 5%.

    That's nonsense. Slashdot ran an article on this just recently. Global wind power in class 3 areas alone could generate 72 terawatts which is 60 times global consumption. Class 3 wind turbines are financially comparable to brown coal. North America has the greatest number of class 3 areas in the world.

    But let's not stop at wind power. A home with solar panels for hot water (not the expensive, dirty and inefficient photovoltaic) saves 50% on heating costs. The panels pay for themselves in 5 years and have a 25 year lifetime. They are maintenance free (they are effectively just black plastic pipes behind glass sheets) and easy to repair when damaged (simple plumbing that a home handyman could do).

    But let's not stop at solar and wind power. Changing your light bulbs from incandescent to energy efficient flouros will save 75% on lighting costs. Modern flouros are compact, come in a variety of shapes, only need to be changed once every 5-10 years, degrade slowly rather than blowing suddenly at inconvenient times, and have equivalent candela output to a 75W incandescent.

    But let's not stop at solar power and wind power and energy efficiency. Your SUV gets 10MPG yet a comfortable Subaru Legacy has equivalent seating and storage but gets 33MPG. Your average driver will save between $750 and $1250 per year while simultaneously slashing their automobile oil consumption by two thirds. That's financially sensible and enviromentally friendlier.

    The solutions are here right now. You need to stop waiting for the magic silver bullet like fusion, or blaming "environmentalists" for preventing fission, or wondering why you're spending $2000+ per year on fuel for your gargantuan SUV, and simply start using the technology that is here right now and is economical right now and is practical right now. You can make the difference right now.

  9. Re:Layer 2 Access Required [Security?!?] by megabeck42 · · Score: 2, Informative

    Getting raw ethernet frames, in Linux, is trivial. For example, to get such a socket from python: (copied from the scapy startup routines)

    self.ins = socket.socket(socket.AF_PACKET, socket.SOCK_RAW, socket.htons(ETH_P_ALL))

    Whats that? You want to only get the traffic from a certain ethernet card, or want to transmit out a certain card? Easy. (In C, from my own code.)

            struct sockaddr_ll device_sa;
            int filedes;
            filedes = socket(AF_PACKET, SOCK_RAW, htons(ETH_P_ALL));
            device_sa.sll_family = AF_PACKET;
            device_sa.sll_protocol = htons(ETH_P_ALL);
            device_sa.sll_ifindex = ifindex;
            bind(filedes, (struct sockaddr *)&device_sa, sizeof(device_sa));

    You can get a device's ifindex with the SIOCGIFINDEX call, or with the rtnetlink RTM_GETLINK request.

    All you need for this is root, or suid root. In fact, I think you can just use CAP_RAW and CAP_NET_ADMIN, I believe.

    --
    fnord.
  10. Re:Stop Blaming Environmentalists (was: Convenienc by Shadow99_1 · · Score: 4, Informative

    Contrary to popular belief not everyone in the US owns a gas guzzling SUV... In fact my american car (Chrysler) gets better MPG than the subaru you mentioned.

    As for everything else...

    I did change my light bulbs, the effect I have to say seems negligable... Either my power company doens't bother to look at my meter to figure out energy use, or my lighting costs are fairly low and therefor the change made to small a difference to notice...

    I don't have solar panels used for hot water... My house is already built (& I didn't build it), so the cost to change things now is to much for my fairly average salary to cover... I could try to get a loan to pay for it, but that would nuke any savings I'd see for years... I also doubt anyone would give me a loan... Winter would also cause some issues for this... Heating bills are largest in the winter and solar isn't very effective in general when the solar panels are covered in snow...

    As for wind.... That's not a change I can make... Even if I did live in a good area (& I think winter would kill any effective use where I live), I'm not likely to be able to put up a tower... First their is the money issue again... Then I think my city would probably frown on it to... When I looked into wireless internet options I found out my area is heavily restricted on building anything over 30" tall... Less would probably not be so good with the number of trees around here...

    Nice ideas, but practicality is questionable...

    --
    we are all invisible unless we choose otherwise
  11. Re:Why are "standby" and "sleep" functions needed? by TeknoHog · · Score: 2, Interesting
    I don't understand why there should be hardware standby and sleep functions. The hardware should be provided with a means of reading and writing its entire state. When you power down, the contents of RAM and all the hardware state should get written to the hard drive. When you power up normally, instead of going through a lengthy boot process, it should read and restore the state the contents of RAM and the hardware state, and pick up where it left off.

    Linux's software suspend does just this, see my other post here. Well, in fact you have to boot the kernel normally (to initialize hardware, etc.) but instead of a lengthy init, you get the restored state.

    --
    Escher was the first MC and Giger invented the HR department.
  12. S1 sucks, S3 is great, if it works for you... by evilviper · · Score: 4, Informative

    I've been trying to get power management to work on PCs for over a decade now, and we're still not there...

    S1 (aka. sleep) works on most every system, since it's been around forever, but it'll only save you maybe 2% over the system being normally up and running (doing useful tasks).

    S3 (aka. suspend) is the damn-good one. It only uses about 0.5 watts more power than your computer being completely off (I suppose it might be different with a more effecient power supply like a Seasonic). However, it's damn near impossible to get it to work. Windows XP, Linux, FreeBSD. Tried on dozens of completely different machines, and I've never seen it work, once. The drivers for pretty much ALL the hardware need to be written with APCI in-mind.

    Hell, if I could just find a list of the motherboards, soundcards, and other components that have drivers on FreeBSD6 that will resume successfully from S3, I'd put together a couple systems with just those componets. Electricity in CA isn't cheap, and I'd be saving lots with instant-on from S3. No more boot-up waits, no more opening-up the same apps every time, etc. Just hit a button, and start working (as soon as the monitor can warm up).

    S5 (aka. hibernate) writes out RAM to disk, and reads from disk upon restart. I'm not a particular fan of this method, as it would take quite a while to resume on a system with a large ammount of RAM. Still, it has the potential to be even lower power provided you're going to be away long enough.

    So, in my experience, you're still screwed... Just shut-off the machine when you're done.

    --
    Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
    1. Re:S1 sucks, S3 is great, if it works for you... by TeknoHog · · Score: 2, Informative
      S5 (aka. hibernate) writes out RAM to disk, and reads from disk upon restart. I'm not a particular fan of this method, as it would take quite a while to resume on a system with a large ammount of RAM. Still, it has the potential to be even lower power provided you're going to be away long enough.

      I use Linux's software suspend (which I've already mentioned too many times in this discussion :) It only takes a few seconds to restore 512MB of RAM. One reason it achieves this is that RAM contents are compressed with LZF when written on the disk, to avoid the I/O bottleneck.

      --
      Escher was the first MC and Giger invented the HR department.
    2. Re:S1 sucks, S3 is great, if it works for you... by Trelane · · Score: 2, Informative
      Linux's support for S3 blows.
      Not really. It works; the problem is usually working around little hardware gotchas. This is where having a team of tech engineers do that work for you comes in, which is why it generally works better with Windows than with Linux--with Windows, HP, Dell, and friends do all the work for you (and, notably, for Microsoft. Great scheme for Microsoft!). With Linux, since it's not supported on the hardware, you have to do it yourself.

      It works anywhere between fine out of the box and totally crappy, depending on the particular quirks of your hardware (for example, ATI cards apparently need VBE information saved, whereas other vid cards apparently don't). A friend of mine's laptop worked 100% with S3; my Inspiron (very similar to his) with a newer card (otherwise same laptop) worked great--except the video was useless after resume. I then found VBEtool tricks, and now it works fine (except I have to do funky tricks due to ATI proprietary driver suckage; it works great with the 2d drivers).

      In conclusion, if you want to run Linux, don't buy a notebook "Designed for Windows"--it will only be guaranteed to work with Windows. Now, due to current market conditions, finding a fully-supported Linux notebook can be an exercise in furstration. If you know of a reliable way to get a Linux notebook, please drop me a line!

      --

      --
      Given enough personal experience, all stereotypes are shallow.
    3. Re:S1 sucks, S3 is great, if it works for you... by olman · · Score: 2, Informative

      Short answer: Don't use VIA chipset motherboards.

      Longer answer: I have never got S3 to work on a box running some variety of VIA chipset mobo. I have got it to work beautifully in every setup I have used that had SIS/Intel/Nvidia/AMD chipset on various W2k/XP boxes at work and home.

      Note that "beautifully" does not imply that getting initially working setup was painless operation. But with correct combination of (WHQ) drivers it has always been doable. On current version of home/work boxes it was pretty much plug + play.

  13. Re:Stop Blaming Environmentalists (was: Convenienc by libcoder · · Score: 2, Funny

    *ahem* I seem to remember a terawatt being one TRILLION watts, not one billion. That's just me, I could be wrong...but I'm not.

    --
    RIAA and the MPAA, putting the "F U" in "fair use".
  14. Kill-a-Watt by TClevenger · · Score: 2, Informative
    If you are truly concerned about what your equipment uses, get a Kill-a-Watt meter. Very easy to use, includes a KWh counter as well as an instant wattage display. You can find it around the 'Net for $27 or less.

    Some things I've tested recently:

    My PC speakers use 40 watts, even when "turned off". Result: they're on a power strip with a switch.
    My HP Laserjet 2100N uses 12-16 watts (depending on the fan), when in Power Saver. Result: it gets turned off when not in use.
    My PIII-650 desktop server consumed about 50 watts when idle. Result: replaced it with a Toshiba Tecra PIII-650 (with a broken screen, cheap on eBay), which draws 14 watts when idle.

    I also realized that my Powerbook power supply consumes less than 1 watt when plugged in but no laptop is connected, or about 2 watts when the laptop is plugged in and fully charged, so I'm not as concerned about unplugging it anymore.

    My next checks: the TV's, older transformer-based clock radios, wall warts and the deep freeze. I will also take running "baseline" checks of my major appliances (fridge, furnace, washer), so I can recheck them once a year and identify when an appliance is running too hard (bad motor bearing, etc.)