Weekly Edition
Return to the Kernel page
| |||||||||||||
Powering down APM
The APM power management interface has never been much loved - even ACPI
was seen as a better alternative. There has been little or no hardware
made which depends on APM for some years; Windows evidently stopped
supporting it in 2006. Linux does still support APM, though, and that
support has a cost, so it is perhaps not surprising that Len Brown would like to remove that support as of
2.6.40.
Removal of APM support on that schedule is almost certainly not going to happen; a number of developers have expressed concerns that there may still be hardware out there in use which would then be unable to run new kernels. In general, the Linux kernel tries not to abandon users running older hardware. So APM may stay for a while, but there is a problem: keeping APM support, it seems, conflicts with some needed changes to the cpuidle code. The need to keep APM working, in other words, threatens to hold back improvements for the majority of users who have more current hardware. The solution to this conflict may take the form of a partial removal of APM support. The most important APM feature for users of old systems is likely to be the ability to power-off the system; other features may be less important. As Andi Kleen noted, idle support probably matters less to such users:
Phasing out APM idle at least would be reasonable. Presumably even
if the old laptops still work they are likely on AC because their
batteries have long died. So using a bit more power in idle
shouldn't be a big issue.
So APM support, as such, may stick around for a while, but it may begin to lose features as the kernel moves on. (Log in to post comments)
Who said anything about APM only being used by laptops? Posted Mar 31, 2011 12:52 UTC (Thu) by rankincj (subscriber, #4865) [Link]
My old P200MMX (~ 1997) has long since been relegated to serving as a broadband switch/router, and it still uses APM. It's my only PC which never gets powered off too.
Who said anything about APM only being used by laptops? Posted Mar 31, 2011 19:28 UTC (Thu) by zlynx (subscriber, #2285) [Link]
If it is never powered off, what does it use APM for?
I don't think the original Pentium even had a power idle mode. Did it?
I've got a dual Pentium 166 still running my network gateway. I'm pretty sure all power management is disabled in the BIOS because all it ever did for me was trigger bugs.
Who said anything about APM only being used by laptops? Posted Apr 1, 2011 14:42 UTC (Fri) by rankincj (subscriber, #4865) [Link]
I do ask it to shutdown or reboot itself occasionally. And also to blank the screen for the large percentage of the time when I'm not looking at its console. IIRC, it always needed APM to do that.
Who said anything about APM only being used by laptops? Posted Apr 3, 2011 1:06 UTC (Sun) by giraffedata (subscriber, #1954) [Link] Good grief, your network switch/router has a screen? And it's normally switched on? CRT, I presume.
Who said anything about APM only being used by laptops? Posted Apr 3, 2011 12:44 UTC (Sun) by rankincj (subscriber, #4865) [Link]
No, it shares a monitor with another PC. (The monitor has both VGA and DVI inputs, and it's trivial to switch between the two.) I am assuming that the VGA graphics card is using less power once it blanks the console; the monitor certainly doesn't appear to be receiving a signal afterwards because it switches itself into "Analog Power-Save" mode.
Who said anything about APM only being used by laptops? Posted Apr 3, 2011 16:42 UTC (Sun) by mgedmin (subscriber, #34497) [Link]
I you bought an embedded Linux router to replace it, I wonder how soon you would recover the cost just from energy savings.
Who said anything about APM only being used by laptops? Posted Apr 3, 2011 18:41 UTC (Sun) by rankincj (subscriber, #4865) [Link]
No idea - but would an embedded device also come with a 80 GB hard drive and have regular security updates for its firmware ;-)?
Anyway, its biggest failing these days is that its ports are all 100Mbps. Eventually, it will need replacing with something that has 1Gbps ports.
Who said anything about APM only being used by laptops? Posted Apr 3, 2011 19:44 UTC (Sun) by Trelane (guest, #56877) [Link]
> would an embedded device also come with a 80 GB hard drive
Just about any would, over e.g. NFS or samba.
Additionally, some wireless or wired routers also have USB capability that you might use for an external storage device (not tried, but seems plausible).
> have regular security updates for its firmware
That's pretty much guaranteed if you're running a mainstream alternate firmware, e.g. dd-wrt or openwrt.
It's a good reason to use alternate firmware, besides the better features.
Who said anything about APM only being used by laptops? Posted Apr 3, 2011 19:46 UTC (Sun) by Trelane (guest, #56877) [Link]
> I wonder how soon you would recover the cost just from energy savings.
Back when I replaced my Pentium 100 + access point combo I'd been using for a while with a WRT54G (using this as a primary motivating reason), I figured it'd be about half a year to a year. This was early on, too (my router is WRT54G v1.1).
I no longer have the basis for the numbers, but that was a back of the envelope calculation done a long time ago.
Who said anything about APM only being used by laptops? Posted Apr 3, 2011 20:08 UTC (Sun) by giraffedata (subscriber, #1954) [Link] I wonder how soon you would recover the cost just from energy savings. The accountant in me has to point out that this (time to recover the cost) is not the proper measurement here. The proper measurement is the present value of all the electricity savings for all time minus the cost of the new hardware. My company once looked at replacing CRTs with LCDs for the energy savings and did the calculations based on time to recover dollar cost and found it would get its money back in 3 years. Since we planned still to be in business in 3 years, it was said to be a good deal. But I redid the calculations and showed that the electricity for the remaining life of the CRTs was cheaper than the price of the LCDs. That's because to buy the LCDs, the company would have to reduce its investment in its core business, thus reducing its income by more than the cost of the electricity. And of course, one assigns a value to the finite labor and hassle involved in changing to a whole different router model, which for me pretty much always exceeds the present value of electricity saved.
Who said anything about APM only being used by laptops? Posted Apr 3, 2011 20:23 UTC (Sun) by Trelane (guest, #56877) [Link]
> (time to recover the cost) is not the proper measurement here. The proper measurement is the present value of all the electricity savings for all time minus the cost of the new hardware.
I'm just a physicist; could you please reformat this as an equation? With the variables and constants clearly explained.
> showed that the electricity for the remaining life of the CRTs
What does this mean, precisely?
> cheaper than the price of the LCDs
These aren't the same units. Electricity is in energy and price of LCDs is in some currency. Presumably you're converting units someplace; where is that taking place?
> to buy the LCDs, the company would have to reduce its investment in its core business
I don't understand why you had to reduce investment in core business.
Who said anything about APM only being used by laptops? Posted Apr 3, 2011 20:34 UTC (Sun) by dark (subscriber, #8483) [Link] I'm not a physicist, but maybe I can act as a bridge here.giraffedata was comparing two possible scenarios: 1. Continue to use the CRTs until they were due for replacement anyway (let's call that time T1), then replace them with LCDs. 2. Discard the CRTs now (let's call that T0) and replace them with LCDs In scenario 2 you use less electricity in the timespan from T0 to T1, which saves money, but you also move the expense of buying the LCDs from T1 to T0, which means you lose the profit you could have made from investments of that money between T0 and T1. For a profitable business, the most natural investment is in its own operations, which is why giraffedata was talking about income from the core business.
Who said anything about APM only being used by laptops? Posted Apr 4, 2011 0:53 UTC (Mon) by Trelane (guest, #56877) [Link]
Thanks; that helps.
Who said anything about APM only being used by laptops? Posted Apr 5, 2011 5:00 UTC (Tue) by cmccabe (guest, #60281) [Link]
There's a psychological cost to having to stare at old and crappy gear all day. The morale boost alone might justify buying LCDs, in the same way it justifies Facebook giving free fancy bottled drinks to its employees.
Also, a lot of companies are sitting on big cash reserves now. For example Apple has about $43 billion in cash right now; Microsoft has about $40 billion. In that case, an extra dollar of capital expenditure will not come directly out of investment in the core business.
Maybe your company is small enough that capital expenditure is a real worry. Even in that case, though, the monitors are probably a good deal. Dropping $100 on a monitor is cheap compared to most of the other perks a company can offer to developers. For example, giving someone a half-day (assuming they're making 80k/year) is $164. I'm sure you can do the math for other stuff like catered lunches, free daycare, bigger offices, and so forth. The monitor is awfully cheap and awfully effective.
Who said anything about APM only being used by laptops? Posted Apr 5, 2011 6:46 UTC (Tue) by giraffedata (subscriber, #1954) [Link] It's true that there are lots of other benefits to replacing CRTs, and costs as well, that we haven't even begun to talk about. But this thread is about the proper way to balance the purchase price of an LCD monitor against the electricity savings. In the example I gave, the proposer said the electricity savings would pay for the purchase. He was wrong. No company has enough cash that it can invest for free. A company has cash because it believes in the near future it will have an opportunity to invest it at a higher rate of return than right now. Spending that cash on LCD monitors would deprive it of that opportunity, so the monitors would have to return more than whatever else the company had in mind for that cash. Also, even "cash" earns interest, so the electricity savings (and whatever other benefits there are to monitors purchased with that cash) must at least exceed that interest. Incidentally, the reason that return on investment in the core business is usually the proper alternative to compare is that management always believes it can make more money off the core business than anything else. Otherwise, it would dissolve the company and return the assets to the stockholders. A business' other investments are normally short term financing for long term core business investment.
Who said anything about APM only being used by laptops? Posted Apr 5, 2011 16:06 UTC (Tue) by nye (guest, #51576) [Link]
>In the example I gave, the proposer said the electricity savings would pay for the purchase. He was wrong.
You haven't shown that though. All you've pointed out is that in a hypothetical scenario you posit, there are other balancing concerns.
In particular, the scenario you describe is valid only for [a subset of] businesses.
Who said anything about APM only being used by laptops? Posted Apr 6, 2011 2:58 UTC (Wed) by giraffedata (subscriber, #1954) [Link] In the example I gave, the proposer said the electricity savings would pay for the purchase. He was wrong.You haven't shown that though. All you've pointed out is that in a hypothetical scenario you posit, there are other balancing concerns. Other balancing concerns? You mean the value of time? That's not a balancing concern, but an intrinsic part of any comparison of an immediate payment versus a series of payments over time. Not considering time in this comparison is like not considering the shipping charge when comparing prices on Ebay. Incidentally, this example is a true story, from a very ordinary business, as opposed to one operating in a black hole where time has no value.
Who said anything about APM only being used by laptops? Posted Apr 6, 2011 15:20 UTC (Wed) by nye (guest, #51576) [Link]
>Other balancing concerns? You mean the value of time?
>... >Incidentally, this example is a true story, from a very ordinary business, as opposed to one operating in a black hole where time has no value.
I honestly have no idea how what you're talking about relates to the value of time. Could you please explain further in words of one syllable?
If it helps, my idea of the value of time is something like this: it takes me X hours to perform task Y; that's X hours I could have spent doing something else and which I'll never get back. Hence, there is a cost to performing task Y.
Who said anything about APM only being used by laptops? Posted Apr 6, 2011 16:30 UTC (Wed) by giraffedata (subscriber, #1954) [Link] Not wanting to do a task because it takes time is part of the concept of time having value that I'm talking about. But more basically, it's the fact that a person generally prefers to have something now than to have it later. It's the reason you have to pay interest on a loan. It's the concept behind the business adage, "expense delayed is expense reduced." I think it's an unsettled question among economists exactly what the psychological basis for that is, but I know it involves the limited lifespan of a person, uncertainty which grows the further into the future you look, and the fact that entropy increases with time. There are myriad ways you can bring this abstract concept into focus in making a business decision by looking at interest you would pay on a loan or income you would make on an investment in pay now vs pay later scenarios, and I tried to do that by describing how my company would end up wealthier by investing money in something other than LCD monitors and continuing to pay, in the future, for the extra electricity for the CRTs. (With all other factors ignored, of course). Even though the total dollar electricity savings exceed the purchase price of the LCD. But the basic concept is well known to beginning economics and accounting students as "present value." $10 paid a year from now is worth less than $10 paid now. (And there is math to tell you how much less).
Who said anything about APM only being used by laptops? Posted Apr 11, 2011 11:19 UTC (Mon) by nye (guest, #51576) [Link]
I still remain highly unconvinced that this applies in the given situation in the real world, but if I find the time (as it were) I'll try to do some research into the relevant economic theory.
Thanks for humouring me.
She said she said Posted Apr 11, 2011 19:07 UTC (Mon) by gvy (guest, #11981) [Link]
> It's the reason you have to pay interest on a loan.
Actually it's not a reason but rather a bait. But then again, given that interest-based loans are forbidden in all of the three monotheistic religions, the kludgery to make this social exploit work looks evil enough. I'm quite happy being free enough not to subscribe to any loans at all.
PS: my first notebook, a Fujitsu-Siemens C6535, still works and APM functions just fine there (suspend, hibernate, poweroff, blank). Wish all of my Thinkpads (A/T/X series) were that durable and dependable.
She said she said Posted Apr 11, 2011 20:24 UTC (Mon) by nybble41 (subscriber, #55106) [Link]
> > It's the reason you have to pay interest on a loan.
> Actually it's not a reason but rather a bait.
If the interest is bait then it must be bait for the one *issuing* the loan, not the one paying the interest, which seems rather the opposite of what you were implying. (Unless you're arguing that predatory borrowers use the promise of interest to bait traps for unwary lenders? That would be quite a reversal.)
Anyway, giraffedata was correct: time preference is why those who take out loans must pay interest. The present value to the lender of the money loaned today is less than the present value of the same money a year from now, so the borrower must pay some additional amount back (the interest) to make up the difference. A truly interest-free loan would be a net loss to the lender, which makes such loans very rare indeed.
Time preference is also the reason the borrower is *willing* to pay that interest, as the present value to the borrower of the larger amount to be paid back in a year is less than the present value of the amount received today. This is possible because different people have different time preferences, which also vary over time. Those with higher time preferences borrow from those with lower time preferences for mutual benefit.
She said she said Posted Apr 11, 2011 20:28 UTC (Mon) by giraffedata (subscriber, #1954) [Link] It's the reason you have to pay interest on a loan.Actually it's not a reason but rather a bait. Bait is something that leads someone into a trap -- i.e. misleads the person into doing something he doesn't want to do. What bait are you talking about? I can assure you that the reason I won't lend you money unless you pay me interest is that I value time. I would have to delay my enjoyment of the things I would spend that money on, and I won't do that voluntarily without compensation. No trap, no bait. I have long wanted to study ancient usury laws, because if they are what they seem to be, it's hard to imagine people wanting them.
She said she said Posted Apr 11, 2011 22:48 UTC (Mon) by cmccabe (guest, #60281) [Link]
Firstly, there are more than 3 monotheistic religions. Secondly, Judaism and Christianity don't forbid interest-based loans.
From wikipedia:
> The Hebrew Bible regulates interest taking. Interest can be
As far as Christianity goes... the New Testament doesn't say anything about charging interest on loans. Attitudes have gone back and forth over the centuries.
The Koran, however, does completely forbid charging interest.
Powering down APM Posted Apr 6, 2011 16:03 UTC (Wed) by mlankhorst (subscriber, #52260) [Link]
a power management interface that makes ACPI feel loved? Surely it can't be THAT bad..
|
|||||||||||||