Linux laptops, revisited (can any sleep like my PowerBook does?)
Michael L Torrie
torriem at chem.byu.edu
Mon Jan 21 22:45:58 MST 2008
Charles Curley wrote:
> On Mon, Jan 21, 2008 at 10:44:43AM -0700, Michael L Torrie wrote:
>> This thread has been enlightening, to say the least. But not super
>> encouraging. What I'm getting from this thread is that suspend to ram
>> is possible on some laptops. But it may require some assembly.
> Ubuntu 7.10 just fell into place on my Thinkpad R51
> (http://www.charlescurley.com/Lenovo.R51.html), including
> suspending. While I haven't tried suspending to RAM, suspending to
> hard drive works just fine.
You've inadvertently confirmed what I am coming to fear. That
suspending doesn't just work. As you experience, hibernating to disk
works for the vast majority of users on most laptops (or even desktops).
Mainly because it has nothing much to do with power management.
Suspending, in my mind, means suspend to RAM (sleep). I would be
interested to hear of your experiences with this.
After using my powerbook for 4 years or more, I don't know how anyone
lives without it, or stands for anything other than fast suspend and
resume. Powering on and off the entire laptop (which is what hibernate
aka suspend to disk still involves) is time consuming. In this day and
age we should expect/demand that our machines go to sleep very quickly
and awake, ready for work, in the state we last left them, in just a
second or two. And I think I can legitimately demand this from linux on
a thinkpad, of all machines. I find it really odd how many of the folks
I've talked to think hibernation is good enough.
I suppose it saves energy in the long run. But frankly it's
embarrassing to pop the lid on a laptop and see it run through the bios,
load the kernel, load the suspend to disk image back, etc before I can
work with it. Especially when all the Apples around me (and even the
windows laptops now) just pop their lids and get up and running in a
second or two.
>> Furthermore from what Thinkpad owners have told me (this extends to all
>> Linux laptops from what I can see) that battery life is well below
>> Windows XP on the same machine. 3 hours is considered "good" for a T61p
>> with a 9 cell battery. So for a variety of reasons, many of which would
> I get more than 4.5 hours out of my 6.6 amp-hour (??) battery.
For a comparison, my 4.0 amp-hour battery, when it was new, lasted for 4
hours on my PB 12". A new battery still would. I'm not sure the volts
on your battery. My battery was something like 11 volts (6 cells in a
3S2P configuration, for any RC junkies out there).
> aptitude install powertop
> Very nice, thank you. It is from Intel, so who knows what it will do
> on AMD processors/chipsets.
Works fine on AMD machines, for the most part.
> Actually it runs just fine while plugged in, which means it is good
> preparation for running on battery. While plugged in, right now my
> processor is running at 600 MHz 97% of the time, and 95% in C2 state,
> whatever that means.
It runs fine, but can't give you a CPU electricity use estimate when the
AC is plugged in.
Idling the CPU down to the lowest speed when not needed is indeed a very
good thing. The bigger the C number (like C4), the less power the CPU
uses in that state, but the longer it takes the CPU to return to a state
where it can process instructions, or something to that effect. I read
somewhere that on laptops, you should shoot for most of your time in the
C3 state when you are trying to max out battery life. See the powertop
FAQ on this.
It turns out that lots of common software, including firefox,
gaim/pidgin, and even gvim can, without correct settings and even
patches, sap your battery. See
http://www.lesswatts.org/projects/powertop/known.php for tips.
> PLUG: http://plug.org, #utah on irc.freenode.net
> Unsubscribe: http://plug.org/mailman/options/plug
> Don't fear the penguin.
More information about the PLUG