Is Linux Power Management Getting Better Or Worse?
Phoronix: Is Linux Power Management Getting Better Or Worse?
With three laptops representing different generations of mobile hardware, we loaded up the past four stable releases of Fedora Linux plus the most recent Fedora 14 Alpha release and then carried out an arsenal of tests looking at how the battery power consumption rate has changed since 2008. If you are concerned at all about running Linux on your battery-powered mobile devices, this article is worth reading.
It's not just about software improving but also hardware. 2.6.35 introduced the intel_idle driver, which makes a massive difference for CPUs that support it but none of those tested do.
I would say that the higher power consumption on the Core2 notebooks are due to the open source graphics drivers not being very well optimized in this regard.
It's too bad there aren't any netbook performance numbers. My guess is that we would see something like what the tests showed with the R52.
I remember installing Fedora 10 on my Eee PC 1000H with a usb flash drive, although it had to be formated in a very particular way in order to boot.
It would be hard to argue that power management is getting worse. When I got my i7 there wasn't even proper P-state support for frequency scaling. A proper test would throw on some minimal distro like Gentoo or Arch, add the linux-phc patchset for dynamic undervolting, enable kernel frequency scaling, do some powertop tweaks, use opensource drivers with supported power management.
Surely power management isn't getting worse. Probably more like a side-effect of whatever happens to be running in a default Fedora install.
Have you tested to make sure that polling acpi every second doesn't have a noticeable effect on (idle) power use? If wakeups have been brought under control in recent releases, could the additional wakeups caused by too-frequent polling by the test suite wipe out some of the gains made?
I have a couple things to comment on with this;
1) pushing the hardware to its limits will disable power management. As a result ONLY the FIRST test was particularly valid. The rest could be only related to how much harder the hardware is actually being pushed by code that is more optimized for it, i.e. making use of hardware that was idle before will improve performance at the expense of power consumption.
2) Fedora 14 is running DEBUGGING KERNELS STILL. Come on -- you've been caught making this mistake MANY MANY MANY times before!
I wish someone released a notebook optimized distro (with optimized meaning battery duration).
Right now I unfortunately don't have time to dig into this, so I would accept a "ready" distro, without need to tweak.
Until I am able to do try the tweaks mentioned in another thread, my notebook's power consumption on linux is way bigger than on Windows.
Thanks for testing anyway. I don't think any distro is optimized for notebooks at the moment. Besides, if the tests were run on a computer with a weaker GPU, the other aspects of power management could be studied in more detail. I think the readers on this forum know quite well where GPU power management is going at the moment.
Unfortunately the measurements taken are meaningless in practice. If I'm going to run something CPU-heavy then I'm most probably connected to mains power.
What is most important for me as mobile user is how long my battery lasts when the CPU is mostly idle and I'm performing light work, like web browsing or word processing.
Unfortunately phoronix is a one man army so he probably doesn't have time to make extensive battery charge tests. Although I agree with you in that it would be far more useful to know how much the battery lasts in practice.
Originally Posted by markus_b
Thank you for the power tests. Could you also test with the close source gfx drivers (idle test + light internet browsing would be enough)? I'm curious about the differences (if there are any).
Tags for this Thread