How Ubuntu 12.04 Is Trying To Drop Power Usage

[phoronix]

Phoronix: How Ubuntu 12.04 Is Trying To Drop Power Usage

After illustrating Linux power regressions and other problems for months, with Ubuntu 12.04 LTS developers at Canonical are finally taking a serious look at Linux power management and how it can be bettered...

http://www.phoronix.com/vr.php?view=MTA0OTU

[ua=42]

Ubuntu consumed more power than Windows 7 during the following separate workload tests: idling, having a web browser open, playing an MP3, running Mozilla Thunderbird, using BBC's iPlayer, and suspending.

Soo... it consumed more power than windows 7. Were there any tests were ubuntu consumed less power? Seriously, that covers about 95% of my netbook use by time spent. (mp3, webbrowser, suspend, idle).

[cking]

Thanks for writing up the work that we've been doing on Power Management. It may be worth holding off a little longer before comprehensively re-measuring the savings since we are still working on reducing wakeup events on a bunch of applications.

[cking]

Soo... it consumed more power than windows 7. Were there any tests were ubuntu consumed less power? Seriously, that covers about 95% of my netbook use by time spent. (mp3, webbrowser, suspend, idle).

Actually with the i915 rc6 workaround webrowsing and mp3 playback is better than Windows but Phoronix just seemed to be reworking my summary and didn't seem to mention the source data produced during the investigation. See http://zinc.canonical.com/~cking/pow...-vs-ubuntu.ods

[FireBurn]

Thanks for writing up the work that we've been doing on Power Management. It may be worth holding off a little longer before comprehensively re-measuring the savings since we are still working on reducing wakeup events on a bunch of applications.

What work is Canonical doing to fix these bugs? It seems the only fix they have is backporting the ASPM fix (which someone else wrote)

Everything else mentioned is talking about other features that aren't ready yet - except for the power.d script (which implements someone else's recommendations)

I'm asking out of curiosity as I know they will be doing something - it's just not clear what that something is from this article

[cking]

Soo... it consumed more power than windows 7. Were there any tests were ubuntu consumed less power? Seriously, that covers about 95% of my netbook use by time spent. (mp3, webbrowser, suspend, idle).

Actually, with i915 rc6 enabled Ubuntu was more power efficient on a Lenovo x220i than Windows - see http://zinc.canonical.com/~cking/pow...-vs-ubuntu.ods - however this data seems to have been not noticed. I suspect Phoronix has just re-worked the summary of the findings.

[cking]

So what are is Canonical doing?

* Crowd-sourced the testing for the PCIe ASPM (https://wiki.ubuntu.com/Kernel/PowerManagementASPM) to ensure we didn't cause regressions by including Matthew's patch in an LTS. Kudos to Matthew for this.

* Measured all the powertop recommendations across a bunch of machines and found a common set of improvements that should not cause any regressions for an LTS and added them into pm-utils. Maybe trivial, but we wanted to ensure we didn't cause any major upsets for an LTS.

* Worked through a list of possible power saving options suggested during UDS: https://blueprints.launchpad.net/ubu...wer-management - this does take time to do accurately and methodically across a range of hardware. We have a lot of details results here and the Phoronix write up is a re-working of the top level summary of this work. I hope readers examine the original source data and understand how much effort has been put into identifying the low-hanging fruit. The data can be found at: http://zinc.canonical.com/~cking/power-benchmarking/ It is useful to understand the kinds of power savings possible and understand exactly where to focus attention. Some of the work was just to identify low-risk low-hanging fruit.

* Identified applications that were sub-optimally doing wakeups and we have engineers working on fixing these right now. See https://bugs.launchpad.net/ubuntu/+b...er-consumption - as you may appreciate some issues are not quick fixes and require engineering time and effort to address.

For those interested, there is a Power Management Wiki page: https://wiki.ubuntu.com/Kernel/PowerManagement

These issues are on our radar and we are devoting effort to this cause. This is work in progress.

As ever, we are always very grateful for the community help in crowd-source testing some of these fixes.

[bwat47]

Glad they are finally making power usage a priority. Ubuntu is pretty sub-par on a laptop right now when it comes to battery life. I get better battery life on arch with laptop-mode-tools.

[grotgrot]

The issues with Canonical's own code for Ubuntu One have been known for over 2 1/2 years. https://bugs.launchpad.net/ubuntu/+s...nt/+bug/475447

[liam]

Phoronix: How Ubuntu 12.04 Is Trying To Drop Power Usage

After illustrating Linux power regressions and other problems for months, with Ubuntu 12.04 LTS developers at Canonical are finally taking a serious look at Linux power management and how it can be bettered...

http://www.phoronix.com/vr.php?view=MTA0OTU

For those interested, there was a thread based upon a Phoronix article a few months ago that had the best power savings suggestions I've yet seen. As a result my idling X220t now consumes something around 7W at about .5 brightness. Also, don't blindly follow PowerTop recommendations. Some, like idling usb components, don't save alot but are a pain b/c wakeup times can be quite large if you've got something using them.
You can also make use of the tune daemon (tuned) which has profiles for several different usage patterns.
Lastly, a new framework was added to the current stable kernel for providing voltage regulation to many standard mobo compnwnts which use i2c (IIRC). When the drivers are updated to take advantage of this, we should see even better idling times.