1. Computers
  2. Display Drivers
  3. Graphics Cards
  4. Memory
  5. Motherboards
  6. Processors
  7. Software
  8. Storage
  9. Operating Systems

Facebook RSS Twitter Twitter Google Plus

Phoronix Test Suite


The Tests Showing Ubuntu 11.04 On A Power Consumption Binge

Michael Larabel

Published on 23 April 2011
Written by Michael Larabel
Page 1 of 7 - 19 Comments

This is the prequel to Mobile Users Beware: Linux Has Major Power Regression and Uff Da! The Linux Power Bug Even More Mysterious. It was written in advance of tracking down the issue to a matter in the upstream Linux kernel. Though as the Phoronix Test Suite stack is presently bisecting and analyzing the kernel during the period in question (Linux 2.6.37 to 2.6.38), this article is being published now and hopefully on Easter Sunday or Monday the actual offending commit will be known along with much more information. This bug has also now been confirmed independent of Phoronix by at least six separate parties that I'm aware of, with reports of Natty either consuming excessive power or a very significant increase in heat output compared to Ubuntu 10.10. These independent reports have occurred on a range of hardware -- including desktops. There is also at least one bug report on the matter for Ubuntu 11.04.

Ubuntu 11.04 "Natty Narwhal" is set to be released on Thursday and while there are a number of new features to talk about in this latest release, the Phoronix Test Suite software has been busy analyzing the performance of this latest release. There is open-source graphics driver improvements leading to some performance improvements (such as Radeon KMS page-flipping), the famous ~200 line Linux kernel patch to improve responsiveness, and various other enhancements that catch our fancy in Ubuntu 11.04. However, one area where there is a frightening regression in Ubuntu 11.04 is with its power consumption. For mobile devices in many workloads, Ubuntu 11.04 is consuming noticeably more power than in any of the past Ubuntu Linux releases. Sadly, no one seems to have noticed in time since continuous integration testing on Linux seems to happen so haphazardly right now.

I have been testing four different mobile devices (notebooks / netbook) under Ubuntu 11.04 and every previous release down to Ubuntu 8.04 LTS or the oldest release where each system is fully supported by a clean stock installation. With each device I've been using Bootchart to analyze the boot performance as well as running a number of Phoronix Test Suite / OpenBenchmarking.org tests to extensively analyze the performance -- including the power consumption under different workloads. Various articles will come from this information in the following days, but today there are the Ubuntu power consumption results.

These results show that, overall, there's a trend of Ubuntu 11.04 burning through significantly more power at least when using Intel hardware. In fact, just how bad Ubuntu 11.04 trips up in terms of its power usage is frightening as it will be noticeable to end-users when the battery life is shorter by up to 10% or more -- in some cases, as much as a 26% difference to Ubuntu 10.10!

For the power consumption tests, the first "idle" test result for each system was measuring the power consumption while each device was running on battery power. During the course of three minutes, the Mozilla Firefox web-browser was opened and then navigated to Phoronix.com, then closed and OpenOffice.org (or LibreOffice in the case of Ubuntu 11.04) launched and opened to a new document, then closed and Nautilus launched followed by navigating through the GNOME menu and the remainder of the time was spent idling. This is just a test to show power consumption during light desktop usage for each computer.

To see the battery power consumption rate when the CPU is fully maxed, we used the OpenSSL test profile to peg all available CPU cores with an encryption workload. For systems where there are open-source GPU drivers with OpenGL acceleration "out of the box" on Ubuntu, there is also OpenArena gaming test results to show the power consumption when the CPU and GPU are busy. In addition, on some systems where the disk is not too slow, there are also PostMark results to show the battery power burn rate when the disk and CPU are taxed.

Here is a run-down of the test systems used for this article:

ThinkPad R52: Intel Pentium M 1.86GHz single-core, IBM 18494WU motherboard SKU with Intel Mobile i915 + ICH6M chipset, 2GB of system memory, 80GB Hitachi HTS541080G9AT00 IDE HDD, ATI Mobility Radeon X300 graphics

ThinkPad T60: Intel Core Duo T2400 dual-core, Lenovo 2613EJU motherboard SKU with Intel 945 + ICH7M chipset, 1GB of system memory, 80GB Hitachi HTS541080G9SA00, and ATI Mobility Radeon X1400 128MB graphics.

ThinkPad T61: Intel Core 2 Duo T9300 dual-core (x86_64), Lenovo 6459CTO SKU, Intel PM965 + ICH8M-E chipset, 4GB of system memory, 100GB Hitachi SATA HDD HTS72201, and NVIDIA Quadro NVS 140M graphics.

Samsung NC10: Intel Atom N270, Intel 945GME + ICH7-M, 2GB of system memory, 32GB OCZ Core SSD, and Intel i945 graphics.

It's just a collection of hardware in the office. Unfortunately, no mobile AMD devices were around for testing. Here is a recap of the key components in each of the Ubuntu Linux releases. With each Ubuntu installation on each system, a clean installation was done each time to occupy the entire disk and the stock package selection and settings were used.

Ubuntu 8.04.4 LTS: Linux 2.6.24 kernel, GNOME 2.22.3, X.Org Server, xf86-video-ati 4.3.0, Mesa 7.0.3-rc2, GCC 4.2.4, EXT3.

Ubuntu 8.10: Linux 2.6.27 kernel, GNOME 2.24.1, X Server 1.5.2, xf86-video-ati 6.9.0, Mesa 7.2, GCC 4.3.2, EXT3.

Ubuntu 9.04: Linux 2.6.28 kernel, GNOME 2.26.1, X.Org Server 1.6.0, xf86-video-ati 6.12.1, Mesa 7.4, GCC 4.3.3, EXT3.

Ubuntu 9.10: Linux 2.6.31 kernel, GNOME 2.28.1, X.Org Server 1.6.4, xf86-video-ati 6.12.99, xf86-video-nv 2.1.14, xf86-video-intel 2.9.0, Mesa 7.6, GCC 4.4.1, EXT4.

Ubuntu 10.04.2 LTS: Linux 2.6.32 kernel, GNOME 2.30.2, X.Org Server 1.7.6, xf86-video-ati 6.13.0, xf86-video-nouveau 0.0.15, xf86-video-intel 2.9.1, Mesa 7.7.1, GCC 4.4.3, EXT4.

Ubuntu 10.10: Linux 2.6.35 kernel, GNOME 2.32.0, X.Org Server 1.9.0, xf86-video-ati 6.13.1, xf86-video-nouveau 0.0.16, xf86-video-intel 2.12.0, Mesa 7.9-devel, GCC 4.4.5, EXT4.

Ubuntu 11.04 Beta 2: Linux 2.6.38 kernel, Unity 3.8.6, X.Org Server 1.10.1, xf86-video-ati 6.14.0, xf86-video-nouveau 0.0.16, xf86-video-intel 2.14.0, Mesa 7.10.2 / Gallium 0.4, GCC 4.5.2, EXT4.

With the Lenovo ThinkPad T61 the Ubuntu 64-bit release was used since the Intel CPU is x86_64 capable, but for the rest the 32-bit version was tested. This was automated using the Phoronix Test Suite and related Phoronix components.

<< Previous Page
Latest Linux Hardware Reviews
  1. A Walkthrough Of The New 32 System Open-Source Linux Benchmarking Test Farm
  2. Habey MITX-6771: Mini-ITX Board With Quad-Core J1900 Bay Trail
  3. OCZ Vector 150 SSD On Linux
  4. Noctua i4 CPU Cooler: Great For Cooling High-End LGA-2011v3 CPUs
Latest Linux Articles
  1. AMD Kaveri: Open-Source Radeon Gallium3D vs. Catalyst 14.12 Omega Driver
  2. 12-Way AMD Catalyst 14.12 vs. NVIDIA 346 Series Linux GPU Comparison
  3. AMD Catalyst 14.12 Omega Driver Brings Mixed Results For Linux Users
  4. 6-Way Winter 2014 Linux Distribution Comparison
Latest Linux News
  1. Linux 3.19 Merge Window Closes Ahead Of Schedule
  2. MIPS R6 Architecture Now Supported By GCC
  3. LowRISC To Feature Tagged Memory & Minion Cores
  4. Intel Skylake Audio Support For Linux 3.19
  5. After 10+ Years, NetworkManager Reaches v1.0
  6. VDPAU Updated To v0.9
  7. An Open Hardware Random Number Generator Proposed
  8. LLVM 3.6 Will Be Branched Next Month
  9. Opera Browser Puts Out Linux Updates For The Holidays
  10. GNOME Shell 3.15.3 Adds Support For High-Contrast Themes
Latest Forum Discussions
  1. The New SuperTuxKart Looks Better, But Can Cause GPU/Driver Problems
  2. XLennart: A Game For Systemd Haters With Nothing Better To Do
  3. Need some hand holding with upgrading xserver
  4. Updated and Optimized Ubuntu Free Graphics Drivers
  5. Debian init discussion in Phoenix Wright format
  6. FPS capped on Linux (AMD fglrx drivers)
  7. Are there an app using HSA ?
  8. Bench specific mount point