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

OpenBenchmarking.org

A Proper Solution To The Linux ASPM Problem

Michael Larabel

Published on 11 November 2011
Written by Michael Larabel
Page 3 of 3 - 40 Comments

Migrating from the Linux 2.6.37 to 2.6.38 kernel caused the power consumption to go up by nine Watts due to this ASPM regression. This is an increase in power consumption by 36%. It works to set pcie_aspm=force on the ThinkPad notebook, but up through the latest Linux 3.2 kernel code as of today the regression still stood in its stock configuration. However, when applying Matthew's patch, the "out of the box" power performance finally returns to pre-2.6.38 levels!

Under graphics load the ASPM regression causes the Core i7 notebook to be burning through six more Watts (an increase of 14%) than when ASPM is active. When trying out today's ASPM patch, the power is back to the levels experienced on Linux 2.6.37. Hell yes!

The results from this ThinkPad notebook are similar to the several other notebooks I have tested the past few hours.

Since the Linux 3.2 kernel merge window closed earlier this week and this is a behavioral change rather than being just a simple fix or reverting the earlier commit, it might not see the mainline tree until the Linux 3.3 kernel. If it does not make it into 3.2, hopefully the patch will at least be carried into the Linux 3.2 kernel package for Ubuntu 12.04 LTS and other distributions.

The patch can be currently fetched from the kernel mailing list and it applied cleanly against today's Linus Git code for the 3.2 kernel head.

UPDATE: More ASPM kernel patches have been published.

That's one regression now addressed, with only several more power management issues left in the kernel...

About The Author
Michael Larabel is the principal author of Phoronix.com and founded the web-site in 2004 with a focus on enriching the Linux hardware experience and being the largest web-site devoted to Linux hardware reviews, particularly for products relevant to Linux gamers and enthusiasts but also commonly reviewing servers/workstations and embedded Linux devices. Michael has written more than 10,000 articles covering the state of Linux hardware support, Linux performance, graphics hardware drivers, and other topics. Michael is also the lead developer of the Phoronix Test Suite, Phoromatic, and OpenBenchmarking.org automated testing software. He can be followed via and or contacted via .
Latest Linux Hardware Reviews
  1. NVIDIA GeForce GTX 980: The Best GPU For Linux Gamers
  2. ROCCAT LUA: A Linux-Friendly Gaming Mouse
  3. Cheetah Mounts: The Affordable Way To Put Your TV On The Wall
  4. Scythe Mugen MAX
Latest Linux Articles
  1. Preview: Radeon Gallium3D Performance For CS:GO On Linux
  2. XWayland Linux Gaming Performance With GNOME Wayland On Fedora 21
  3. EXT4/Btrfs/XFS/F2FS Benchmarks On Linux 3.17
  4. Fedora 21 Alpha First Impressions: It's Great
Latest Linux News
  1. Microsoft Announces... Windows 10 With A Start Menu
  2. Borderlands 2 Launches On Steam For Linux
  3. Debian Jessie Might Get Rid Of The kFreeBSD Port
  4. Fedora Might Try A New Scheduling Strategy For Its Releases
  5. AMD's Catalyst Working On A GLSL Shader Cache
  6. OpenMP 4.0 Offloading Is Closer For GCC 5
  7. Wayland Presentation Extension Added To Weston
  8. Intel Skylake Support Rolls Out To Mesa's DRM
  9. VA-API's Libva 1.4.0 Brings VP8 Encoding Support
  10. Operating System U Fails To Live Up To Its Goals
Latest Forum Discussions
  1. FSF Issues Statement On Shellshock Bash Vulnerability
  2. Updated and Optimized Ubuntu Free Graphics Drivers
  3. X.Org Women Outreach Program Only Turns Up Two Applicants So Far
  4. Advertisements On Phoronix
  5. NVIDIA Alerts Nouveau: They're Starting To Sign/Validate GPU Firmware Images
  6. New AMD Catalyst drivers out today
  7. Take the Steam Survey results with a grain of salt. It is flawed.
  8. AMD Wants To Know What's Wrong With Catalyst