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 Benchmarking Platform
Phoromatic Test Orchestration

How Well Is AMD CnQ Working For You On Linux?

AMD

Published on 27 October 2012 08:02 AM EDT
Written by Michael Larabel in AMD
18 Comments

One of the latest possible Linux power-related regressions I've heard about is that AMD Cool 'n' Quiet may no longer be functioning too well on Linux-based systems.

Last week in Prague I heard from a Linux developer who was saying AMD Cool 'n' Quiet was no longer working out too well on his AMD-based systems. CnQ, of course, is the AMD technology for dynamically lowering the clock frequencies and voltage when the processor isn't under load in order to lower the power consumption and heat output -- AMD's equivalent to Enhanced Intel SpeedStep Technology (EIST).

Last year when testing the AMD FX-8150 was the last time I explicitly looked at the AMD Cool 'n' Quiet performance under Linux. In concluding that article I wrote, "Cool 'n' Quiet is good for dropping the CPU clock speeds and voltage when idling or with minimal load, which lower the operating temperature and system power consumption. When the FX-8150 is being stressed, having Cool 'n' Quiet enabled was not detrimental to the system's performance."

I've also tested Cool 'n' Quiet in previous years to find similar results that it was useful for reducing the power consumption by a small but measurable amount when the CPU is under minimal load; this technology has been found in AMD CPUs going back to the Athlon 64 days.

After hearing that Cool 'n' Quiet is now less beneficial under Linux, I'm going through on some different AMD hardware to see if it has indeed regressed or is some more isolated problem specific to only some CPUs/motherboards/BIOS. So far I did run a few tests from the brand new AMD FX-8350 Vishera with its eight Piledriver-based cores when toggling Cool 'n' Quiet support from the ASUS Crosshair V Formula motherboard. The results in this article is not only to be on the look-out for problems but also to see how well the latest-generation AMD hardware is performing power-wise with CnQ support.

As shown from this graph, the minimum power consumption when going from idle to full load (of building the vanilla Linux kernel) was 91 Watts in its default configuration (Cool 'n' Quiet disabled) or 96 Watts when CnQ was disabled. Monitoring was done via a USB-based WattsUp power meter of the overall AC power consumption. So there is a difference and it seems to be working for the FX-8350 with a ~5% increase in power consumption if CnQ is not enabled.

The performance difference between the two configurations is very small -- since there is a very slight latency in the CPU needing to ramp up its clock speeds when the processor initially encounters load.

CnQ provides a slight power-savings while gaming and the CPU not being fully taxed.

CnQ doesn't really cost much in terms of lost performance due to the switching of frequency steppings.

Between idle and various different loads from the Linux desktop, the power consumption with CnQ disabled for the AMD FX-8350 was about 4% higher. So it appears that CnQ at least for the AMD Vishera/Piledriver processors are indeed working. Additional testing will be done on older AMD hardware to see if CnQ has regressed there with the latest kernel compared to older releases.

If you're using AMD hardware for some time, have you noticed any changes in the power consumption potentially concerning Cool 'n' Quiet technology?

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 Articles & Reviews
  1. Trying Out The Modern Linux Desktops With 4 Monitors + AMD/NVIDIA Graphics
  2. Turning A Basement Into A Big Linux Server Room
  3. NVIDIA's $1000+ GeForce GTX TITAN X Delivers Maximum Linux Performance
  4. OS X 10.10 vs. Ubuntu 15.04 vs. Fedora 21 Tests: Linux Sweeps The Board
  5. The New Place Where Linux Code Is Constantly Being Benchmarked
  6. 18-GPU NVIDIA/AMD Linux Comparison Of BioShock: Infinite
Latest Linux News
  1. GNOME 3.16 SDK Runtime Now Available
  2. Initial Intel Braxton Support Might Come To Linux 4.1
  3. Why KDE's KWin Doesn't Integrate Weston/QtCompositor For Wayland Support
  4. Clang Now Supports Targeting The NaCl OS
  5. PC-BSD Updates Its Lumina Desktop (v0.8.3)
  6. Fedora 22 Alpha Now Available For AArch64 & PowerPC64
  7. Systemd Developers Did NOT Fork The Linux Kernel
  8. PulseAudio 7.0 To Enable LFE Remixing By Default
  9. Features & Changes Coming For Mir 0.13
  10. How Far Valve Has Come: Three Years Ago They Needed OpenGL Linux Help
Most Viewed News This Week
  1. Introducing The Library Operating System For Linux
  2. Improved OpenCL Support For Blender's Cycles Renderer
  3. Allwinner Continues Jerking Around The Open-Source Community
  4. Open-Source Driver Fans Will Love NVIDIA's New OpenGL Demo
  5. GNOME 3.16 Released: It's Their Best Release Yet
  6. Systemd Change Allows For Stateless Systems With Tmpfs
  7. Ubuntu 15.04 Final Beta Released
  8. Nuclide: Facebook's New Unified IDE