Page 4 of 4 FirstFirst ... 234
Results 31 to 36 of 36

Thread: ATI Gets Dynamic Power Management & Profiles Too

  1. #31
    Join Date
    Dec 2008
    Posts
    57

    Default

    lol, i don't agree with you. It's all about motivation. :P My brother has dyslexia, but now he has a wellpayed job and he doesnt seem to be a robot either. Sure, our school system isn't the best, but you sound like it would the worst on planet...


    But really thats offtopic. We shouldn't talk about school when we can talk about Dynamic Power Management on open ati drivers. :P I'm really excited.

  2. #32
    Join Date
    Aug 2008
    Location
    Finland
    Posts
    1,673

    Default

    Quote Originally Posted by curaga View Post
    Top scores in that study, yet no ability to do maths without a calculator.
    Then again, things you can do with a calculator are pretty trivial anyway should they lean towards a more mathematical education later on.

  3. #33
    Join Date
    Jun 2009
    Posts
    2,932

    Default

    OK, I've finally had time to play around a bit. How do I test whether PM is working?

    I have drm-radeon-testing from today, and I've passed radeon.dynpm=1 to the kernel.

    I have the Mesa, libdrm and ddx from git.

    dmesg says:
    Code:
    [    0.536165] [drm] Internal thermal controller with fan control
    [    0.536329] [drm] radeon: power management initialized
    But it also says this (much later):

    Code:
    [  441.708267] You have old & broken userspace please consider updating mesa
    This is what happens:

    Code:
    # echo profile > /sys/class/drm/card0/device/power_method 
    # echo default > /sys/class/drm/card0/device/power_profile
    # cat /sys/class/hwmon/hwmon0/temp1_input 
    69000
    Code:
    # echo low > /sys/class/drm/card0/device/power_profile
    
    ***** wait for a while... *****
    
    # cat /sys/class/hwmon/hwmon0/temp1_input
    62000
    So, it seems to be working. It's not a huge difference, but it's a fanless card, so I guess it does run a bit hot in general.

    Code:
    # echo dynpm > /sys/class/drm/card0/device/power_method
    # cat /sys/class/hwmon/hwmon0/temp1_input
    60000
    
    # glxgears
    3203 frames in 5.0 seconds = 640.533 FPS
    3735 frames in 5.0 seconds = 746.953 FPS
    3984 frames in 5.0 seconds = 796.798 FPS
    3991 frames in 5.0 seconds = 798.084 FPS
    4217 frames in 5.0 seconds = 843.314 FPS
    3589 frames in 5.0 seconds = 717.598 FPS
    3459 frames in 5.0 seconds = 691.586 FPS
    4012 frames in 5.0 seconds = 802.205 FPS
    3649 frames in 5.0 seconds = 729.521 FPS
    3484 frames in 5.0 seconds = 696.674 FPS
    4091 frames in 5.0 seconds = 817.972 FPS
    
    *** flicker flicker flicker ***
    
    # cat /sys/class/hwmon/hwmon0/temp1_input 
    65000
    So, it seems to be working. There is flicker while reclocking, and the glxgears numbers are a bit odd, as if it keeps reclocking up and down. I know it's not a benchmark, it was easier to check than OpenArena or something.

  4. #34
    Join Date
    Jun 2009
    Posts
    2,932

    Default

    BTW, I get jerky mouse with dynpm, it goes away when I switch to profile pm.

  5. #35
    Join Date
    Jan 2007
    Location
    Germany
    Posts
    2,156

    Default

    Quote Originally Posted by pingufunkybeat View Post
    BTW, I get jerky mouse with dynpm, it goes away when I switch to profile pm.
    This a radeon pm bug? I have my mouse going away after 2-3 seconds when I don't move it using Ubuntu 10.10 "Maverick"-daily.

  6. #36
    Join Date
    Jun 2009
    Posts
    2,932

    Default

    I don't know if it's related to radeon pm, I just notice that the mouse cursor sticks at different parts of the screen when dynamic powersaving is activated (dynpm on). This does not happen when I use profiles (dynpm off). It didn't happen with the 2.6.34 kernel either. I don't know, maybe it's reclocking too often?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •