Page 2 of 6 FirstFirst 1234 ... LastLast
Results 11 to 20 of 53

Thread: Radeon Driver Power Management Has Room For Improvement

  1. #11
    Join Date
    Jun 2011
    Posts
    9

    Default

    Quote Originally Posted by RealNC View Post
    The article has a funny title. "Room for improvement". lol. Seeing the performance and power consumption charts, the only conclusion is: the open drivers totally suck ass.
    My only conclusion is "just do it, it's free software".

  2. #12
    Join Date
    Jun 2009
    Posts
    2,929

    Default

    Quote Originally Posted by joepjoosten View Post
    I had also the same experience of slow switching between profiles with dynpm, and sync in blancing period is off when in low profile.
    Dynpm does not switch between profiles.

    You have to switch between profiles manually.

    What dynpm does is it adjusts the clocks and voltages on the fly, trying to exactly match the expected workload. It does not switch between predefined profiles. This means that it's more flexible, but also far harder to implement.

  3. #13
    Join Date
    Jun 2009
    Posts
    2,929

    Default

    Quote Originally Posted by RealNC View Post
    The article has a funny title. "Room for improvement". lol. Seeing the performance and power consumption charts, the only conclusion is: the open drivers totally suck ass.
    So do it better, if you can.

  4. #14
    Join Date
    Feb 2011
    Posts
    55

    Default

    Quote Originally Posted by RealNC View Post
    The article has a funny title. "Room for improvement". lol. Seeing the performance and power consumption charts, the only conclusion is: the open drivers totally suck ass.
    Frankly, after having read the article, what I took away was that power management was actually pretty good. Performance needs (lots of) work, but that's no surprise and that's being worked on everyday.

    Also, IIRC a mesa developer said that normally there is a programmable power management module that isn't used in radeon atm, so all things considered, congrats mesa-devs

  5. #15
    Join Date
    Jul 2008
    Location
    Greece
    Posts
    3,791

    Default

    Quote Originally Posted by pingufunkybeat View Post
    So do it better, if you can.
    I can't. But the point here is, neither can the radeon devs, lol

  6. #16
    Join Date
    Jan 2010
    Posts
    105

    Default

    I've been giving the open driver an extended try for the past few days on my 5770 for the first time. My results are nearly identical to these results for the 5750 except that not only does dynpm not adjust itself properly (keeping itself never lower than the same frequencies of the mid profile) but it also creates very noticeable blinking of the screen whenever it does switch. I keep it in the mid profile mostly, which is just fine for everything short of playing games, or other significant 3d rendering. I just wanted to point out, however, that using this driver is a far better experience than catalyst for, again, everything other than gaming. I can actually scroll on chromium with it maximized with out any lagging (more than I can say for the catalyst driver with tear free on) or tearing (again more than I can say for catalyst). All video playback is perfect as well. I seriously hope someone works on getting dynpm to a useful state soon, because as soon as that works, without question I prefer the open drivers after using them a few days. The only bother I have is using the mid profile keeps me at a steady 55C, which isn't awful, but the fan noise really does kill me. Thats compared to somewhere in the mid 40's while using catalyst....Also unity and the new compiz in natty work much much better on the open stack. Using catalyst gives a much more jerky experience, plus the mouse corner lag bug in 11.5, and 11.6
    Last edited by chronniff; 06-28-2011 at 12:03 PM.

  7. #17
    Join Date
    Oct 2007
    Location
    Under the bridge
    Posts
    2,144

    Default

    Also unity and the new compiz in natty work much much better on the open stack. Using catalyst gives a much more jerky experience, plus the mouse corner lag bug in 11.5, and 11.6
    Run ccsm, click OpenGL and disable the 'vsync' option. It conflicts with the 'tearfree' option in fglrx ('tearfree' will ensure you never get any tearing).

  8. #18
    Join Date
    Jan 2010
    Posts
    105

    Default

    Quote Originally Posted by BlackStar View Post
    Run ccsm, click OpenGL and disable the 'vsync' option. It conflicts with the 'tearfree' option in fglrx ('tearfree' will ensure you never get any tearing).
    already did that, and yeah that does improve it a lot, but stuff like workspace switching with expo can still get very choppy, especially when using the curve deformation, or adding more workspaces

  9. #19
    Join Date
    Dec 2009
    Location
    Ontario
    Posts
    145

    Default

    Great article.

    I'm curious now what effect the different power levels have on the desktop? (e.g. Gnome shell). I rarely ever run 3d games so I don't need that much horsepower, just enough to composite my desktop smoothly. Curious now whether running low power mode is enough.

  10. #20
    Join Date
    Feb 2011
    Location
    France
    Posts
    197

    Default

    Quote Originally Posted by Welsh Dwarf View Post
    The other thing I find bizzar is that according to the DRM, my card doesn't have a low profile (or at least, it's the same as mid) but according to the benchmarks (comparing voltage) it's the mid profile that my card is lacking.

    Being on a laptop, these things have increased importance.

    David
    How do you see this ?

Posting Permissions

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