Announcement

Collapse
No announcement yet.

ATI Kernel Power Management Moves A Bit More

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • ATI Kernel Power Management Moves A Bit More

    Phoronix: ATI Kernel Power Management Moves A Bit More

    On Sunday we reported that ATI in-kernel power management was moving along after AMD's Alex Deucher spent some time in recent days building upon RafaĆ? MiĆ?ecki's initial power management support. Alex's patches added GUI idle IRQ support, support for changing the GPU clocks when the engine is idle, support for turning down the number of active SIMDS when running in a lower power stage for the ATI R600 ASICs and later, and new ASIC specific callback functions. This morning though he's taken the support a bit further...

    http://www.phoronix.com/vr.php?view=ODA3Mw

  • #2
    As soon as this gets going i'm going to jump ship from the ATI blob.

    Comment


    • #3
      I'm sorry for the total n00bishness, but does 'agd5f' stand for Ati Graphics Department (office/cubicle) 5F?

      *ducks and runs*

      Comment


      • #4
        Nope, but good idea. Alex was agd5f long before ATI/AMD came along

        Comment


        • #5
          im going crazy xD
          cant wait for all this to work!
          thanks alex and keep up the good work!

          Comment


          • #6
            To leave a comment on agd5f's blog, you need to log in, but it isn't possible to register there...

            Comment


            • #7
              Wow, this is just awesome, Alex, thank you so much.

              Comment


              • #8
                btw
                do you guys expect the oss powermanagement to be as good as the fglrx powermanagement?

                Comment


                • #9
                  It's hard to say. We're trying to make sure that the open source drivers are able to do all the same power-saving things as fglrx (as we discover them ) but it's going to be a bit like 3D -- it may not be worth adding all of the complexity if you can get 80% of the benefit with 20% of the work.

                  On the other hand I wouldn't be surprised if some scenarios specific to consumer users were handled *better* than fglrx so that the overall results were considered superior. Power management is one of the areas where the environment differs widely across the OSes, so a Linux-specific open source solution could also have an edge by not having to be based on common, OS-independent code.

                  I think it's safe to say it will be "pretty good" anyways

                  Comment


                  • #10
                    Originally posted by bridgman View Post
                    It's hard to say. We're trying to make sure that the open source drivers are able to do all the same power-saving things as fglrx (as we discover them ) but it's going to be a bit like 3D -- it may not be worth adding all of the complexity if you can get 80% of the benefit
                    with 20% of the work.

                    On the other hand I wouldn't be surprised if some scenarios specific to consumer users were handled *better* than fglrx so that the overall results were considered superior. Power management is one of the areas where the environment differs widely across the OSes, so a Linux-specific open source solution could also have an edge by not having to be based on common, OS-independent code.

                    I think it's safe to say it will be "pretty good" anyways
                    sounds good!
                    since im using an ati chip on my notebook.
                    today i checked how long my notebook would last with lucid + open source drivers and had a drop from 4hours ith fglrx to maybe 1hour and 30 minutes.
                    plus, the notebook fan annoyed the shit out of me

                    Comment


                    • #11
                      My recollection was that dynpm was turned off by default in the drm still, so you wouldn't have seen much in the way of power savings there. Not sure if the Lucid (2.6.33) drm is ready to have dynpm turned on or whether you need newer (ie this week) code for good results.

                      Comment


                      • #12
                        Originally posted by bridgman View Post
                        My recollection was that dynpm was turned off by default in the drm still, so you wouldn't have seen much in the way of power savings there. Not sure if the Lucid (2.6.33) drm is ready to have dynpm turned on or whether you need newer (ie this week) code for good results.
                        since im playing heroes of newerth quite frequently the oss drivers are unfortunately not yet a viable option for me.

                        Comment


                        • #13
                          Originally posted by phoronix
                          With Alex's brand new patch, hwmon drivers can now be loaded for the I2C thermal chips found on many ATI graphics cards. In other words, we finally have thermal monitoring and fan speed support for the open-source ATI stack!
                          Correct me if I'm wrong. But as I understand it I2C is only used by older cards. For the newer cards, Evergreen included, all the stuff is internal, and ATI has yet to release any documentation or code about that.

                          Comment


                          • #14
                            Is there a way to use that with kernel 2.6.33?

                            Comment


                            • #15
                              Originally posted by monraaf View Post
                              Correct me if I'm wrong. But as I understand it I2C is only used by older cards. For the newer cards, Evergreen included, all the stuff is internal, and ATI has yet to release any documentation or code about that.
                              Pretty much, although the transition seems to be fuzzier than that. Internal temp/fan logic was added midway through the 6xx generation but board partners didn't start using it immediately. AFAIK the situation is more or less :

                              - all pre-6xx boards use external chips
                              - "some" 6xx boards used on-chip controllers
                              - "most" 7xx and Evergreen boards use on-chip controllers

                              Dual-GPU boards are interesting in the sense that most of them share a cooling solution between the GPUs, so the fan speeds need to be a function of multiple temperature sensors. In those cases you again see an external fan/temp controller, with sensor diode inputs from both GPUs being combined to drive one or more fans at the same speed.

                              Comment

                              Working...
                              X