Announcement

Collapse
No announcement yet.

AMD Radeon HD "Cayman" Finally Moves On With Acceleration

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

  • #11
    I just ordered the AMD 6970M for my laptop to kick out the Nvidia 460M. The main reason is to see the progress as I did with my former HD4770 from no proper rendering the desktop to having composite enabled


    With the AMD HD4770 in my desktop I was jealous of the Nvidia guys and wine support, now that I have an Nvidia I am jealous of xrandr etc. you can never win

    Comment


    • #12
      Speed is OK, fan kills my ears

      Built in and up and running, it actual needs the BARTS ucode as 6970M chip.
      I guess you can only control the GPU speed, not the memory clock or the fans speed (I used Waermeleitpaste and the other cooling block, that came with the card)?

      This is on "low" profile and I had no hanging or anything in videos, browser or something:
      Code:
      # cat /sys/kernel/debug/dri/0/radeon_pm_info 
      default engine clock: 680000 kHz
      current engine clock: 299980 kHz
      default memory clock: 900000 kHz
      current memory clock: 900000 kHz
      voltage: 950 mV
      Whatever power options I change, the fan doesn't go down...
      Code:
      radeon-pci-0100
      Adapter: PCI adapter
      temp1:        +64.5?C
      The 60 degree celsius are ok, I had the same with the Catalyst (but is was able to set the fan speed to 30%, which could be fake because it was as noisy as with the radeon driver). Overall works good, I can even play OilRush with ~7-10 fps and graphic glitches. Finally, I can kick my deskopts around between the two screen using xrandr (this is a pain with nvidia and ati binary drivers).
      Last edited by disi; 14 August 2011, 07:59 AM.

      Comment


      • #13
        The power tables in VBIOS often contain a separate entry for >1 screen where the memory clock is kept high since refreshing multiple large displays requires more memory bandwidth. Might be worth trying with a single screen, rebooting to make sure the driver picks up the appropriate table entry, to see if memory clock goes down on "low" then.

        I'm not offering this as a solution, just a way to better understand what is going on.
        Test signature

        Comment


        • #14
          Originally posted by bridgman View Post
          The power tables in VBIOS often contain a separate entry for >1 screen where the memory clock is kept high since refreshing multiple large displays requires more memory bandwidth. Might be worth trying with a single screen, rebooting to make sure the driver picks up the appropriate table entry, to see if memory clock goes down on "low" then.

          I'm not offering this as a solution, just a way to better understand what is going on.
          You are a genius

          Code:
          disi@disi-bigtop ~ $ cat /sys/kernel/debug/dri/0/radeon_pm_info 
          default engine clock: 680000 kHz
          current engine clock: 99990 kHz
          default memory clock: 900000 kHz
          current memory clock: 150000 kHz
          voltage: 900 mV
          Code:
          radeon-pci-0100
          Adapter: PCI adapter
          temp1:        +48.5?C
          Using only the laptop screen and it is quiet as a lam

          p.s. it would be cool, if one could manually control this...
          //edit: is this going to change at some stage or hard wired into the vbios?
          Last edited by disi; 14 August 2011, 09:07 AM.

          Comment


          • #15
            Originally posted by disi View Post
            You are a genius
            I just remember stuff, I think agd5f told me about it

            Originally posted by disi View Post
            //edit: is this going to change at some stage or hard wired into the vbios?
            I doubt the power tables themselves will change but over time I expect you'll see more flexibility in how the tables are interpreted. Right now the main goals are "be stable" and "don't melt the chip" and those are hard enough already.
            Last edited by bridgman; 14 August 2011, 11:17 AM.
            Test signature

            Comment

            Working...
            X