Powerplay

[koolmanoncampus]

Besides 2d/3d stuff everyone is eagerly awaiting; does anyone know of any plan to release power scaling documentation?

[airlied]

With the atombios code we should be able to do some powerplay stuff, within some constraints..

I'll hopefully play around when I get my hands on some sort of r500 laptop.

I'm not sure how r600 works but the r500 tables in the atombios look decodeable, actually using the info to set the functionality on the card properly will be where the fun part is.

[Michael]

Bridgman has confirmed that PowerPlay information will be released later on.

[koolmanoncampus]

Dave, do you have a rough plan for the user implementation yet?

Basically I am curious whether its slated to be user control or automatic with a changed powersource(battery vs. plug).

[Dandel]

With the atombios code we should be able to do some powerplay stuff, within some constraints..

I'll hopefully play around when I get my hands on some sort of r500 laptop.

I'm not sure how r600 works but the r500 tables in the atombios look decodeable, actually using the info to set the functionality on the card properly will be where the fun part is.

powerplay should work on the r500 based desktops from what i see... just look over the closed source driver for details.

[airlied]

powerplay should work on the r500 based desktops from what i see... just look over the closed source driver for details.

I don't have fglrx installed on this box, but I don't see any powerplay tables in my bios image when I try and look for them... granted this is a cheap x1300, my x1900 is in the office..

[airlied]

Dave, do you have a rough plan for the user implementation yet?

Basically I am curious whether its slated to be user control or automatic with a changed powersource(battery vs. plug).


It'll most likely wait for randr 1.3 to happen and be provided with randr 1.3 GPU properties..

so then it would be up to the desktop environment to set things appropriately..

[koolmanoncampus]

that seems like a good way of doing it.

[Berniyh]

With the atombios code we should be able to do some powerplay stuff, within some constraints..

I'll hopefully play around when I get my hands on some sort of r500 laptop.

I'm not sure how r600 works but the r500 tables in the atombios look decodeable, actually using the info to set the functionality on the card properly will be where the fun part is.
I always wanted to try out the radeon driver, but this one (missing) feature has kept me away from it.
I've got fglrx installed currently and powerplay works fine.
If you need some testing help (or stuff like that) I might help you as much as I can. ;)

Another thing:
How is mergedfb+xinerama status with the radeon driver?
I've read a while ago on bugzilla, that it kills displays. (Which honestly I couldn't really believe ;))

My card is a X200M (aka X1150 aka whatever (there are 1000s of names for that thing, aren't there?)) or to be more precise, a RS480M (says lcpsi, or was it RS485M (says MSI)?

Bridgman has confirmed that PowerPlay information will be released later on.
What bothers me about this is, that I can fully understand, that they try to get out the stuff for the newer chips first, but also Bridgman (or was it you, I can't really remember) said, that r300/r400 support will maybe be dropped from fglrx at some time this year. When that will happen, there will be only the open source driver for those chips, so I would really appreciate, if they would ensure, that these chips at that time work very very good with the open source driver.
While for most stuff (2D and also 3D) this seems to be already true, the open source driver _is_ lacking some of the features, the chips have and it would be sad if that will still be the case when fglrx drops them.

[bridgman]

What bothers me about this is, that I can fully understand, that they try to get out the stuff for the newer chips first, but also Bridgman (or was it you, I can't really remember) said, that r300/r400 support will maybe be dropped from fglrx at some time this year. When that will happen, there will be only the open source driver for those chips, so I would really appreciate, if they would ensure, that these chips at that time work very very good with the open source driver.
While for most stuff (2D and also 3D) this seems to be already true, the open source driver _is_ lacking some of the features, the chips have and it would be sad if that will still be the case when fglrx drops them.

I think you can rest easy there.

[Berniyh]

I think you can rest easy there.
Cool, thanks. ;)