ATI Kernel Power Management Moves A Bit More

[pingufunkybeat]

What does the ForceLowPowerMode option do? After enabling it I still see with rovclock that the core and memory clocks are at their maximum values...

With KMS, it doesn't do anything.

With UMS, it halves the clock speeds, I think.

[Louise]

There are definitely differences from generation to generation but I *think* the main breakpoints were in the middle of the 6xx generation and then more changes between 6xx and 7xx. We'll know more in a month or so, but AFAIK Evergreen should be able to leverage most of the 7xx code.

Remember, there is no R800, just Evergreen

Ok, thanks.

[yotambien]

With KMS, it doesn't do anything.

With UMS, it halves the clock speeds, I think.

That's a weird one. As I said, rovclock still reports the clocks at full whack.

[agd5f]

Last time I looked, rovclock did not take into account the post divider. Also, rovclock only works on legacy radeons, (r1xx-r4xx).

[yotambien]

I probably should have posted in another thread to avoid confusion: my card is an R300. I just thought ForceLowPowerMode also worked for legacy cards. Should owners of these cards keep using rovclock then?

[agd5f]

I probably should have posted in another thread to avoid confusion: my card is an R300. I just thought ForceLowPowerMode also worked for legacy cards. Should owners of these cards keep using rovclock then?

ForceLowPowerMode works on all asics. rovclock should be avoided.

[yotambien]

ForceLowPowerMode works on all asics. rovclock should be avoided.

So rovclock is reporting the wrong numbers? If it's not, why is the card at its highest frequency settings with the ForceLowPowerMode option activated and why should this tool be avoided?

[agd5f]

So rovclock is reporting the wrong numbers? If it's not, why is the card at its highest frequency settings with the ForceLowPowerMode option activated and why should this tool be avoided?

Last time I looked at rovclock it did not take into account the post divider for the engine or memory clocks so if the post divider is anything other an 1, it will set/report the wrong value. Also, IIRC, some of the vram setup stuff in the memory reclocking code is wrong for r2xx+ chips.

[agd5f]

Stupid $^$^&% edit limit. If the post divider is initially set to 1 and you half the engine clock, the post divider will likely be 2. since rovclock does look at the post divider, it will still report the old clock.

[yotambien]

I can't say I fully understood you, but I take it that there are some technical drawbacks with using rovclock. The point is that if ForceLowPowerMode is just halving the clocks, that's a bit conservative, isn't it? This is, I use rovclock to set the core and memory frequencies to the values reported by fglrx in its lowest power settings, and those are 1/4 and 1/3 their original values, respectively.