[ltp] radeonfb swaps core and memory clock?

[Alex Deucher]

On 12/9/05, Paul RIVIER <paul.rivier@eleve.emn.fr> wrote:
> obi wrote:
>
> >Hello,
> >
> >I just noticed something weird on my T41 (radeon M9). radeonfb reports
> >
> >kernel: radeonfb: Reference=3D27.00 MHz (RefDiv=3D12) Memory=3D252.00 Mh=
z, System=3D200.00 MHz
> >
> >but both rovclock (Core: 252.0 MHz, Mem: 200.25 MHz) and aticonfig (when
> >I used fglrx, I don't use them anylonger) reported the opposite: core 25=
2
> >and memory 200.
> >
> >Is it only on my machine? Is it a bug? I tried to trace it down in the
> >kernel, but my ability to read kernel code is close to NULL ....
> >
> >graziano
> >
> >
> Hi,
> I do not know, but that reminds me something I wonder :
> Does the radeonfb driver handle powerconsumption management facilities
> in ati recent cards ?
> Like ati/X handles it throught DynamicClock for exemple.
> Actually I work a lot in emacs without X but I can not notice a big
> difference in the power consumption,
> while the load on the GPU should be much lower.
>
> Does someone know something about that ?
>

radeonfb does (or did at least) have some dynamicclocks code, but it
might only be enabled on ppc.  in fact that's where I initially got
the code to add dynamicclocks to X (ati later gave me some improved
code).

Alex

> Thanks
>
>
> Paul