Fan control for PowerMac7_3 (#3)

H. Peter Anvin hpa at zytor.com
Sat Oct 16 16:01:08 EST 2004


Benjamin Herrenschmidt wrote:
> On Sat, 2004-10-16 at 15:32, H. Peter Anvin wrote:
> 
> 
>>It's the backside fan that oscillates; backside_fan_pwm varies between 
>>30 and 100 in what is pretty much a squarewave.  See attached graph (and 
>>note how the other fans vary with workload.)
>>
>>I probably need to write a "power virus" program for the G5 to really 
>>test out the high end (a power virus is a program which keeps the chip 
>>running as hard as it can; generally keep all pipelines stuffed.)
> 
> think about also banging FPU and Altivec units then :)
> 

Those would be included in "all pipelines."  I need to learn more about 
the specifics of the G5 -- and general PowerPC stuff -- before I can 
write such a program, though.

> Since it's low oscillation point is 30, I suppose it properly detects
> U3H (can you verify that in the code, adding a printk for example in
> init_backside_state()).
> 
> I'll double check the values used for the PID in darwin

I'll do that and compile with debugging enabled, and send you a log from 
hell.

	-hpa



More information about the Linuxppc64-dev mailing list