On Tue, Nov 22, 2011 at 09:41:52AM +0100, Andreas Herrmann wrote: > On Fri, Nov 18, 2011 at 07:47:03PM +0100, Lucas Stach wrote: > > Hi, > > > > I'm trying to use the fam15h_power driver to measure my processors power > > consumption, but it gives me clearly wrong values. With a completely > > idle system the processor claims to be using 125261718µW, which would > > mean the processor operates at it's TDP limit. With all cores 100% > > loaded it claims to be using 130675781µW which I can't believe as this > > would mean the processor violates it's own spec. > > > > This is a desktop system with an AMD FX-8150 processor on a brand new > > Gigabyte GA-970A-UD3 board with the latest (F4) BIOS. > > > > Is this a known bug? If not I would be glad to help track this down. I > > had a quick look at the BIOS and Kernel Developer’s Guide (BKDG) for AMD > > Family 15h (#42301) but couldn't determine if the math done in the > > fam15h_power driver is correct or if we need to set some registers to > > specific values to make this work as intended. Any hints would be > > appreciated. > > FYI, I am looking into this. I've recently also seen that the values > reported in idle state seem to be bogus values as the power > consumption measured for the entire system is often less then the > value reported by the driver for CPU power consumption. > > Hope that I can provide an explanation for this oddities soon. Sorry for the late update. It's possible to get reasonable idle values by adapting the power averaging period. (If interested in details, see description of D18F5xE0 Processor TDP Running Average in "BKDG for AMD Family 15h Models 00h-0Fh Processors".) (At your own risk) you can try the following: (You have to be root to do this) # sensors fam15h_power-pci-00c4 Adapter: PCI adapter power1: 95.41 W # setpci -s 18.5 0xe0.l 00fffffe # setpci -s 18.5 0xe0.l=9 # setpci -s 18.5 0xe0.l 00552f69 # sensors fam15h_power-pci-00c4 Adapter: PCI adapter power1: 21.89 W Note that this might have implications for other users of this power information, like Advanced Platform Management Link (APML). Until this is clarified I won't come up with a driver quirk. Note that I'll soon send two updates for the driver. But they are not related to the unreasonable idle values. Regards, Andreas -- Operating | Advanced Micro Devices GmbH System | Einsteinring 24, 85609 Dornach/Aschheim, Germany Research | Geschäftsführer: Alberto Bozzo Center | Sitz: Dornach, Gemeinde Aschheim, Landkreis München (OSRC) | Registergericht München, HRB Nr. 43632, WEEE-Reg-Nr: DE 12919551 _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
