mumak

Could you please provide 2 more SMBus dumps at different 3rd fan speeds and note what was the fan speed (%) at the time you created each of them? That should help to figure out how the 3rd fan can be read.

Could you please also make a dump when fans are running at max speed?

As I said before, it's a bug in NVIDIA drivers, so NVIDIA must fix it.

AFAIK this was a problem of recent NVIDIA GeForce drivers (r397).

This can happen if SVID is disabled in BIOS - in such case CPU power monitoring cannot work. Please check the BIOS settings.

Some older HP Envy series (based on Quanta ODM design) used to have fan speed at EC [0xB2-0xB3], [0xD2-0xD3] (FAN_FEEDLo / FAN_FEEDHi). Not sure if that might apply here, but it's rather unlikely as the above models were notebooks using ENE ECs. An Embedded Controller dump (with known high and low fan speeds) might shed some light... And maybe a DSDT dump too...

It's the highest temperature among all diodes on the CPU (cores, GT).

I believe not all voltages you see in E-LEET can really be monitored. That screen is used to set values only, for values really monitored check the "Monitoring" tab.

Please post the ISA and SMBus sensor dumps.

It still might be a B3 chipset. There's a feature in Intel chipsets called Chipset Compatibility Revision ID. If this feature is activated by the BIOS, then the chipset reports a Compatible Revision ID (which is not the true hardware revision of the chip). On the 6-series chipsets this would cause all the identifiers on a B3 chipset to report a B2 value ! Unfortunately in this case there is NO UNIVERSAL WAY how to determine the true Hardware Revision ID - the only possible method would be if the BIOS saves the original ID in some register, then activates this feature and offers the reading of this ID to its own tools (which might be the case of the Acer tool). When this issue came out some vendors have released BIOS updates which obviously do not enable this feature anymore and so allow a proper reading of the ID - the result was that machines with a B3 chipset have been reporting B2 in all tools unless the user performed such BIOS update.

Thanks. Note for Fiery: VIN1 = DIMM VIN2 = NB VIN3 = +3.3V (*2)

Please post a photo of the BIOS System Health screen, or write down the values it displays.

Indeed that could be the issue. The BIOS is unable to inform ME of POST end and that can cause several issues. Possible solution: BIOS upgrade (with ME) or RMA

I don't see a relation between HPET or C-States and a badly written driver which spends more time at IRQL >= DPC_LEVEL than allowed.

The question is whether they would be willing to release that info. If yes, an NDA would solve the confidentiality concern. It might be worth a try..