My old B3 Q6600, CPU_Z and other VCORE readings

[BonzaiDuck]

Quickie question, you Anandtech geeks.

I built this Q6600 system back in fall, 2007. I pushed it to 3.2 Ghz, where it sat for several months. Then I just decided to back it off to the "easy" 3.0 Ghz settings.

The "setting" of the VCORE is at about 1.3125V. BIOS shows it around 1.31V. Everest Ultimate shows it at 1.31V. But my latest version of CPU_Z (Feb, 2009) shows ~1.25V at idle.

Whichizzit?

 

[nevbie]

does the everest reading vary? (under load etc) if it doesn't, i would guess CPU-Z. in any case 1.31 is 1.3125 rounded, so it sounds like the "supposed" voltage instead of actual.

 

[daw123]

I would go with BIOS VCore reading as per Gillbot's thread here.

 

[BonzaiDuck]

Well, I've read a lot here about the reliability of CPU-Z.

And -- yes -- the Everest reading varies with load. So does the CPU-Z reading.

The reason I say this: I've never really pushed the VCORE on the Q6600 as far as some benchtest reviews (overclocking with air-cooling). And I used to use the "set" value as a cautionary limit. Now, I have some more experience and less confusion.

For instance, I think I pushed my E8600 system to its ~4.3Ghz with a VCORE "set" value of 1.3625V. 1.3625V is the upper-limit of the Intel "safe-range" spec. But I think there was a lot of agreement that the CPU-Z reported voltage is more relevant for the comparison to "safe-range." At idle, the E8600 is closer to 1.335V under CPU-Z. At load, it's more like 1.29+V

I don't think I ever paid much attention to the CPU_Z reading with the Q6600 maximum overclock.

 

[Idontcare]

BonzaiDuck - I believe in CPU-Z Vcc reporting as it (and more importantly the changes in it) has always tracked well with my power-consumption data.

When I watch my kill-a-watt it will jump around slowly in a quantum fashion between two power levels at full load for example, say 206W and 210W. Simultaneously watching the Vcc reported in CPU-z shows that the Vcc was jumping between its discreet levels, say 1.3125 and 1.3500 in 100% correlation with my kill-a-watt readings.