Weird Voltage Readings

[TheRyuu]

Check this out:
Voltage's

Now look at the vcore in both of the programs. One is 1.40 and the other is 1.46. In my bios I have it set to 1.40 and I don't know if my CPU might be getting more than that.

Second, I don't think its dangerous to run the CPU at that voltage right? Now also listen to this. I primed it for 6 hours on about 1.375 volts and it came out ok. However Quake 4 started to freeze so thats why the voltage is on 1.4 (just to let you know). Temps seem to be ok, only goes to around ~46c under full load.

Third, is that OC safe? I think it is, just wondering what the all knowing people at anandtech think...

 

[Markbnj]

There are several reasons why this happens. The first is that the sensors aren't 100% accurate. Often they are far from accurate at all.

The second reason, and probably more germain, is that this isn't like an analog meter, that responds continuously to variations in electrical signal strength. The sensor itself is analog. The chip it is connected to samples the sensor and writes those samples into a register. The programs reading the chip poll this register. They may just put up whatever is in it, or they may average the readings over time. There are lots of different approaches.

 

[TheRyuu]

But it is safe, right?

 

[TheRyuu]

??

 

[suszterpatt]

Don't trust software readings. They would only give you accurate readings if the voltage of the rails were dead on. Get a multimeter and see if the rails are stable.

 

[themusgrat]

In any case, even 1.5 is good. Try to not go much farther, though many do without any harm. Then again, there are horror stories.