VID In Core Temp vs. Bios

[Myrkul23]

I have a Q6600 that is overclocked to 3.0 GHz on room air, and my voltage I put it on in the BIOS is 1.35. I got to that number after a lot of Prime95 testing. Anyway, whenever I run core temp the VID is always 1.300. What is the discrepancy? And which is the real voltage? Thanks for the help

 

[Evilonigiri]

The VID shown in core temp is what the cpu's stock voltage is. So it doesn't change.

The VID shown in the bios is what you are changing. If you want an accurate reading, use CPUz or HWMonitor.

 

[Shadow703793]

^+1

VID is what voltage Intel says the CPU will need at stock speed. The one in CPUZ is the current voltage.

 

[Myrkul23]

Ok, so I checked in CPUz and it reads my voltage as 1.28 but in the BIOS it's set at 1.35. Any idea why that would be different?

 

[Lupiron]

Yup. VDrop and droop.

When you manually select a Bios VCore, it's automatically "droped" down to a lower value upon booting into windows. Thats to make sure your Processor never receives a voltage exceeding that Bios setting.

Also, when you run a powerful program, as your processor uses more currant to simply opporate at higher usages. (IE 100% of your maximum frequency.) It is reduced further, to protect against both types of spikes. Thats VDroop.

The power supply can not instantly stop sending power to the chip when it gets the signal to. "some" VCore is still being sent to the chip for that small time frame, and can cause a spike in power, because your chip is no longer using it.

However, when over clocking, the bigger the vdrop and droop, the more Bios VCore you have to give it, just to maintain a "loaded" vcore of whatever.

Do you have some sort of Load Line calibration on that board? if so, enable it.

--Lupi

 

[Myrkul23]

Ok, thanks for the info! That clears it up. One question I have though is that when I run Prime95 overnight, I get no errors in the program. The problem though is that my screen's graphics get messed up and even though I can see my mouse cursor moving I cannot click any icons or start windows explorer. The only thing I could do is run command prompt and shutdown through a manual command. Is this a stability issue?

 

[Lupiron]

It can be... since you ran it over night, any one of the windows background programs could have failed. Normally it leaves a small message saying that it failed. But windows explorer can bomb out without a peep!

Have you ever watched it happen? I mean, been sitting there when it happened?

You never mentioned anything else. Whats up with your memory? What have you changed to over clock to 3.0?

--Lupi

 

[Myrkul23]

I have my Q6600 at stock multiplier (9) with the FSB of 333 MHz and RAM running in a 1:1 ratio. I ran memtest with no errors. I haven't seen it actually happen, but the one thing that I can see is that when I look at my keyboard's LCD (I have the Logitech G15), the normal readout is replaced with a bunch of memory addresses (looks like a crash).

 

[Lupiron]

More than likely it is. Just add a notch of VCore.

Remember, when you run prime, its using all your capabilities, however, windows has to maintain a huge array of smaller programs as well, and sometimes when its at 100%, because it prioritizes the background processes weakly, system junk will bomb out. Windows explorer, Super fetch, pre fetch, desktop manager, etc.

You have a high VID chip, and they just need more voltage. The higher the starting voltage, the higher the ending voltage for a speed, in a sliding scale. As in, a 1.3000 VID like yours will need not only the difference in VID. Say, against my 1.2000 VID chip.

I get 3.6 at 1.312 volts, loaded. You'd think that you would just need 1.412 for a 1.3000 VID, but nope, you need a 3rd more, so more like 1.43 volts loaded for the same speed, and that gets worse as you up the voltage yourself.

Try it with another notch of VCore, and see if it happens. Usually if thats all that the error was was a background system file, then a single notch of VCore should be all you need!

--Lupi