So these are the most basic settings for overclocking. And yes this is for people who have read the guide numerous times and have some idea of what they are doing.



The MOST important things to note for the upcoming 45nm is this. CPU VTT is changing to a stock of 1.1v and a maximum of 1.4v.

Other than that everything should be the same. I love the new BIOS and it's options.

The 0906 brings OFFICIAL support for 45nm for the P5K-E and a lot of fun new options. The ONLY draw back is the vDrop issue. It drops voltage as much as 0.11v from the BIOS setting. The maximum in the BIOS is 1.7v so you can still get in the high 1.5s-1.59ish.

Anyways.. Asus has a pretty simple BIOS.. I do like it.

Edit: Fixed First link.

I am bothered by the Load-line calibration.

According to Anandtech, removing the Vdroop is a bad thing:

This means that Vdroop ensures that the CPU VID isn't exceeded.

Removing it is a terrible thing, apparently.

Finally anandtech points out the load line calibration is a bad thing.

Source: http://www.anandtech.com/cpuchipse [...] i=3184&p=5

So what's what?

Considering that on this BIOS when I set the VID to 1.525 in the BIOS and I get 1.416 in Windows and a drop to 1.400 during Load, I think it will be ok.. and not spike over VID.

In addition, I've known plenty of people who pencil mod and remove vDroop all together.

Also- LoadLine Calibration is different on the P35 and X38 (I think). Previously it was known as the Voltage Damper on the P35. Either way I've never noticed it having a bad effect on my CPU or board.

So Vdroop is a bad thing?

Lol... I think it's overstated so much in the OCing world.

12-24mV is nothing to worry about.

When I was getting 48-100mV vDroop on my DFI I was worried.

vDroop is a normal and intended part of the motherboard. I use LoadLine because it is there. It was implemented by Asus engineers and said to be safe by people smarter than I at Asus.

I think people just like keeping voltages low, and the points we run at it's tough if it droops too much under load. I need 1.400 to be stable under load at 3.6 GHz. If I was having to set it at 1.45 volts I wouldn't be happy. Luckily I only have a 16mV Droop.

After all of the explanation and all, they concluded with the statement above.

Argh I'm so confused!

Without the load line calibration, you're VID would be higher? But because of the Voffset and Vdrop, the true voltage should be lower anyways right?

Nah... Load Line just eliminates some vDroop.

VID - vOffset (vDrop)= Windows Idle Voltage (WIV).
WIV - vDroop = Windows Load Voltage (WLV).

Your processor has to be stable at 100% or WLV...

I kind of look at it like this.. Ok.. VID is the max you want the processor to go to. The stock VID on my chip is 1.2875. I'm already running 1.416 so it's already over the VID.

What Load-Line and Pencil mods do is move the spike up basically. Instead of occuring from a lower voltage and spiking to the VID it starts higher and spikes over the VID. I'm not really concerned about this.. it's not spiking to 1.9v..

So let's say you set up this, assuming a 34mV Vdrop and 50mV vDroop, needing 1.400 volts to be stable at 3.6 GHz.

1) Without LL Cal: 1.400 Load and 1.450 Idle and 1.484 VID.
2) With LL Cal: 1.400 Load and 1.416 Idle and 1.45 VID.

In case 1 the voltage might spike to 1.47 volts. It never went over the VID though, right? In case two it spikes to 1.47 volts, but it passes the VID... What's worse?

Oh now I get it. Thanks bunch cnumartyr!

So basically you just don't want the VID so high, that it right?

Oh and before I meant Vdroop, not vdrop.

Kind of.. I meant it's going to take the same voltage to get something stable. The question (in my mind) is how high the voltage has to go after the Droop during idle conditions. The less the theoretical better.

It's going to spike all the same from load to unload.. but who cares? It might go over the VID, but we're already WAY over the VID..