Yes it is best to clarify but have a look at mine below, it probably is the stock.
faugusztin I don't expect the lowest idle voltage has changed that much compared to sandy - don't get your hopes up.
From what I've read on the Interwebs it seems ivy has trouble with anything <0.9V. Maybe an effect of the new 3d transistors.
While I was doing my overclocking tests I accidentally put in too much of a negative offset causing me to go to ~0.8V at idle and I had the system freeze pretty quick.
At the same time, note Billy's power results - hardly any change at idle. Here's my results (all measured from AIDA64, not BIOS setting - which ASUS recommend):
Idle: Volts, Multiplier, power draw (AC)
0.9V, 16x, 78W
0.99V, 16x, 79W
1.024V (my stock), 16x, 79W
So, basically no difference.
The real power savings only come in load. I didn't actually look to optimise that as I went to mildly OC it instead. Here you can see some scaling up instead:
Load (IBT on very high): Volts, Multiplier, power draw (AC)
1.145V (my stock), 36x, 144W
1.176V, 42x, 157W
1.2V, 43x, 163W
1.24V, 44x, 175W
I think ASUS should consider letting us edit the multiplier voltage tables directly *per multiplier*
The offset method + load line calibration is a bit crude. It seems to work off a preset voltage table per multiplier, the offset just affects all of them equally and you rely on the LLC setting to ensure you have enough volts at the higher multipliers.
Initially my LLC setting set me at way too high (1.36!) a voltage while under load, so I have to actually run with a negative offset to counter that.
Effectively you have to "stretch" out the voltage table at both ends even though you only wanted a bit more headroom at the high multipliers (my initial offset to counter the LLC caused my idle to drop below <0.9V)
A tool like CPUgenie might let you edit each p-state voltage... but I haven't tried it on here, however I used CPUgenie on an old ASUS CULV notebook that helps me get around 10 hours of web-browsing battery time on it