Question The Offset voltage is changed to 0.1250V by BIOS of Ryzen. The cause of its bug and a mechanism were specified.

[KeMuケミュー]

The Offset voltage is changed to 0.1250V by BIOS of Ryzen. The cause of its bug and a mechanism were specified.

-Offset problem -
Basic : 0,0250V << Probably.
XFR: 0.0500V
TB: 0.0500V

Question) Why is Offset changed to 0.025+0.05+0,05 = 0.125V?
1) Vcore and 2) Base Offset 3) TB and 4) XFR

That's a principle of the program mistake added up:.
MOD which makes "P0 State" that.
But because the voltage is input double, 2.5V hangs.
There is a voltage limiter, so the voltage it doesn't take for that is corrected automatically.

⇒ That's the numerical value remains bug of the "power" in an equipped exchange "defense" (It's the bug work which occurs in SFC: GADYURIN.)
⇒ It's exchanged for Excalibur after an equipped exchange is removed once when it's original.
⇒ When exchanging it directly without removing a weapon, because you don't return it to the beginning, a parameter will be power 50 + weapon 166 + weapon 166 (Power increases in 380.)

In case of a TB boost program
(X) the way to change for that is impossible as a program.
(◎) It's a way right to increase after it's returned to the original numerical value.
Without returning the numerical value, TB and XFR change the numerical value just as it is, it's developed into the numerical value remains bug by the thing you (have added).

Offset is the mechanism added up.
One with an X does an X nothingness, and lay is also the trick to need the voltage wastefully 0.05 V (It's a presence of XFR.)
XFR and TB become effective at default. And it develops into Offset0.1250V problem ⇒ CPU is lack of the voltage and is a boot, and it isn't put on.
To refer to more biggest clocks of 1.40V or 1.50V for default, a fixation kill⇒ mother is dead (The voltage be reset for some reason, a bumpkin includes it, and.)

－－－－－－
Offset increases 0.125 V by a defect of a TB boost. When a PC made an error, CMOS will be default automatically. But the voltage isn't reset. A TB boost will be "ON".
When it's Ryzen 3600, the SoC voltage should be set as more than 1.05 V.

Ryzen 1700 3.5Ghz 1.0750V + Offset 0.1250 = Vcore 1.2250V
Ryzen 1600AF 3.8GHz 1.10V + Offset 0.1250 = Vcore 1.250V
Ryzen 3600 0ffset -0.10V 4.0GHz ( 1.15V+Offset 0.1250V) = Vcore 1.2750V
Ryzen 4600 Offset -0.10V 4.2Ghz ( 1.15V+Offset0.1250V) = Vcore 1.2750V
This bug is form March 2017 to 2020 Ryzen All BIOS

Post 26 May 2020

 

[KeMuケミュー]

Update : 27 May 2020

Is Ryzen's P state disabled forcibly? (Commentary)
● Verification: X370 Taichi

(Report from blog)
For the time being, "AMD Cooling function" & "C State function", so the voltage is in the state only below 1,55GHz
P2 State is AGESA prepared by AMD. Whether it is fully adapted depends on the manufacturer.

"Intel pressure"
The P2 State function can execute programs. However
Whether "Custom P State" is functioning properly depends on the manufacturer (⇒ It is implemented, but buggy.)
Offset 0.125V problem is also "Intel's pressure", XFR and TB boost were implemented, but there is a bug

<want to lower the voltage>
1: P State from 1.55GHz to 3.8GHz cannot be set. The CPU clock is forced to P0 when the power plan balance is exceeded.
2: In case of MSI-X370, you can use P2 State
3: Most motherboards can only turn "ON" in states below 1.55GHz or 2.2Ghz
4: Custom P State "AUTO" = "ON" (but not Enable) ⇒ It is correct to leave it in AUTO
5: CPU clock must be specified! ⇒ If you leave it in AUTO, the CPU will be forcibly booted at 4.0GHz to 4.3GHz and the PC will not boot.
6: For voltage, "Do not specify voltage for MSI" or "There is a risk that the mother will die only in MSI. On that assumption, specify the CPU clock and then input + voltage"

<Is the CPU clock actually going down? >
There is a strong possibility that the CPU clock has dropped internally to 2.2GHz.
If you want to use a voltage less than 1.55 Ghz, "only when you sign out of Windows 10" it becomes valid.
CPU clock dropped to 0.9GHz on MSI motherboard for that motherboard which cannot set voltage

It is a benchmark of 7Zip.
"Only for power saving plans" State function was used
(The score will be 2.2 Ghz.)

When the power plan is used more than the balance ⇒ PC becomes compulsory P0 State

(Q) I bought the 1600AF for some reason. Where is the voltage setting? ..

⇒ Spec guarantee "It is the same voltage as Ryzen 1700 of 14 nm"
First is 3.5Ghz-1.2375V (Offset "-0.10V").
⇒ Perform voltage survey on Windows 10! !
⇒ This is not ZEN +. MAX is about 4.0GHz, 4.1Ghz
With ZEN +, it is 1.10V (Offset 0.000V) at 3.8GHz.
The 1600AF has a higher voltage of 3.6GHz than ZEN +.
The voltage of the Ryzen 1600AF increases significantly from 3.7GHz.
When the CPU becomes 3.8ghz, "Vcore 1.30V Offset -0.10V" is a standard.
The voltage of 3.7GHz is done with Windows 10.
You use the MSI After burner and Ryzen Master. The model that cannot reset the voltage must not be done on the motherboard side.

4GHz-1.275V is NG ⇒ 1600AF fits that. ZEN1 requires 1.425V or more.
What about 4.1GHz? ⇒ That is a false report. Not in the first place. It requires a voltage higher than ZEN + at 3.8 GHz.
(When written in English, the tone, the wording when writing a lie, and the atmosphere are indistinguishable in a foreign language.)

 

[therealmongo]

Friend,

people are probably not answering your forum post because the translation (I am assuming Google translate) is not easily readable, expressions are not making very much sense etc etc

 

[VirtualLarry]

Note that I'm pretty sure that the voltage tables are per individual CPU, and just because your 1600AF requires Zen-like voltages for 4Ghz OCs, doesn't mean that it's not 12nm fabrication. There's a reason that it wasn't binned as a 2600(X), and that reason, is probably because it needed a higher voltage.

Don't confuse that with the idea that 1600AF is a "lie" that it's Zen+, that it's somehow really Zen (1). I don't think 14nm is still made anymore at GF. Certainly not in qty., for Ryzen (Zen 1) CPUs.

PS. Zen1 (my original 1600AE CPU), did have a bug, in some boards, that attempting to OC to 4Ghz, and boosting voltage, would sometimes result in CPU clocking around 2.2Ghz, and voltage being bumped back down. Maybe that's the mechanism that you identified? Clearly, it seemed like a bug in the AGESA or whatever, but perhaps it was also a fail-safe to prevent the CPU from burning.

Edit: If you want more-or-less steady all-core OCs of 4Ghz, get a Zen2 Ryzen R5 3600 CPU, and run it at stock speeds.

Edit: I don't have any personal experience OCing Zen+ CPUs (2600, 2600X, 1600AF), so I can't comment directly on those. But I tried OCing Zen 1 (1600AE), and couldn't really get it much higher than 3.9Ghz @ 1.400V or 1.425V, when I tried for 4Ghz, it didn't work so well.