Vcore exceeds 1.55V limit of 0x12B microcode.

[SmartOne_2000]

Can you please explain this behavior on my less than 2-days old 14900K cpu running Intel defaults profile (Extreme) on my Asus z690 TUF gaming D4 motherboard?

Is the 1.55v the Core requested maximum (Core VID) or supplied Vcore maximum? 

This is my 3rd RMA'd cpu so I don't think its the cpu issue but a firmware/microcode issue.

Thanks!

[SmartOne_2000]

Hello Intel?

[Gessler]

I have the same issue - though in my case the Vcore hasn't gone past the 1.55V threshold after about a month of usage (I've set my HWiNFO to cycle every 1000ms), but the VID has been routinely going past the limit since Day 1.

I RMA'd my original 14700K for this same issue (though it wasn't exhibiting any apparent symptoms of degradation, it too was only a couple months old though) but the replacement chip does the same thing so I too am led to believe this could be a flaw in the microcode that they haven't addressed.

My motherboard is a Asus STRIX Z790-E Gaming WiFi II and I too am on Intel Default Settings.

I don't think there is a BIOS issue because I've reinstalled the latest BIOS with the microcode several times and it successfully completes each time. Both HWiNFO as well as Intel's own Processor Identification Utility confirm that I'm running microcode version 12B so the correct microcode is loading.

I guess at this point the only thing I can do is to continue using the CPU until & unless it degrades.

Hope Intel looks into this but it seems they've gone absolutely silent on the 13th/14th gen microcode issue since the 0x12B dropped.

Hit up this thread if you ever find a solution.

[SmartOne_2000]

I think VID is what the processor Vcore asks...and for what I understand, it can ask for the moon, if necessary. But the 0x12B microcode limits this Vcore ask to 1.55V maximum. In my case, this limit seems to have been exceeded. So, I don't know if this reading is within acceptable voltage reading error of the Nuvoton NCT6798D system monitoring chip or not.

This is my 3rd RMA chip, so I doubt its the chip itself (hopefully not) but something within the microcode not quite up to snuff yet.

[Gessler]

Hi again,

Has your CPU shown any signs of degradation in the couple weeks or so it has been since you got it? Did you RMA your two previous CPUs because you had stability problems or simply based on the VID/Vcore behaviour being out of spec?

Over the last couple weeks I've been running HWiNFO in the background all the time when my PC is on (about 8 hours a day) and what I've noticed is that the peak Vcore value I'm seeing has increased steadily by ~10mV increments from 1.518V to 1.527V, then 1.536V and today I just saw 1.545V.

The next 'increment' would probably take me over the threshold as well (though the 1.55V limit is a VID limit like you said, not a Vcore limit - and my chip, like your's, has been going past that since Day 1).

I guess what I wanted to ask was is your chip also showing an increase in the peak Vcore values over time?

It's concerning that nobody from Intel seems like they're even looking into this.

[SmartOne_2000]

No observed signs of degradation yet since the CPU is still new. The prior CPUs were RMA'd due to annoying instability problems. I never monitored the increase in the peak Vcore values, only set an alert and notification in HWInfo to inform me when that value exceeded 1.55v.

Now, I changed IA VR Limit in my bios from Auto to 1550mv and it dramatically reduced the number of alerts. No degradation in performance observed. Then changed it to 1500mV and no alerts observed and my peak Vcores are 1.505v

 The voltages I watch out for and set alerts for in HWInfo are:

1. Vcore ...peak voltage so far is 1.48v

2. VR VOUT (Vcore) ... peak voltage so far is 1.505v. I figure that last rounding error is due to the error tolerance of the voltage monitor chip for my mother board.

But you may change the IA VR Limit to values between 1500 and 1540. Others, as a precautionary measure, believe its too high a value and set their to 1400mV. There's a slight performance impact on artificial stress tests but none observed in real life.

[Gessler]

Understood. Did either of your two previous chips degrade even after running the recommended BIOS settings with the correct microcode since Day 1 or were you using them since before the 0x125 or at least 0x129 updates came out?

Also, how do I go about setting the IA VR Limit? I can see the option for it in my BIOS but it's grayed out and I'm unable to change it.

Is there a guide you followed? Does it involve deviating from Intel Default Settings profile?

[PC1997]

Hello,

IMHO. It's been my experience that running anything more than the performance profile will lead to degradation over time. The Extreme profile (although, supported with the KS SKU) is not officially supported with the K series itself. Furthermore, and in either case, it is too much power for the CPU to handle and WILL degrade over time, even with the latest microcode.

The trouble is there are two parts to the equation: the microcode issues that are supposedly now solved...AND motherboard manufacturers pumping TOO MUCH POWER through the chip - destroying itself.

If you're running the latest microcode you do not have to worry about that side of the equation anymore. If you hooked up an oscilloscope, it would go over 1.55V to some degree for a very short period of time, but it is not meaningful. You do not have to worry about that anymore,(assuming the latest BIOS is present when the CPU is new). However when you allow the wattage - 320W with the Extreme profile, that level of heat dissipation causes the internals of the CPU to basically... slowly melt itself, at a rapid rate depending on use cases. While it is not an 'unlimited' setting, it's still high enough to cause you problems.

And here's what really throws people off... the better the cooling, the faster your CPU will potentially degrade. Here's why: if you use an air cooler, you would not be able to boost as often, because predictably, you would run into TJ Max limits all day. On the other hand, if you have strong liquid cooling, your CPU will be able to boost to a higher frequency more often and for longer durations... and it's just that causing degradation IF your BIOS setting allow excessive PL1 & PL2 performance. That's causing CPUs to degrade more rapidly. Intel has clearly stated this as well. In fact, Arrow Lake is locked down for precisely this reason. Hope this helps ppl.

TL;DR: You are pumping the nitro through your CPU with 320W PL1 & PL2. All-The-Power, All-The-Time. Big Power, Big Performance, Big troubles with Raptor Lake.