Idle CoreParking vCore // 1700x+X370TaiChi

The C-States that I can see reported in windows only show C1, C2, and C3, of which 0% time is spent in C3.
I don't know of any 3rd party tools that will correctly show which C-state a given core is in.

This was the same for Linux Mint and a CLI power tool I downloaded. I can't recall the tool, but it showed C0, C1, and C2 states only.

I would hazard to guess that Windows completely looses the plot and puts the Power Plan to "Ryzen Balanced" and refuses to let me out of that. When I switch to Ryzen Balanced, the behaviour and voltages are VERY similar to when the system does this strange thing by itself.
SV12 Core voltages don't drop below 1.187v, nor does the VRM voltage, but vCore from the MoBo still goes down to ~0.5v ish.

Tried CMOS resets, no fix.

I think it's behaving in Linux, but the lack of decent tools for detecting Ryzen voltages makes it hard. Checking Core clocks and P and C-state seems to indicate normal behaviour though.

I am not beyond ruling this to being a Windows mess up somewhere, but it does seem strange that after a reboot it does not happen within the time frame until I turn off at night.

It's as if lower P states are not being properly entered. Damn annoying and without checking in on ParkControl or constantly looking at TaskManager, one wouldn't know that the CPU is only dropping to around 1.2v at idle.
IS that a big deal? No, probably not, perfectly safe to run 24/7 under 1.3v. But still it's basically not working as it should do.

Would have come back earlier, but I've been busy messing with an Amazon Fire tablet. Getting root, removing the junk-ware, setting up iptables rules etc. As much as can be done without a full TWRP/MAgisk switch. Too many little projects!

I've disabled the Rightmark PPMP Service which is a service installed when the Rightmark Power Profile editor tool is used. I am not sure of it's impact on the system and why it needs to be running as a service - I don't think I am using any features of that side of things, except for deep editing and creating a power saving version of Ryzen Balanced Power Profile.

We'll see how this plays out - couldn't test that long today as a BSOD caused a reboot dxgmms2 - something to do with DX graphics? - whatever.

As far as testing in Linux is concerned, everything looks normal but I don't have enough insight into sensors to tell for sure.
We get the Nuvoton sensor reporting, which is what we see in BIOS and most GUI tools in Windows. That wasn't hard to get working.
However, SV12 TFN on the CPU and the VRM OUT from the IR35204 chip are not reported in anything except for HWiNFO in Windows.
Watching the C-state in Linux, most time is spent in C2, which is what I would expect, idle frequency is reported ~2GHz as expected.

I'll continue to test in Linux and in Windows (with PPMP disabled).

@cx5 Please don't take this the wrong way. But at this point I have no real interest in making anything run beyond spec.d parameters. It's probably easy and a 'free' overclock, but it's not something I am too fussed about at all for the time being.

Back to my original post:
Nope, setting the BCLK to a locked 100MHz does nothing to affect a change in this behaviour. It seems to stop it wandering so much when watching HWiNFO (is it correctly reported? Why would the BCLK not stick rigidly to 100MHz?), but doesn't cure my issue.
I have updated the chipset drivers, again no change.

Short of a CMOS reset, I am not sure why this is happening?

ZoneAlarm firewall/anti-virus is a known entity which does cause some strange behaviour. Could it be this that might be the culprit?
I wouldn't bother so much with a 3rd part firewall, except that the alerts are rather informative with ZA and non-existent with Windows own protections. I prefer to be alerted about every small thing happening rather than an overall program accessing the internet. ZA informs me of registry access as well, which is handy to stop startup programs being added.

But why would a Windows restart cure the issue so swiftly, and for it not to happen again after the first reboot?

I am very tempted to boot into Linux Mint Cinnamon 19.x from USB stick to see what happens on the next cold boot. A cleaner system which is up to date should give me some better indications of what is happening.
Sure, no Windows Core Parking, bit I might be able to read voltages with the applicable tools accessed.

VUMeter wrote: @cx5 Hi thanks for your suggestions, but I have no desire to overclock my system at present. I intend for this system to last about 10 years, as that is my usual upgrade cycle.

@VUMeter, Hey, it's the same for me. 10 years life from a PC. Now, if only I can resists to play with it too much.

Actually I also don't OC CPU and RAM, exactly as you say, having enough.

Just trying to match the bought spec, i.e. 3000mhz ram in my case, should run 3000 or 2933....

But no harming playing once in a while. 10 years is a long time......