2 of 2 people found this helpful
This is standard operational behavior when the R7 series is set to auto. As an example, Auto settings such as (CPU/SOC/Memory etc.) Load Line Calibration (LLC), Maximum Power Allotment, Power Phase Control to name a few, are being continuously adjusted on the fly to what the system judges is necessary to maintain stability. Thus a VCore 1.45 V isn't unheard of and is quite common.
Now in terms of the 1.35V you speak of, that's not the VCore, but the VID which is completely different. I'll explain the fundamental difference between VID and Vcore as seen in HWInfo64, so you understand the differences for your future reference when viewing the data.
The VID reading being reported in HWInfo64 is the value the CPU desires or should I say, what it believes it's set to. Thus 1.35V, which would be standard operating maximum voltage sans scenarios such as XFR requested voltage increases or general voltage spikes (which is also fairly regular with chips using Pstates, custom or manufacturer standard). This isn't a reliable reading.
The Vcore reading is the actual voltage the cpu is receiving at any moment based off of sensor data and is a reliable reflection of cpu core(s) V status. Thus a peak of 1.45VCore being represented.
So, in HWInfo64 you've simply been reading the incorrect data column, which happens and many make the same mistake. Scroll down past "memory timings" and directly under that, you should see the category "CPU [#0]: AMD Ryzen". Once there under the category mentioned, locate "CPU Core Voltage SVI TFN" or you can also search in the category "Insert your motherboard here/look for your motherboard" and locate "Vcore". Either of those two readings will provide a fairly accurate representation of your actual "CPU VCore" measurement as seen below.
So, with that being said, you have nothing to worry about. There are simply autonomous voltage adjustments being applied, to ensure your CPU's stability at any given moment. Your temperatures are fine, because a peak of 1.45V is likely only momentary and is simply another Pstate, like any other. If the autonomous settings determine this is the correct value for that moment, it has also determined that it is under a safe operating scenario to do so.
Thank you for the really good explanation. Just one thing: The VID in HWInfo is at 1.55v and the "real vcore" is at ~1,35v, but only when my CPU has a high load. When the load drops, I have short peaks of 1,45v. I just discoverd this and I have no idea why the vcore is getting higher when the cpu is nearly not used. I think it really doesn't matter, because the vcore peaks are very short.
Anyway, my question is answerd, thanks again for the detailed answer.
cm035 Glad to see that you have found the correct answer to your question. In order to help others with a similar problem to find this answer quickly, we have locked the discussion to keep it focused on the issue you reported.
Ryzen sucks. I can't believe they going to the market with out test ram, motherboards and even processors. It's a shame, I want really to upgrade my 4690K to R5 1600, but with this problems...
I even regret of buy my RX480.
AMD you need new people working on drivers, testing, etc.