This is going to be a little long, I apologize. Since the exploding Ryzen CPU issue popped up recently, I've been watching my voltages like a hawk. I've noticed that every once in a while (perhaps once a day?) there is a momentary spike in three values: VDDCR_SOC, VDDCR_VDD and CPU Die (average). They all happen at exactly the same time, and tend to exceed safe values. The temp doesn't always (I've seen it hit 110 C, but sometimes it's only 70 or 80). The safe SOC voltage is 1.3 V, and it always goes well beyond that, generally 1.5, 1.6, 1.8 or even 2+. A key finding that I think is suspicious is that the average die temp spikes, but no other temperature value in the whole computer does. I suppose it could be the memory controller which maybe doesn't have a dedicated sensor, so it only shows up in the average. I'm deeply alarmed, and I'm hoping someone can help. If it was one value, I might not be so concerned, but considering all three are spiking at precisely the same time, I can't imagine it's not a real problem.
I have wondered if this is a sensor glitch, or a quirk of the way data is parsed by HWinfo, or if it's a real event that's going to damage things. I thought that turning on DOCP/EXPO profiles might be causing it, but it just happened and I have DOCP turned off. I'm having a tough time establishing a pattern because it's so infrequent. I don't think I can use logging to get more detail, because I'd have to be logging for many hours or days. I'm going to try turning off PBO next, and waiting until I see it. I'm hoping this isn't a problem with AGESA, the Asus BIOS or a physical voltage regulator.
I've been told not to worry about it, that if it was real, it would immediately destroy the CPU. I'm not convinced, so I'm continuing my efforts to figure things out. I have alerts set
on these values, so that'll hopefully help me. I have noticed one thing: the last two times it happened, it seemed to be responding to user input. The first time, I was in another application (maybe watching Netflix), and I switched to HWinfo, which immediately showed the spikes. The second time was just a few minutes ago, when I clicked on my remote desktop (I'm accessing my machine from work), as soon as clicked on the window the spikes/alerts appeared. Could it be that the processor is over-responding to a sudden user input, and is ramping up voltages too fast in an effort to respond?
I could use any help anyone is willing to provide, but I think I have two major questions initially:
1. Is this likely to be a real event, or is it HWinfo just displaying erroneous values?
2. If it is real, is it likely to cause damage, or is it ok because it's so short-lived?
Thanks very much in advance for any feedback.
