So I have a 5800X that I bought about 5 months ago and it seems to always crash my PC and restart when I have low load on the CPU. Now I have tried changing out the Memory, GPU, PSU and then took the CPU and then tried it on a different board and had the same problem every single time. Whenever I put a low constant load on it it will crash. For example when I try beatsaber or Ys or any game/program that puts light constant load on the CPU it always crashes. I have tried turning PBO on and off and increasing the voltage offset(did not help) and LLC(helps a little bit) Increasing Voltage with a positive PBO offset also does not help(increased temps though)
Any ideas on what can help because I'm getting pretty tired of this as I cannot even play the new Ys game for 30 minutes without it crashing but if I play cyberpunk it will not crash.
Hunch, it sounds like your processor, or one of it's cores is starved for current. But we need more info.
Using HWMonitor64: What is the max temp of the CPU? Memory sticks, VRM? Chipset? GPU?
What type of Case do you have, ITX or ATX?
Does this happen more when you go from a High temp, and then cool down quickly? Or just idle for a while at low temps? The reason I ask, is that a lot of people think that validating their processor with Prime95 is the right way to go. However what they fail to realize is that a lot of errors are caught when the processor first cools down.
Why? Consider a processor that is cool and idling sipping current fine. User pushes many threads of Prime95 and all is well. However when the CPU goes back to idle, it tends to crash. Why? Well, the CPU could sip current at idle when it was cold, but that same CPU can not sip current at idle when it is Hot. You need more voltage to push through a hot chip than a cold chip.
Next step: Information gathering; Bring up event viewer. You want to bring up Administrator events. You will be looking for WHEA errors. Sometime you get more than one for a Crash. You want to examine each. Normally it will identify the "APIC" that caused the problem.
Go to CPU-Z. Click the "About Tab" Click Save the report .TXT. Examine the report for APIC's. Each APIC will be associated with a given CPU. Now look at all failing events. What you are trying to determine is if it is only one CPU failing or Multiple CPU's failing. Remember two different APIC ids will share the same CPU (2 threads in a Core)
Good Luck, What you see should determine your next steps.
Based on what you find, I would have two different approaches.
If only one core is failing, it could be a problem with only that core. However, there might be some doubt that that core is the one that is definitely at fault if that core is your most favored core as specified in Ryzen Master, or HWMonitor64 summary CPPC core list. It could be that that just happens to be the first dispatched and the most likely the victim of another CPU setting. (Playing with Positive core displacement or a slight positive VCore displacent would probably be the right way to proceed)
Maybe nicer, is if many different cores are failing. That would leave me to believe, or have a hunch, that there might be a problem with either the Memory system (Necessitating a slight increase in Dimm voltage) or a problem with the Internal Memory controller. This would probably suggest increasing VSoc a bit.
Anyway, Good Luck. Let me know how it comes out.
I had the same issue and you know what fixed problem? Turning off hard drive hibernation and all power saving options in power plan. Right click on windows logo, power options, select balanced or high performance, advanced options. Now in every tab turn off all power saving options (turn on performance or disable) and most important, ok hard drive tab, turn off sleep after time.
I had fresh windows install, just like hardware, to make sure windows won't be a problem... But it still was a problem.
Mostly my PC was ceasing in desktop, or while I was leaving it on while just playing yt music. Oddly enough, blue screen code was indicating hardware failure related to CPU / memory. Also, try to turn off memory oc profile. In my work I had problem with bs and it stopped when I turned off docp
you need to tweak the pbo curve and positive the gold selected core cause of low load it boost the gold core too high ... this can be fix by own manual set the core (loss single core performance for air cool but very powerful for aio) and voltage vcore or tweak the gold core curve to be positive .... (positive it boost lower speed)
board boosting too high is the cause for my 5800x ...
for the temp it can be fix by changing the ppt 120 tdc 82 edc 110 (my current setting) so we limit the cpu temp without loosing the performance...with air cool........ you can tweak own setting and find the best speed and temp by running cinebench....
use the corecycler to run 24 hours or more to detect any error.... hope this help