PC Processors

[Tufio]

I beg you i'm going crazy

Stock cooler (had some cheap coolermaster with the **bleep** "pin" block, was very hard to set it and it spilled paste on the sides, can this be the problem? inside is clean)
msi b450 gaming plus max
msi get force 1660 super 6gb
crucial ssd m2 250gb mx 500 2280 (windows 10 pro here) and seagate 1tb
be quiet pure power 11 500w
be quiet pure base case
hyper x  dimm fury refresh ram 3200 8x2
Pc is from early october, made it myself

Problem: mostly fast reboots, rare blue screens
dump: https://easyupload.io/gycby0
Is the WHEA_UNCORRECTABLE_ERROR event 18

Those reboots mostly always happens right after system start, by opening a video or using the search in windows, or else after a couple hours during some low activity like browsing, youtube or league of legends.

I never crashed in bios.
Never crashed during heavy gaming load or OCCT. ( bu i suppose it was jsut "stable day")

Now, after those reboots the ryzen behavior.. change and i have the core number 6 stuck like this: (also voltage stuck over 1.4 and higher temps)

This is idle but core 6 IS ALWAYS over 4mhz.
For some reason, the pc is 100% stable with this, never crashed or rebooted.
After restart it go back to normal (all cores go in idle, 35-40 idle temp) and the reboot lottery start.
I'm 100% clueless.

I updated everything, updated the bios, did 2 cicles of memtest and all OCCT stuff for 1h, 0 errors.
Windows is fully updated, i do dism/sfc often and no errors.
Bios settings are stock default, only xmp 1 enabled for 3200. (but disabling or going 2 for 3000 change nothing)

Now, most days the pc is stable, had even 18 consecutive days of no reboot until few days ago, i use it 14h/day.

Pc is recent, fresh windows pro 10 recent iso, the extra stuff at start are just mouse program, mozilla maint, realtek audio and Nvidia display.

Do pc load with something wrong?
Is cpu defective/damaged? (but still can work well for 18 days?)
Maybe the mobo or something is sending wrong voltage? How do i check? I don't have more ryzen or recent mobo around.

[mstfbsrn980]

I do not know the answer to your questions. Because I don't have an AMD component right now. I saw the error messages that you were experiencing, and I conveyed what I know will be the definitive solution even if I am not using an AMD system. That's all... 

[Tufio]

---

@mstfbsrn980 wrote:

I do not know the answer to your questions. Because I don't have an AMD component right now. I saw the error messages that you were experiencing, and I conveyed what I know will be the definitive solution even if I am not using an AMD system. That's all... 

---

At this point i'm almost sure my problem is this boost thing getting some voltage setting wrong and rebooting.

Then restarting/defaulting with a different set up, with core a single core always boosting 4200mhz, no one can explain that.
Maybe ryzen master bugged something.

[Tufio]

---

@mstfbsrn980 wrote:

---

Ok core ratio multiplier set to 36.

The voltage in auto is 1.2 from the 1.1 at 34.

Did some OCCT tests, no problems but in the power one the cpu temp got to 105°, did i lost the throttling function? The unstable stock setting never passed 95°.

Idle/browsing temp 35-39, youtube 42, 50-70 gaming, all ok.

Well, let's see in the next days.

[Tufio]

Very well, 30 days of 36 multiplier and 0 crash/reboots.

The problem is the oc boost thing.

I'm ok like that, i guess i will wait for windows 10 to get destroyed by the usual bad update, then format/update everything and try stock again.

I will update.

thank you all, gotta rep now

[Tufio]

Well, i had another reboot just now.

Same as before, no blue screen.

On Tue 23/03/2021 22:54:16 your computer crashed or a problem was reported
crash dump file: C:\Windows\LiveKernelReports\WHEA\WHEA-20210323-2254.dmp
This was probably caused by the following module: ntoskrnl.exe (nt+0x97E766)
Bugcheck code: 0x124 (0x0, 0xFFFFD20F9E65EC00, 0xBEA00000, 0x108)
Error: WHEA_UNCORRECTABLE_ERROR
file path: C:\Windows\system32\ntoskrnl.exe
product: Microsoft® Windows® Operating System
company: Microsoft Corporation
description: NT Kernel & System
Bug check description: This bug check indicates that a fatal hardware error has occurred. This bug check uses the error data that is provided by the Windows Hardware Error Architecture (WHEA).
This is likely to be caused by a hardware problem.
The crash took place in the Windows kernel. Possibly this problem is caused by another driver that cannot be identified at this time.

Last reboot was in november.

I still run ryzen with the 36 multiplier, no updates, no problems until now.

Now, is normal for a ryzen to do this rebooting a couple times a year, or i still have a problem?

[Tufio]

I have now disabled the precision boost overdrive. (it was always auto), so no more boosting over 3600.
The pc seems stable (for today), did some occt tests for power and cpu, voltage is even lower than under multiplier 36, got max 1.08 on vid single core, 1.1 on tfn and 1.125 vid effective from HWiNFO64.

Now, what to do? And is cpu truly the problem here?

My dumps: https://1drv.ms/u/s!AnQ4g2Vrik9YggYGnIhtvdc2oagC?e=4FUf87

[Tufio]

Now, what to do? And is cpu truly the problem here?

My dumps: https://1drv.ms/u/s!AnQ4g2Vrik9YggYGnIhtvdc2oagC?e=4FUf87

[Tufio]

On Tue 30/03/2021 18:16:19 your computer crashed or a problem was reported
crash dump file: C:\Windows\LiveKernelReports\WHEA\WHEA-20210330-1816.dmp
This was probably caused by the following module: ntoskrnl.exe (nt+0x97E766)
Bugcheck code: 0x124 (0x0, 0xFFFFD60504F4E840, 0xBEA00000, 0x108)
Error: WHEA_UNCORRECTABLE_ERROR
file path: C:\Windows\system32\ntoskrnl.exe
product: Microsoft® Windows® Operating System
company: Microsoft Corporation
description: NT Kernel & System
Bug check description: This bug check indicates that a fatal hardware error has occurred. This bug check uses the error data that is provided by the Windows Hardware Error Architecture (WHEA).
This is likely to be caused by a hardware problem.
The crash took place in the Windows kernel. Possibly this problem is caused by another driver that cannot be identified at this time.

The current ones are always the same, clean reboot, no blue screen, windows integrity ok.

I updated everything again, updated windows/chipset/bios, tried again at bios stock settings and reboot right away.

I have now disabled the precision boost overdrive. (it was always auto), so no more boosting over 3600.
The pc seems stable (for today), did some occt tests for power and cpu, voltage is even lower than under multiplier 36, got max 1.08 on vid single core, 1.1 on tfn and 1.125 vid effective from HWiNFO64.

Now, what to do? And is cpu truly the problem here?

My dumps: https://1drv.ms/u/s!AnQ4g2Vrik9YggYGnIhtvdc2oagC?e=4FUf87

[Tufio]

---

@mstfbsrn980 wrote:

Set the CPU core ratio to 36 and manually set the CPU core voltage with the BIOS. Try to find the correct value by increasing the CPU core voltage starting from 1.4 to 1.5. Frankly, I would try to find the correct voltage with 1.46 or 1.47 values.

---

is that for testing? At 34 my "auto" core voltage cap is only 1.1

[rainingtacco]

For monitoring use HWINFO64. You get all info you need there. Do you get any WHEA errors before restart? Have you checked event viewer in windows? Check your RAM with memtest and check your drives with a software that seeks for bad sectors[ssd have trim, but sometimes it also fails]. 

[Tufio]

---

@rainingtacco wrote:

For monitoring use HWINFO64. You get all info you need there. Do you get any WHEA errors before restart? Have you checked event viewer in windows? Check your RAM with memtest and check your drives with a software that seeks for bad sectors[ssd have trim, but sometimes it also fails]. 

---

Yes, i got those whea from event viewer, i check often.
No bluescreens.

Already did 2 cicles of memtest.

Looking for this trim.

edit

Trim is already enabled, any program i can use?
I have crystaldisk, it is normal for the ssd m2 to get very hot?

[rainingtacco]

You can use HDD scan/hd tune to test SSD. SSD can fail too, even though they have inbuilt error correction[trim] like ECC RAM. The intermittent random bsods that happen usually when system is loading points to PSU, memory or the storage[ssd] itself. I doubt its motherboard, but who knows? It would be nice if you could somehow find a debugger loader with windows, that will analyze step by step what process, and driver is called and make a log in case of crash. 

Oh and do you get Kernel Power events in event viewer? These can be PSU or memory related. 

Have you tried booting in safe mode and check if reboots/bsods persist?

[Tufio]

---

@rainingtacco wrote:

You can use HDD scan/hd tune to test SSD. SSD can fail too, even though they have inbuilt error correction[trim] like ECC RAM. The intermittent random bsods that happen usually when system is loading points to PSU, memory or the storage[ssd] itself. I doubt its motherboard, but who knows? It would be nice if you could somehow find a debugger loader with windows, that will analyze step by step what process, and driver is called and make a log in case of crash. 

Oh and do you get Kernel Power events in event viewer? These can be PSU or memory related. 

Have you tried booting in safe mode and check if reboots/bsods persist?

---

Critical unexpected shutdown kernel power event id 41? Yes, the reboots often make those.
Other kernel power are normal info, no yellows.

I don't really crash/reboot during windows load (well in october had a couple reboot little loops) and i never han any problems in bios.

At the moment i only reboot after a couple hours, during low activity. (like league of legend or browsing), but for some days or even weeks i'm fine.

I did not really used safe mode much, as i said is not like i have those reboots everyday, i can't force them to test.
Also the pc is new, i don't really have extra stuff to cause problems.

edit:

OK did hd-tune error scan and no problems.