[SOLVED] Pc rebooting AFTER heavy loads

[PaFu]

Hello everyone, I start saying that I've always used my pc pretty heavily, doing 3d renders long even 7-10 hours, without any issue. It's been a couple of weeks that my pc, right after finishing the render, shuts down and reboots, making me lose the final render. I replicated the issue with Blender and Premiere, rendering projects not heavier than usual. The strange thing is that it reboots after the finish of the heavy loads, not during, so I don't think it can be a temperature issue. The first thing I thought is that there was a problem with the psu delivering the wrong ammount of power to the system, so searching online I found OCCT, by Ocbase, to test the psu. The power test stressed the entire system, and it didn't found out any problem with the psu, but right after the finish of the 1h long stress test (2-3 seconds after), the pc rebooted, exactly like the Blender and Premiere renders. I updated windows and the Nvidia studio drivers and nothing has changed. In the last weeks I updated the ram from 2x8 to 4x8gb, so I even thought that the new ram could be the issue, so I tried rendering with just the old 16gb, but the problem persists. What could it be?
The pc has:
Psu Corsair TX750M
Nvidia gtx 1060 6gb
32gb ram ddr4 3200Hz Corsair Vengeance LPX (4x8)
Ryzen 5 3600
Mobo Msi b450 gaming plus

Thanks in advance for the help!

 

[Lutfij]

Welcome to the forums, newcomer!

The issue you speak of happened to me when I had an unstable overclock on a friend's/client's PC. Might wan to make sure you're on the latest BISO version for your platform. As for your OS, what version are you currently on?? How old is the PSU in your unit? You're on an X.M./P profile or a manual overclock with your rams? You should also see if your SSD(if the build has one) is pending any firmware updates as well.

 

[vov4ik_il]

If there is a dynamic boost (Turbo Core or whatever name they use) disable that and try...

 

[desperate PaFu]

Welcome to the forums, newcomer!

The issue you speak of happened to me when I had an unstable overclock on a friend's/client's PC. Might wan to make sure you're on the latest BISO version for your platform. As for your OS, what version are you currently on?? How old is the PSU in your unit? You're on an X.M./P profile or a manual overclock with your rams? You should also see if your SSD(if the build has one) is pending any firmware updates as well.

Hi Lutfij, thanks a lot for the answer, I'm the same PaFu as before, I'm writing this with another pc because the one I addressed befor stopped working and I was logged there and couldn't recover the user.
Anyway, a step at a time.
I'm on Windos 10 version 20H2, I bought the psu (and the entire pc) in september 2019, so it's pretty recent. I checked the ram and I'm on xmp profile. These are the info I checked yesterday, so today I wanted to check the rest, but something happened. Yesterday I had some trouble entering the bios, because pressing DEL or F12 during boot brought me to a black screen without anything, and the only thing I could do was to shut down the pc. I searched online and found an alternative way to enter the bios, pressing down ctrl+f5 during boot. This brought me to a screen saying: "system will auto reboot and enter flash mode. Do you want t enter flash mode?" And saying yes brought me to another screen saying "Select bios rom file to flash" with an x in the top right corner. Closing that window brought me to the Bios. A very uncomfortable way to enter the bios, but yesterday it worked perfectly. Today I wanted to check the Bios version and update it, but after the first screen regarding the flash mode like before and after my "yes" the system rebooted (like yesterday) but the screen went just black, without anything and nothing I could do. I left it like that for a couple of minutes hoping something would happen but nothing, so I shut down the pc. I then tried turning it back on, but now it is with this black screen (the monitors don't even wake up from standby, but don't say absent signal or anything) and I don't know what to do. I tried pressing Ctrl+F5 during boot or pressing repeatedly Del or F12, but nothing seems to change. Could it be that the pc is now always in flash mode? How could I solve it without losing all my data?
Thanks in advance

 

[vov4ik_il]

Running BIOS flash seems to be about the worst way to adjust settings.
I would try

• resetting CMOS to defaults (check the manual for procedure)
• check for POST codes (in the mobo manual too)
• contact MSI regarding warranty and RMA
• if not under warranty, writing a bios file manually with flasher or getting a new board

How could I solve it without losing all my data?

Pull the hdd power cable or the ssd out, you did not specify what is used at this system. As long as platform and chipset stay, you should be able to migrate w/o reinstall.

 

[desperate PaFu]

Running BIOS flash seems to be about the worst way to adjust settings.
I would try

• resetting CMOS to defaults (check the manual for procedure)
• check for POST codes (in the mobo manual too)
• contact MSI regarding warranty and RMA
• if not under warranty, writing a bios file manually with flasher or getting a new board

Pull the hdd power cable or the ssd out, you did not specify what is used at this system. As long as platform and chipset stay, you should be able to migrate w/o reinstall.

Thank you for the reply, do you think that updating Bios with the "flash bios" button could help before trying to reset the cmos?

 

[vov4ik_il]

Thank you for the reply, do you think that updating Bios with the "flash bios" button could help before trying to reset the cmos?

If you can get to a screen where it prompts you to. Consult the manual. It seems you got it messed up.

 

[desperate PaFu]

If you can get to a screen where it prompts you to. Consult the manual. It seems you got it messed up.

Thank you a lot, it worked! Clearing the CMOS made me possible to see and enter the BIOS! The strange thing now is that in the Boot priorities I can't select my sata ssd, on which I have the os, it just doesn't show on the list, where it shows my secondary hdd. I tried to make the boot override (in which it does show my ssd) and windows boots up perfectly and works fine, but obviously if I reboot the pc it chooses automatically the hdd and brings me to the Windows screen where the only wise thing between the options I see is to shut off the pc. (there are the things like make a recovery copy of windows or reset the pc or see advanced options) It's strange because the ssd does show in the system status and in the boot override, but not in the boot order priorities. Could it be that the hdd is in the sata1 port and the ssd is in the sata2? My pc has always booted properly with the ssd, and I never changed the ports in which the hard drives are located.
Thanks again for your patience, really appreciate it!!

EDIT: Ok I found the way to use the ssd, there was an option to choose between the two sata which one to prioritise. Tomorrow I will update the bios and check if the shut downs after heavy loads are still there. Thanks again for the help!

 

[vov4ik_il]

You would still want to turn the boost option off, otherwise it will likely behave the same way.

 

[PaFu]

Hey, sorry for the late update, so, I checked the ram with MemTest86, I checked the PSU with OCCT, flashed the latest BIOS, cleared the CMOS, checked all the driver updates and windows 10 updates, and the problem persisted. Then I decided to try one more thing, I changed the electrical socket I was connected to, done the heavy load tests and no reboots happened. So I thought that the issue was the socket, and I decided to test the pc for a couple of days plugged to the new one to be sure that was the problem. Nothing happened for three days, until now. It rebooted for no reason, I wasn't even doing heavy tasks. So I checked the event viewer and found this:
Error WHEA-Logger
Event ID: 18
A fatal hardware error has occurred.
Reported by component: Processor Core
Error Source: Machine Check Exception
Error Type: Cache Hierarchy Error
Processor ID: 13

Do you have any clue about what could it be? Thanks

Edit: I did again a little stress test and the pc rebooted after it. In the event viewer it says Event 41, Kernel Power. I really don't know what to try anymore..

 

[vov4ik_il]

I would turn XMP off and remove overclock if any.

 

[PaFu]

I would turn XMP off and remove overclock if any.

Thanks for the answer, I tried it but the issue persists.

 

[vov4ik_il]

There was a BIOS option called "Power Loading", try enabling that. It will level out power dumping when loads change. See if that helps. In your board, it might have a different name.

Edit: the option "Power Supply Idle Control" in MSI BIOS...

 

[desperate PaFu]

There was a BIOS option called "Power Loading", try enabling that. It will level out power dumping when loads change. See if that helps. In your board, it might have a different name.

Edit: the option "Power Supply Idle Control" in MSI BIOS...

Sorry for the late response, I tried both options available in that bios setting, but in both cases it reboots.

 

[vov4ik_il]

Are the CPU options in BIOS all at stock clocks and voltages?

 

[desperate PaFu]

Are the CPU options in BIOS all at stock clocks and voltages?

Yes, I haven't changed any bios setting since I've cleared the cmos a couple of weeks ago. Maybe the only thing I can do is to buy one component, try to replace it, see if the issue it's still there and in case it is, return it and try with another component. I even brought the pc to a local pc hardware shop to check if everything was alright and with their voltage and temperature tests they didn't find anything strange, even under heavy load.. The day I will discover what the heck happened with my pc will be a day to remember haha

 

[vov4ik_il]

To me it sounds like insufficient decoupling circuitry causing power dump spikes… either on the motherboard or on the video card, whichever works hard. It maybe fixed with software (gradual loading and unloading) or with hardware (swap for another board or gpu). If my assumption is right - would not be a simple fix on the hardware side.

 

[desperate PaFu]

To me it sounds like insufficient decoupling circuitry causing power dump spikes… either on the motherboard or on the video card, whichever works hard. It maybe fixed with software (gradual loading and unloading) or with hardware (swap for another board or gpu). If my assumption is right - would not be a simple fix on the hardware side.

Thank you, I'm gonna try to replace one by one each component (I actually started with the cpu, let's see how it goes) and I'll keep you updated. If it ends up being the thing you are talking about, do you know how to prevent it from happening again? Like a ups maybe or something like that, or I've been just really unlucky?

 

[vov4ik_il]

how to prevent it from happening again?

It is not an aging defect, it would be a design flaw with regards to this specific application... And it is just a hypothesis.

 

[desperate PaFu]

It is not an aging defect, it would be a design flaw with regards to this specific application... And it is just a hypothesis.

Ok so I changed the cpu and stressed the pc for 10 days today and no reboots happened after heavy loads, it seems solved. Thanks again for your support, all this situation has been quite stressful so I'm so glad it's over. Still don't know what actually caused the issue though