PC Processors

[Pabloz]

Hi

This PC is used for video processing and can work all night long. It looks like reboots only happens when processing videos (CPU heavy load), can reset after minutes or hours. It has been working without trouble for 7 months. No overlock.

Event viewer entry (bug check code is zero) :

• Event ID: 41 Description: The system has rebooted without cleanly shutting down first.

System Configuration: (9 months old)

• AMD Ryzen 7 5800X with Deepcool Gammaxx 400 v (now on NZXT AIO X63)
• MSI MPG B550 Edge WiFi
• 4*8GB Crucial DDR4 3600
• Nvidia Geforce GTX 1060 6GB
• C: WD Black SN850 NVMe 500Go (system drive)
• AMD RAID 0 with 2x4To WD Blue (data drive)
• Antec 650W 80 Plus Gold

Here are the unsuccessful steps I've already done :

• Connected power supply to APC UPS
• Disconnected everything but keyboard, mouse, wired network
• Ran CPU stress test for hours with no error (max temp 92° C)
• Ran GPU stress test for hours with no error (max temp 79° C)
• Tested RAM with no errors
• Removed 2 memory sticks to run only with 2x8
• Ran CHKDSK /F on both drives with no errors
• Ran SFC /scannow with no errors
• Disabled RAID
• Updated BIOS (E7C91AMS.1A0 - 05/25/2022)
  
• Updated chipset drivers
• Updated audio and video drivers
• Downgraded from Windows 11 Pro to 10 Pro
• Replaced graphic card with an AMD Radeon RX570
• Opened case to increase cooling
• Reseted BIOS to defaults settings

The only thing that helped had been to disable CPU speed boost and keep CPU on default speed (3.8 ghz and xmp on) but it’s a big slow down.

Any help will be welcome !

[BigAl01]

You've done a lot of testing / troubleshooting already.  If it's still rebooting randomly, I would look to the PSU as a possible source of the problem.  Perhaps a stronger PSU is in order for extensive CPU loads?

[Pabloz]

I just received another power supply and will test it tomorrow. I will post an update here.

[Pabloz]

New power supply (750W 80+ Gold) installed today and same thing happened again
The only thing that helps is to disable CPU speed boost and keep CPU on default speed (3.8 ghz and xmp on) but it’s a big slow down.

[TDM]

Pabloz,

The 5800X likes to set auto PBO and overlocking to fast as possible. So at this speeds it sounds heat related. I would say try resetting the X63 AIO, with a re-paste. If that isnt it watch the temps in the bios and watch what they do. My 5800X will jump up to almost 60C while in the bios before the fan speeds ramp up.

[Pabloz]

I set Windows Power mode to Balanced and also disabled turbo following these instructions (yes it also works with desktop CPU).

Letting the CPU automatically overclock when processing videos, it heats up to 88-90C (with both AIO and heatsink) while it caps to 80C when running on regular speed...

I will try another thermal paste tomorrow, just in case...

[johnnyenglish]

---

@Pabloz wrote:

Hi

This PC is used for video processing and can work all night long. It looks like reboots only happens when processing videos (CPU heavy load), can reset after minutes or hours. It has been working without trouble for 7 months. No overlock.

---

Hi, is it only with this task you said? Do you mean while rendering? Which program do you use?

If it just with that specific application, I'd suggest you check the vendor's forum if it has one.

Lets imagine for instances, it would be DaVinci Resolve, then Black Magic Design forums. It could also be nvidia drivers since some video edit applications are GPU intensive.

Good Luck

---

[Pabloz]

Good idea ! I will

[taskforce131]

same for me i had to put in on 3.8 because if was over clocking itself to 4.7 something is wrong with last bios update + windows update and the driver update all 3 of them missed this things up 

[Pabloz]

It happened with the BIOS 7C91v18 published in 20-12-2021 and worked flawlessly the first 7 months. I only updated it recently as part of the troubleshooting...

[cpurpe91]

Humor me. What are your VRM temps under load? I know the X570 Gaming Edge Wifi has terrible VRM cooling as I have one. 

[Pabloz]

Good question !

Will check tomorrow

[Sundowner]

I don't know if your situation was the same as mine, but I've had reboot issues with two different Ryzens (a 1700X and a 5600G). Often the systems would reboot at idle, but it could also happen during browsing or watching videos. It drove me mad because the PCs were great, both were fast and I loved them, but the random reboots were so hard to diagnose. No error message, useless data in the event logs, no pattern based on chip, motherboard, or heat. Just sudden reboots.

There are many others that have had similar issues and have like you tried many things like turning off overclocking, swapping power supplies, and even replacing the CPU or mobo. Some claim to get more stability from increasing CPU core voltage, others from increasing power to memory, others from turning off the power saving profiles in Windows. 

Although I can't conclusively explain what caused it, what helped me gain more stability was resetting Windows. This is where Windows basically reinstalls the OS and wipes out the apps but saves your personal files. You'll have to reinstall apps and browser settings but it did restore stability for me (and I got to keep CPU boosts and XMP enabled in BIOS!). Despite various testimonials from others that would indicate some issue between the AMD chips and power settings, my estimation is that there is something off with the chipset  drivers or video drivers, since a Windows reset seems to work so well for me. The last time I had this happen, it was shortly after I'd installed the latest Adrenaline software and updated the chipset driver. It doesn't seem to be happening to everyone, or there would be a mass revolt and we'd all go to Intel, but there seems to be a small but definite group that is seeing this issue.

If nothing else you try works, you may want to give the Windows reset a try. After the reset, be careful about what drivers you update and study the stability with each one for a while before updating the next.  

[Pabloz]

Thaks for sharing your experience.
I already downgraded from Windows 11 to 10 and it's still happening.
Today I changed the Power Supply and it happened again.

The only thing I found that improves things is to disable the speed boot and keep it at default speed.

[MADZyren]

I would start testing by connecting the computer straight to wall socket. While UPS:s are meant to stabilize electric flow, protect equipment and help keep it on during short blackouts, they can also cause problems (or be faulty)

Next I would try lowering memory speed from 3600 to 3200 while keeping latency settings the same they are now. Same memory can be stable at certain speed when CPU is running at a lower clocks speed, but become unstable when CPU is running faster or is overclocked. If this seems to fix the issue, you might want to try set memory speed back to 3600, but raise memory voltage to maybe 1.375-1.4.  I would not go higher than 1.4V on every day use.

If problem persists, you could try removing one stick of memory at a time to see if one of them is faulty. If memory is only just a bit unstable and does it only under specific circumstances, they are difficult to diagnose.

If these do not work, I would propably manually set an all-core clockspeed with manual CPU voltage. It is often thought that undervolting/underclocking CPU is bad because it reduces performance, but in practice affects performance so little, you are unlikely to notice and long, heavy processing might actually benefit as CPU runs way cooler, makes less noise and consumes less power. Currently I'm running a 3800X at 3.7GHz all-core @0.95V  and before this 4.2GHz all-core @~1.29V. With latter, in most tests my CPU did better than at stock, where it should peak at 4.5GHz.

You could also check

Windows Event Viewer -> Windows logs -> System

Windows Even Viewer -> Windows logs -> Application

to see what happened when the machine crashed. Maybe there is an error message pointing to some device of application.

[Pabloz]

Hi Thanks for helping.

I already tried to slow down memory to 3200 and tested sticks by pairs with no difference unfortunately.

Well, the event viewer is almost empty... The entry is :
Event ID: 41 Description: The system has rebooted without cleanly shutting down first. (bug check codes are always zero)

Thanks a lot for the explanations about the CPU power. I will also try that

[MADZyren · Accepted solution]

"Event ID: 41 Description: The system has rebooted without cleanly shutting down first." is recorded after the crash and reboot. If you scroll to earlier events, preceding reboot, it doesn't mention anything specific either? 

But it is starting to sound like your CPU is broken in a sense that under heavy load, it can't handle full advertized speed. I feel like warranty should cover this.

[MiNGv00]

Having a similar issue, but I am 5800x3D which is not able to tune the ratio.

Lastly, I tune the CPU nb/soc to 1.15V to fix it. Hope it can help you

[ThreeDee]

I had a similar issue with a 3700x on an ASRock X470D4U server board. My Plex jail would crash when PBO and/or CPB where enabled .. but was stable 24/7 with them disabled.

I replaced the CPU with a 3900x and now I can run PBO/CPB without issue.. yet the 3700x runs fine on another motherboard though which is what has me scratching my head still.

Regardless if it would actually help or not, I'd go for a 750wtt 80+ Gold (at least) from a reputable brand .. check reviews before purchasing anything

Which Antec model are you running?

Are you using any cable extenders inside your setup?

Have you tried your RAM on auto w/PBO/CPB enabled for testing purposes?

Have you tried running just one stick of RAM for testing purposes?

[Pabloz]

Which Antec model are you running?
Antec HCG650 Gold, 650W Fully Modular replaced now with rand new Super Flower Leadex III 750W 80+ Gold

Are you using any cable extenders inside your setup?
No, it's plugged directly into a brand new APC UPS

Have you tried your RAM on auto w/PBO/CPB enabled for testing purposes?
Yes. Tested successfully RAM with MemTest86. Also removed 2 memory sticks to run only with 2x8 and switched pairs.

Have you tried running just one stick of RAM for testing purposes?
No

[MADZyren]

"Are you using any cable extenders inside your setup?
No, it's plugged directly into a brand new APC UPS"

I think he meant inside your PC case - between PSU and motherboard and the rest.

[crystalyser]

Maybe this thread can help...?

If you are on anything newer then agesa 1203c you should turn of PBO because EDC  ....

[Pabloz]

Will check, thanks !

[Pabloz]

Should I downgrade my BIOS ?

[ThreeDee]

My daughter runs a 5800x on her x570 Steel Legend with 2 x 16GB 3200 CAS16 sticks @ 3600 CAS16 1.385v without issue .. PBO/CPB enabled with latest Beta AGESA 1.2.0.7 BIOS installed (it's actually a one newer Beta that adds better compatibility with RTX 4090's .. but she hasn't had any issues on any BIOS's). be quiet! Dark Rock Slim CPU cooler. RX 6700xt. EVGA GQ 850wtt 80+ Gold. 2 x M.2 NVMe drives and 1 x SATA HDD.

You might have a dud CPU on your hands

OR .. maybe your Raid0 array is being fickle

[MiNGv00]

Having a similar issue, but I am 5800x3D which is not able to tune the ratio.

Lastly, I tune the CPU nb/soc to 1.15V to fix it. The max voltage is 1.2V, but the auto-tuning is 0.9X which may cause the ram unstable.

Hope it can help you.