cancel
Showing results for 
Search instead for 
Did you mean: 

Processors

Faultz
Adept I

3700x unexpected shutdown during stress test.

My PC's current specs.

CPU:3700x (Stock cooler)

Mother board: Aorus b550m pro

Ram:16Gigs of G Skill trident neo running in XMP

GPU: RX580

PSU:NZXT 750w 80+ gold

My issue is after trying everything I can think of from bios updates, To driver updates. and a few other things I can't remember that I did. After trying to stress test the CPU using either cinebench or furmarks CPU burner my pc shuts down after mere seconds after hitting run. maybe a minute or 2 with cinebench but that its. I've already repasted the CPU with higher quality thermal paste to no change at all. And whilst using Ryzen master the temps don't even say they get over 75c at any point. on top of that, I can't replicate the affect with AMD's own "Stress test" in Ryzen Master. Its died on me whilst trying to launch a game as well, but once I'm playing the games it doesn't seem affected. IF any one can help me id be greatly appreciate it.

One last thing, whilst trying the research this people were talking about it doing it idle. But i don't leave my pc on long enough to know if it does that too. and Half of the stress tests were done with XMP off and the ram running at its default 2600mhz with no change in characteristics 

 

0 Likes
9 Replies

Make sure the CPU Cooler is installed correctly. It should be tight, even, and making good contact with the CPU surface.

Remove the side panel to the computer case and see if the temperature lowers. IF it does that indicates poor air circulation.

Also make sure nothing is blocking or interrupting the CPU fan's air flow nor any cable or wires are hitting the fan blades.

Download and run OCCT CPU stress test.

The first CPU Stress test use large Packet then the second CPU Stress test use Medium Packet and finally with the third CPU Stress test use Small Packet.

Check the Temperatures and PSU Outputs during the tests. You should also run the PSU test afterwards.

The 3700X Maximum Operating Temperature is 95C. After 95C it starts to throttle and slow down. If it can't reduce the temperature it will automatically shut down the computer to prevent damage.

Set OCCT Global Overheat Temperature to 96C in Settings. That way when the CPU reaches 96C OCCT will automatically stop the test thus preventing a shutdown of the computer during the testing.

I also have a 3700X in which I have a much more powerful CPU Cooler then the stock CPU Cooler that came bundled with the CPU.

I have a CoolerMaster Hyper212 EVO (150 TDP Rated) with two fans of 1950 RPM.

When I ran OCCT CPU Stress test with both Large and Medium Packets my CPU didn't overheat. It got hot but never overheated. But the moment I ran OCCT CPU Stress test with Small Packets the test would automatically stop within seconds due to overheating. I had set OCCT Temperature in Settings to 96C.

I replaced one of the 1950 RPM CPU fans (Pull) with a high CFM FAN (200 CFM). Now I can run OCCT CPU Stress with Small packets without overheating. The temperature doesn't go above 70C..

OCCT-Screenshot-20201220-001031.png

I have Run every test with no issue's to report, Whilst the CPU is getting toasty showing a max of just over 80c, I never experienced a unexpected shut down like I do when I run Cinebench or furmarks CPU burner which shut my PC almost instantly. The screen shot is after testing the PSU. Which i never was expecting to be the issue since it will shut itself down whilst just running the cpu tests. 

0 Likes

Wish I could delete that. But it appears I can't, so i hope you see this before you reply. I ran the psu test once more after I posted that and now it is powering off consistently in less the 3 minutes of me starting the test, same story for when I run the small packet test. But the Temps don't ever go higher then say 82 which is nowhere near the set 96°c critical limit which should disengage the test all together.Taken right before Sudden shutdown,Taken right before Sudden shutdown,

 

0 Likes

I see that you have a RX580. In Wattman set the Power setting to 50% or max and see if that helps with the shut downs.

@pokestercan help you configure your RX580 to be more stable.

Obviously your CPU is not overheating but something else is causing the computer to shutdown.

Generally if a computer shuts down by itself it is either Power or Heating issue or incompatible hardware or defective hardware. Sometimes it might even be a driver issue.

What are the PSU Outputs during the test showing in OCCT?

I have heard in previous Thread that NZXT Cam, the monitoring software can cause those types of issues. Are you using CAM by any chance?

It's a NZXT C750 80+ gold (750 watt), which is way more then enough to cover both my 3700x and my rx580 when running at full tilt. I bought it specifically for upgradablitly since I won't be running the 580 for hopefully much longer once gpu avaliablitly is once again a thing. As for CAM. I have no idea. I shouldn't be. But I idk where to look if I do. 

As for tuning my RX580 I'll certainly give it a try cause why not. But I am doubtful that will solve anything as the shut downs all happen (aside from the psu tests) with the gpu at idle. With me only running the small package test, cinebench, and cpu burner. 

0 Likes

most likely you don't have that NZXT monitoring software installed then.

The OCCT PSU basically runs both the CPU and GPU Stress test at the same time to put maximum stress on the PSU.

While running the PSU test check Temperatures and PSU Outputs.

Yes, I agree, 750 Watt Gold rated PSU is sufficient for your PC setup unless the PSU is going bad.

If you CPU is not going over 85C under heavy stress than it isn't a overheating CPU. But can be some other hardware overheating like the GPU or PSU or a power issue.

Did this problem start from the moment you built this PC or upgraded your PC and was your PC working fine before and then started unexpectedly having these shutdowns.

Or after changing some hardware or Windows update?

Which BIOS Version do you have installed? Try doing a CMOS CLEAR, resetting BIOS back to "default" to see if it is a BIOS setting causing the issue.

Is your RAM Memory listed under Gigabyte's RAM MEMORY QVL List for Matisse processors: https://www.gigabyte.com/Motherboard/B550M-AORUS-PRO-rev-10/support#support-doc

Literally everything is new from the psu to the ram. Only thing that isn't "new" is the gpu that was purchased earlier this year and has given me zero issues till this new build. On top of that. I'm only home in the weekends as of rn due to work. So my time tinkering with my pc is very limited.

My Ram is supported and was specifically designed for Ryzen since it loves memory speed. 16 gigs of gskill trident z neo @ 3600mhz with xmp enabled (which I do have enabled)

All drivers are up-to-date along with the most recent bios

 

0 Likes

In Radeon Settings go to the performance tab, tuning tab, and near the bottom find the "Power Limit" Slider.

Move that slider to the right to its maximum of 50.

Then go to your fan settings and switch to manual. You want to move those dots to have them run at 100 from 70 degrees on up.

If that helps you can experiment with back off that fan curve a bit if it is too loud.

https://community.amd.com/t5/blogs/performance-tuning-with-radeon-software-adrenalin-2020-edition/ba...

Good Luck!

0 Likes

I'll do that on Thursday since I'm currently away from home due to work. 

Thank you guys for your suggestions.

0 Likes