[SOLVED] Older System Freezing

jcwatts13

Honorable
Oct 14, 2017
6
0
10,510
Greetings,

I built this system in 2016 so it is showing its age, but it was really good at the time, and I think there may yet still be some life in it. However fairly recently, it has started developing a problem of freezing up. It used to do it very rarely, but has been doing it more often. This kind of freeze up is not a BSOD, it's not a black screen. The screen just freezes in place with whatever I was doing, usually gaming. The fans stay running, the GPU load indicator light goes to zero. The event viewer just shows event 41, kernel power. The only way to get out of it is to do a hard reset by holding the power button down.

I was fairly confident it was the GPU, and I figured since it was fairly old I'd replace it. I replaced it with the same model as I had before and it runs a lot cooler (stays around 60C at most), and FPS definitely improved. But, it froze up again last night, while playing eve online.

Specifications are as follows:

Windows 10 Home 64-bit
AMD FX-9590 (with Thermaltake 3.0 AIO CPU cooler)
16 GB RIPJAWS 11-11-11-28 4 x 4 DDR3
ASUSTeK SABERTOOTH 990FX R20 Motherboard
RADEON R9 FURY X 4GB HBM GPU
120 GB SanDisk OS SSD
500 GB Crucial SSD
1 TB WD Mechanical Drive
EVGA SUPERNOVA G2 1000W

I've ran windows memory diagnostic tool with no erroneous results. GPU drivers are up to date. At this point I'm thinking it may be the power supply, as both my CPU and GPU suck down the power like no other. However, a 1000W EVGA is quite a beast of a power supply, and I think EVGA is likely the best in the business when it comes to these. Before buying another, I'd like to do some troubleshooting, but I'm not sure where to begin. I've been monitoring temps/metrics using AMD's software, and I see nothing out of the ordinary, except with multiple game clients running I'm using up around 3.6 gigs of the 4 gigs available of video memory, which I think is par for the course because I'm running 3 displays, two of them 1080p and one at 4k.

I'd appreciate any advice, I'd like to narrow this down before spending further money on parts that aren't the problem.
 
Solution
Consider release date of 2013 for the CPU, 2015 release date on the GPU. This system has given exceptional longevity. Anything you do in regard to troubleshooting this system should be focused towards parts that would be useful to a new(er) build. At this point, if it remains an intermittent issue, deal with it as you save money. Consider your next build.

If you do as good a selection on your next which meets your needs this long, I would say job well done (both times).

Lutfij

Titan
Moderator
I built this system in 2016 so it is showing its age, but it was really good at the time, and I think there may yet still be some life in it.
AMD FX-9590 (with Thermaltake 3.0 AIO CPU cooler)
ASUSTeK SABERTOOTH 990FX R20 Motherboard

I doubt considering that the processor you have was capable of destroying most motherboard's.

It's possible that your PSU might not be capable of delivering power to your entire system after all these years of stress. See if a replacement PSU(borrowed from a friend or neighbor, not bought) with at least 850W of power at the entire system's disposal changes your issue. EVGA have a reputation behind them, true, but it's been known for a long while now that they have duds even in their top tier lineups.

What BIOS version are you on for your motherboard at this moment of time?
 

jcwatts13

Honorable
Oct 14, 2017
6
0
10,510
I built this system in 2016 so it is showing its age, but it was really good at the time, and I think there may yet still be some life in it.
AMD FX-9590 (with Thermaltake 3.0 AIO CPU cooler)
ASUSTeK SABERTOOTH 990FX R20 Motherboard

I doubt considering that the processor you have was capable of destroying most motherboard's.

It's possible that your PSU might not be capable of delivering power to your entire system after all these years of stress. See if a replacement PSU(borrowed from a friend or neighbor, not bought) with at least 850W of power at the entire system's disposal changes your issue. EVGA have a reputation behind them, true, but it's been known for a long while now that they have duds even in their top tier lineups.

What BIOS version are you on for your motherboard at this moment of time?

First I've heard of it destroying motherboards, I know it can run hot if you let it, but it's been water-cooled since Day 1. I've run the CPU stress test on CPU-Z and it gets nowhere near maximum temp even if I let it run for long periods. I don't have a PSU I could borrow unfortunately. I was thinking of just getting another 1000/1200 EVGA, and hoping the modular version would fit my current cables so I wouldn't have to reroute them all.

According to CPU-Z my bios version is: 2501 - AMD AGESA OrochiPIV1.5.0.2 with a release date of 04/08/2014.

I went to the Asus web site and tried to find more up to date drivers, but the only thing offered was realtek audio (which I'm not using) and WLAN drivers. Not sure if I'm looking in the right place.
 

punkncat

Champion
Ambassador
Please do not reuse connectors from a different PSU. Even if the same model a newer revision could be different in some way. Certainly, get verification from the manufacturer if you decide to do so.

Just in reading and with no other indicator it would point at PSU. At the same time though, event 41 points to any unexpected shutdown. It happens to be related to power enough to consider a good indicator but isn't always correct.

Might be worthwhile to explore a PSU tester. Anything inexpensive isn't going to be foolproof, but you could bet that if it fails the most basic tests those provide you know where the answer lies. I would balance that against the cost of a good power supply. In a case such as this if it turns out to be something else at least you have a good PSU to base the troubleshooting on.
 

jcwatts13

Honorable
Oct 14, 2017
6
0
10,510
Please do not reuse connectors from a different PSU. Even if the same model a newer revision could be different in some way. Certainly, get verification from the manufacturer if you decide to do so.

Just in reading and with no other indicator it would point at PSU. At the same time though, event 41 points to any unexpected shutdown. It happens to be related to power enough to consider a good indicator but isn't always correct.

Might be worthwhile to explore a PSU tester. Anything inexpensive isn't going to be foolproof, but you could bet that if it fails the most basic tests those provide you know where the answer lies. I would balance that against the cost of a good power supply. In a case such as this if it turns out to be something else at least you have a good PSU to base the troubleshooting on.

I would purchase a tester, I'm just not certain it would even show anything, because this problem seems to be intermittent, and only under heavy load with both GPU and CPU, as stress testing the CPU or running benchmarks for the GPU do not reproduce the problem. Would under volting the GPU perhaps help out with a weak power supply?
 

jcwatts13

Honorable
Oct 14, 2017
6
0
10,510
Please do not reuse connectors from a different PSU. Even if the same model a newer revision could be different in some way. Certainly, get verification from the manufacturer if you decide to do so.

Just in reading and with no other indicator it would point at PSU. At the same time though, event 41 points to any unexpected shutdown. It happens to be related to power enough to consider a good indicator but isn't always correct.

Might be worthwhile to explore a PSU tester. Anything inexpensive isn't going to be foolproof, but you could bet that if it fails the most basic tests those provide you know where the answer lies. I would balance that against the cost of a good power supply. In a case such as this if it turns out to be something else at least you have a good PSU to base the troubleshooting on.

I've even tried running CPU-Z and Furmark together to stress both the CPU and GPU, and I cannot get it to malfunction. Super frustrating.
 

punkncat

Champion
Ambassador
Consider release date of 2013 for the CPU, 2015 release date on the GPU. This system has given exceptional longevity. Anything you do in regard to troubleshooting this system should be focused towards parts that would be useful to a new(er) build. At this point, if it remains an intermittent issue, deal with it as you save money. Consider your next build.

If you do as good a selection on your next which meets your needs this long, I would say job well done (both times).
 
Solution

jcwatts13

Honorable
Oct 14, 2017
6
0
10,510
Consider release date of 2013 for the CPU, 2015 release date on the GPU. This system has given exceptional longevity. Anything you do in regard to troubleshooting this system should be focused towards parts that would be useful to a new(er) build. At this point, if it remains an intermittent issue, deal with it as you save money. Consider your next build.

If you do as good a selection on your next which meets your needs this long, I would say job well done (both times).

Thanks, I ordered a G3 1000W and if it persists, I figure I can use that PSU on a new build as well. Fingers crossed, as I don't play too many cutting edge games, so this thing has been chugging along just fine for the most part.