• Pardon our dust as we work on some regularly scheduled forum maintenance. You may notice some missing features during this time. Thank you for your patience!

System reset when gaming/GPU stress testing, but is GPU definitely the problem?

Jan 18, 2019
11
0
10
0
Hello all,

In 2017, I built a mini-gaming rig for my spouse. (Please refer to the bottom of this message for complete list of components used.)
Every component is still within its warranty except for the 980 Ti GPU, which was bought in the Summer of 2015, initially part of my old rig.

The gaming rig had been working perfectly since it was first built.
It's seldom used except for checking e-mails and playing games.
Nothing is overlocked it's not really used for long periods of time.
The OS and BIOS is up-to-date. There's no malware or virus to speak of.
Drivers are all up-to-date. There's no dust.
I have 10 years+ experience building rigs.

For the last couple of months she's been playing Battlefield 1.
There have been zero performance issues to speak of.

2 days ago, when she tried to play Battlefield, the PC suddenly reset roughly 2 minutes after she'd loaded into a server.
There was no warning. No BSOD. No error message. No minidump.

There were a few errors in the Event Viewer, but nothing that seemed out of the ordinary.
They were "DistributedCOM" Event ID: 10016 errors, such as:

XML:
The application-specific permission settings do not grant Local Launch permission for the COM Server application with CLSID 
Windows.SecurityCenter.WscDataProtection
 and APPID 
Unavailable
 to the user NT AUTHORITY\SYSTEM SID (S-1-5-18) from address LocalHost (Using LRPC) running in the application container Unavailable SID (Unavailable). This security permission can be modified using the Component Services administrative tool.

----------------------------------------------------------------

The application-specific permission settings do not grant Local Launch permission for the COM Server application with CLSID 
Windows.SecurityCenter.SecurityAppBroker
 and APPID 
Unavailable
 to the user NT AUTHORITY\SYSTEM SID (S-1-5-18) from address LocalHost (Using LRPC) running in the application container Unavailable SID (Unavailable). This security permission can be modified using the Component Services administrative tool.
 
----------------------------------------------------------------

The application-specific permission settings do not grant Local Launch permission for the COM Server application with CLSID 
Windows.SecurityCenter.WscBrokerManager
 and APPID 
Unavailable
 to the user NT AUTHORITY\SYSTEM SID (S-1-5-18) from address LocalHost (Using LRPC) running in the application container Unavailable SID (Unavailable). This security permission can be modified using the Component Services administrative tool.

----------------------------------------------------------------

The application-specific permission settings do not grant Local Activation permission for the COM Server application with CLSID 
{2593F8B9-4EAF-457C-B68A-50F6B8EA6B54}
 and APPID 
{15C20B67-12E7-4BB6-92BB-7AFF07997402}
 to the user MALWINE\Malwine SID (S-1-5-21-2893702971-324590147-2545585346-1001) from address LocalHost (Using LRPC) running in the application container Unavailable SID (Unavailable). This security permission can be modified using the Component Services administrative tool.

----------------------------------------------------------------

The application-specific permission settings do not grant Local Activation permission for the COM Server application with CLSID 
{6B3B8D23-FA8D-40B9-8DBD-B950333E2C52}
 and APPID 
{4839DDB7-58C2-48F5-8283-E1D1807D0D7D}
 to the user NT AUTHORITY\LOCAL SERVICE SID (S-1-5-19) from address LocalHost (Using LRPC) running in the application container Unavailable SID (Unavailable). This security permission can be modified using the Component Services administrative tool.
Giving those services permissions did not fix anything.

So, naturally, my first thought was the check the temperatures.
When I checked the temps, everything seemed normal.
I kept monitoring the temps when starting up BF1 and nothing was out of the ordinary.
In fact - and probably due to not overclocking - the temps were very good.
The CPU remained in the 40c's under load and the GPU didn't get too hot either averaging about 60c - sometimes less - again, everything at stock.
Still, once I got into game, I could play for maybe 2-3 minutes and then the PC would suddenly reset without any error message.

Wondering if it was limited to Battlefield, I tried some benchmark programs like 3DMark and RealBench.
Both reset the PC after a few seconds-minutes. There were no voltage spikes or high temps or anything, though.
My research led me to believe that it could be the PSU, but I'm not sure. Everything is pointing toward the GPU.

Prime95, AIDA64, IntelBurnTest, and OCCT (currently been running OCCT for over an hour) haven't caused any issues/error messages.
Neither has Cinebench15. All of them have ran from start to finish without resetting the machine. And temps can be seen here:



If I don't go into game, I can use the PC, seemingly, just fine. Videos and music play fine, too.
I have also made sure that everything was connected properly inside the rig - it was. There's no dust or anything.

I'm going to run one or two more tests just to rule out something like RAM (memtest85), but I'm confident it's either the GPU or PSU.

Question is, would anybody be able to help me pinpoint the exact cause? Perhaps it ISN'T a hardware issue, but something else?
Also, seeing as the PSU is still within its warranty, I'm not too worried about replacing it.
However, if it's the GPU, I'll have to buy a brand new one - 980Ti didn't come cheap!
I'd like to see if there's anything else I can do just to be sure at this point.

------------------------------------------------------
CPU: Intel i5-6500 @ 3.20GHz
CPU COOLER: Noctua NH-U12S
Motherboard: Gigabyte H270N-WIFI-CF
RAM: Corsair Vengeance 8GB (2 x 4GB)
GPU: EVGA NVIDIA GeForce GTX 980 Ti
PSU: Corsair CX550M PSU
MONITOR: ASUS VN248
CASE: Fractal Design Node 304

OS: Windows 10 Home 64-bit
Version: 1809
OS build: 17763.253

NVIDIA Driver Version: 417.71
All other drivers up-to-date via Drive Easy
BIOS: F8d (latest)
AV: Windows Defender
 

jankerson

Illustrious
BANNED


That would be a problem. :D

You could pull the MB out of the case and hook it all up that way, but that's a bit of work.

Best just to get a new GPU and be done with it, and throw that old PSU in the garbage.

Not sure who recommended it to you in the 1st place for that system, but it was a mistake from the start. I am surprised that it lasted that long before something happened.

High end GPU's and budget PSU's are never a good mix, nothing that is good ever happens in the end.
 

jankerson

Illustrious
BANNED
Sounds like the PSU is going bad.

And that PSU is too light for a 980Ti in the 1st place and not ment for a gaming PC.

Need to get a better quality PSU and 650W.

Budget PSU's and high end GPU's are never a good mix.
 
I think trying the GPU in another machine (if available) might be the easiest way.

....but another thing I noticed.....AIDA64 doesn't give the voltage for the 12V rail.

It does give the +5V and + 3.3V (which are good by the way)....but if that 12V rail was going low....that could cause what you are seeing as well as a bad GPU.
 

jankerson

Illustrious
BANNED



It's the PSU giving out, the CX series isn't great at all and not ment for that type of PC.
 
Jan 18, 2019
11
0
10
0




Granted CX is a Bronze grade PSU, but that wattage is more than sufficient to run a 980Ti on a non-overclocked machine. I've seen it corroborated as such on several forums and there are a number of rigs using the same components as me on PCPartPicker. Making sure the CX550M would run the 980Ti was a query I had from the get go and it was answered. This rig has been running perfectly fine for 1.5 years.
 

jankerson

Illustrious
BANNED



It's a POS PSU that is a budget unit that isn't ment for gaming machines.

That is the bottom line.

Should have gotten the TXM or RMX.


 

jankerson

Illustrious
BANNED


Both are garbage, but still get recommended due to mainly politics and ignorance.

Anything cheap gets recommended too often and for the wrong uses.
 
Jan 18, 2019
11
0
10
0


I'll take your word for it. Honestly, I picked it due to its relatively cheap cost and that it had been recommended to me on more than one occasion. This was also going to be a non-overclocked build so I didn't worry too much. What would you guys recommend in its stead? Of course, it would need to fit within the Node 304 - I'm sure there was a limit as to the total size PSU it could take.
 

jankerson

Illustrious
BANNED


EXACTLY... Cheap.... That's is the problem and why they get recommended way too often. (Always has been a major issue in the PC world recommending cheap parts due to price) Now you don't have to buy $150+ PSU's either, but quality units really don't cost much more than the budget units in general.

You just need a better quality unit:

Seasonic Focus Plus, this would be the one to pick for your case.
 
Jan 18, 2019
11
0
10
0
So I received the new Seasonic PSU today. Definitely looked more up market compared to the Corsair CX550M.
Everything installed nicely - even if a little tight.
Unfortunately - and I say that gritting my teeth - he PSU was not the problem.
I ran a RealBench stress test and tried to play Battlefield 1. Each time the PC reset.
Temps were still fine and zero error messages.

After all the of the rhetoric and "bottom line" assurances I received yesterday about how it just *had* to be the "cheap" little Corsair PSU, that had been working fine for 18 months, as you can probably imagine, I'm pretty damn livid after the hallowed gold star Seagate PSU didn't fix the problem.

Does anybody else have any other suggestions as to the cause of the restarting?
Figuring it must just be GPU now, but I don't fancy spending £300-£500 on something else that does not fix the problem.
Perhaps it's software related?

**EDIT** Ran memtest85 - RAM. 0 errors detected.
 

jankerson

Illustrious
BANNED


I doubt that it's the GPU.

Might be the MB also.

As suggested try the GPU in another system, if it's fine then that leaves the MB.

However keep the new PSU either way because something will have to be replaced, either the MB or the GPU.

The old PSU is still suspect in the end, being that it was really too weak for the setup in the 1st place it could have damaged the GPU or the MB. Don't want to risk damaging the new MB or GPU.

It's defiantly a hardware issue and you replaced one suspect part already that could have damaged something else and or could damage anything new that is replaced.

Think you are livid now, how would you feel if you sent back the new PSU and got a new MB or GPU and in 2 weeks when the new part goes also because you kept that suspect PSU? One that you spent real money on.

 

jankerson

Illustrious
BANNED


Why would he send back the NEW PSU?

When the old PSU is still suspect and highly likely damaged the MB or GPU.

And it's VERY possible it could damage anything new that would be replaced.

 

mortemas

Reputable
Feb 11, 2015
506
19
5,165
70
While hardware might predominantly be the reason for sudden restarts, it does not mean it must always be the cause. I am convinced that my sudden restarts were fixed by either the Win10 October update or another windows update called "Intel Corporation - SoftwareComponent - 11/25/2018 12:00:00 AM - 2.2.100.47975" which I believe was a microcode update. If you have not already received it unprompted, try requesting the October update and look for the Intel update in your update history to see if Microsoft has pushed it to your PC. Besides the restart problem, it also fixed my screen saver, which was not working for months. See my post here for the whole story: http://www.tomshardware.com/answers/id-3861800/hard-stops-restarts-mins.html#21672261
 

jankerson

Illustrious
BANNED


Any Microcode update would also be on the MB website in the driver section.
 
Jan 18, 2019
11
0
10
0
Okay. I installed the GPU on my personal rig. Temps were still fine. Booted up Call of Duty Black Ops 4. Boom. My rig reset.

I was reluctant to do this from the get go because a) it's tedious and b) I'd rather not put potentially damaged components in my otherwise flawlessly running enthusiast system.

Is it safe to say the GPU was the only problem or could the motherboard still be at fault or even the CPU via, possibly, latent damage caused by an allegedly awful PSU?

Forgive me for being snarky. Having a bad day. Ha!
 

jankerson

Illustrious
BANNED


I would say the GPU is damaged.

However put your GPU in the other machine that you installed the new PSU in and see if it resets.

Make sure the NEW PSU is still in there, don't hook up your GPU to that old suspect PSU whatever you do, not worth the risk.

If it runs fine, throw that old PSU in the garbage and get a new GPU.

And be thankful that the whole machine wasn't wasted.
 

jankerson

Illustrious
BANNED


That would be a problem. :D

You could pull the MB out of the case and hook it all up that way, but that's a bit of work.

Best just to get a new GPU and be done with it, and throw that old PSU in the garbage.

Not sure who recommended it to you in the 1st place for that system, but it was a mistake from the start. I am surprised that it lasted that long before something happened.

High end GPU's and budget PSU's are never a good mix, nothing that is good ever happens in the end.
 
Jan 18, 2019
11
0
10
0
I'll buy her a new GPU. Probably a 1070 or 2060 assuming the new PSU can handle either. Needs to be something for ITX Node 304 case, though.

The CM550X was actually recommended for the 980Ti by about a dozen people in two separate threads on this forum back in 2017. I then saw it being used in a number of completed builds on PCPartPicker. No one mentioned that it wouldn't sufficiently power a 980Ti. Quite the opposite in fact, that it was more than enough. My initial experience with the newly built rig was that everything was working perfectly well. My decision to buy it was based on over a dozen peoples say so on a forum. Everything to correct the problems I've been experiencing recently is pretty much based on your word only. I'm following the advice of people that volunteer their expertise. Not always easy to tell who knows what they're talking about and who doesn't, especially if there's no voice to the contrary.
 

jankerson

Illustrious
BANNED


The Seasonic Focus Plus you bought will handle any GPU you put in it with zero issues.

Yeah.. I am not surprised at all about the recommendations here, I see that sort of thing all the time and it gives me a headache every time I see it. One can only beat their head against the wall so much you know.

The PSU is the MOST important part in the PC, everything is hooked into it. It's the last thing one should ever skimp on, that's both wattage and quality.
 

ASK THE COMMUNITY

TRENDING THREADS