GPU turning off while gaming

[Mr Shea]

Hello, hopefully someone can help me with my issue. I have been having an issue where my computer's GPU turns off when im playing games, such as CSGO. I cant make it through a match without it happening once. The GPU turns off and flashes white but the rest of the computer stays on. the cpu stays one and spinning, all the fans stay on and spinning, and the mobo stays on and lit up with leds. I have to force restart to get it working again. I was told it was the GPU overheating even though my censers said it wasnt, so i got 3 more 120mm fans. The problem continued... Then i was told it was my GPU failing so i got myself a brand new RX580... The problem continued. Then i was told my PSU is faulty and should upgrade it. Now i have a 850W Rosewill Glacier PSU, and the problem continues... Soon i will have 2 full computers before fixing this problem... I now have a program called HWINFO 64 that suppose to help me understand but i cant find a program that can read the saved documents saying what was what, it just comes out all scribbled and matrix like. The most confusing part to me is if i benchmark my computer with cinebench or furmark or heaven or valley or others that pound both the cpu and gpu at the same time, i cant get it to black screen like when i game.
I shall list my computer specs below and hope someone can help me figure out my issues.

PSU - Rosewill 850w
GPU - MSI RX 580
Ram - 16gig ram (2x8)
MOBO - msi gaming pro
CPU - Ryzen 1200 @3.8 (soon will be a ryzen 5 1600 cause why the hell not)
Fans- 6 120mm (air goes in the front and out the back and top)

 

[PHaS3]

Hi

So you've replaced the GPU completely and still have the issue. You only give CSGO as an example, do other games exhibit this behavior?

I assume you've applied no overclock to the card, since you replaced it. Have you installed the latest AMD drivers? If not, start there.

If that doesn't help, use DDU to completely remove the AMD drivers, and reinstall them fresh again.

If that doesn't help, try set your CPU back down to stock (as well as anything else you've OCd, e.g. RAM).

 

[Caring1]

..... I now have a program called HWINFO 64 that suppose to help me understand but i cant find a program that can read the saved documents saying what was what, it just comes out all scribbled and matrix like.

There's no need to save any document from that program, just install, and open, it displays all the info you require on the screen, if you want to save a copy, screen shot it.
Post a copy here for members to look at, also GPU-z sensor tabs might help.

 

[HD64G]

A friend was forced to change his dying PSU when he had many shutdowns while playing heavy games on his Nitro RX480. He never had any problem like that since then.

 

[DeathtoGnomes]

I would look at the logs from windows event viewer, specifically the ones with bright red circle with an X. If you dont know how to troubleshoot, now is good time to start learning. Help us help you by posting logs and screenshots of logs.

Throwing parts at the problem only damages the savings account.

 

[Mr Shea]

I would look at the logs from windows event viewer, specifically the ones with bright red circle with an X. If you dont know how to troubleshoot, now is good time to start learning. Help us help you by posting logs and screenshots of logs.

Throwing parts at the problem only damages the savings account.

I have checked the windows event viewer and the only red with x (critical) i see is one with the source of "kernel Power".
The crash or sort of crash happens with most of the games i play, fallout and even non demanding ones like polybridge
When im playing i can check my temps and neither the cpu nor the gpu get higher that 60c.
I reinstalled the drivers again and still same problem.

 

[Bill_Bright]

A friend was forced to change his dying PSU when he had many shutdowns while playing heavy games on his Nitro RX480

The OP said he already replaced the PSU. That said, every home and every computer user should have access to a AC Outlet Tester to ensure your outlet is properly wired and grounded. I recommend one with a GFCI (ground fault circuit interrupt) indicator as it can be used to test bathroom and kitchen outlets (outlets near water) too. These testers can be found for your type and voltage outlet, foreign or domestic, (like this one for the UK) at most home improvement stores, or even the electrical department at Wal-Mart. Use it to test all the outlets in the home and if a fault is shown, have it fixed by a qualified electrician.

If you have access to a multimeter, and know how to use it safely, you might check your mains voltage to make sure it is not excessive. If too high, it can result in stability issues.

The tap on the transformer failed in my neighborhood last year and was sending all the homes on this block 292VAC resulting in 146VAC at our outlets! It was confusing at first because all my UPS were beeping and supplying battery power to my systems, but all my house lights were still on. Then I checked the display panel on my UPS and saw the input voltage was 146VAC. Not good. Starting to panic, I verified this with my multimeter. It should have been around ~120VAC.

I quickly called the power company, told them I was a tech, and they would be paying for all the neighborhood's replacement air conditioners and refrigerators and whatever else fries if they don't come fix this quick! No exaggeration - they were at my house within 15 minutes after I called. They were able to bring it down to 126VAC that night by moving to a different tap temporarily. Within a week we got a new transformer and we are now at ~120VAC.

What that is certainly a rare event, it could happen so worth checking.

The most confusing part to me is if i benchmark my computer with cinebench or furmark or heaven or valley or others that pound both the cpu and gpu at the same time, i cant get it to black screen like when i game.

You only mentioned CSGO. Does it happen with other games and apps?

and the mobo stays on and lit up with leds.

Check your motherboard to see what the pattern of lit LEDs mean. And I agree to stop overclocking - until this is resolved.

You might try running with just 1 stick of RAM and see what happens. If no problem, swap sticks.

 

[Zyll Goliat]

Hmmm.....seeing that you already tried the usual suspect like PSU&GPU there might be a possibility that is something wrong with your mobo or the actual PCI-E slot if you have another slot you can take out your GPU and put her in that other slot and see if the problem persist.....

 

[eidairaman1]

Put rig on another circuit, reset the mobo bios too

 

[Mr Shea]

The OP said he already replaced the PSU. That said, every home and every computer user should have access to a AC Outlet Tester to ensure your outlet is properly wired and grounded. I recommend one with a GFCI (ground fault circuit interrupt) indicator as it can be used to test bathroom and kitchen outlets (outlets near water) too. These testers can be found for your type and voltage outlet, foreign or domestic, (like this one for the UK) at most home improvement stores, or even the electrical department at Wal-Mart. Use it to test all the outlets in the home and if a fault is shown, have it fixed by a qualified electrician.

If you have access to a multimeter, and know how to use it safely, you might check your mains voltage to make sure it is not excessive. If too high, it can result in stability issues.

The tap on the transformer failed in my neighborhood last year and was sending all the homes on this block 292VAC resulting in 146VAC at our outlets! It was confusing at first because all my UPS were beeping and supplying battery power to my systems, but all my house lights were still on. Then I checked the display panel on my UPS and saw the input voltage was 146VAC. Not good. Starting to panic, I verified this with my multimeter. It should have been around ~120VAC.

I quickly called the power company, told them I was a tech, and they would be paying for all the neighborhood's replacement air conditioners and refrigerators and whatever else fries if they don't come fix this quick! No exaggeration - they were at my house within 15 minutes after I called. They were able to bring it down to 126VAC that night by moving to a different tap temporarily. Within a week we got a new transformer and we are now at ~120VAC.

What that is certainly a rare event, it could happen so worth checking.

You only mentioned CSGO. Does it happen with other games and apps?
Check your motherboard to see what the pattern of lit LEDs mean. And I agree to stop overclocking - until this is resolved.

You might try running with just 1 stick of RAM and see what happens. If no problem, swap sticks.

As it happens i work at home depot! so i shall pick up one of those testers to find out the answer! and its happens in most games i play. so fallout series, the forest, gta 5, and yes, polybridge the next time it happens i will check the mobo for a post code, but i do not think it will show one because it has only showed post codes in the past when it turns off. When black screen thing happens, the mobo stays on and working, so do the cpu and hard drives keep making noise.
and FINE!! ill take of my over clocks thanks MOM! ruin my fun!

 

[John Naylor]

Well if ya already have HWiNFO here's what I would be looking at ....

1. Start up HWiNFO, "sensors only"..... move to right side of screen
2. Run RoG Real Bench (benchmark takes about 8 minutes)
3. When completed, Record max temps and min / max voltages on each core as well as the 3 voltage rails (3.3, 5 and 12v)

4. Let system chill for 10 minutes and rest all mon / max settings in HWiNFO
5. Close RoG RB ... open Furmark and run the flying donut until the temp curve flattens
6. Record max temps in Firmark as well as min / max on each rail.

For the rails ...

..... anything > 5% fails to meet standards an the PSU shud be replaced under warranty
..... anything < 2.0 or even 2.5% should be good for moderate overclocking
..... anything > 1% should be your target for serious overlcocking.

 

[DeathtoGnomes]

I have checked the windows event viewer and the only red with x (critical) i see is one with the source of "kernel Power".
The crash or sort of crash happens with most of the games i play, fallout and even non demanding ones like polybridge
When im playing i can check my temps and neither the cpu nor the gpu get higher that 60c.
I reinstalled the drivers again and still same problem.

As I said, post logs and pics if needed. Saying just "kernel Power" is not enough to go on. What else does it say?

But, if you prefer we can continue to speculate without any hard knowledge, so lets stick with crappy answers for you and just say its the PSU again! There solved.
We all know thats not exactly the solution here. Be a little more forthcoming with actual information, not all of us want to be armchair quarterbacks.

 

[Laurijan]

Maybe a bad hdmi cable? i had lots of problems with hdmi cables and 4K resolution. Most likely were not shielded enough for my EM mess at my computer desk.

 

[Bill_Bright]

and FINE!! ill take of my over clocks

It's just while testing. You might find just returning all your voltages/clocks to the defaults resolves all your stability issues.

 

[Mr Shea]

It's just while testing. You might find just returning all your voltages/clocks to the defaults resolves all your stability issues.

OK so a few days with no overclocked cpu or ram and the problem no longer occurs. Does this mean i just needed a little more voltage for the cpu or something?
I have a ryzen 5 1600 coming after the weekend so i am looking forward to some overclocking fun. Any pointers or hints that could help me in my journey?
Thanks everyone for helping me with my issue, it is now solved great teamwork dudes or dudettes

 

[eidairaman1]

OK so a few days with no overclocked cpu or ram and the problem no longer occurs. Does this mean i just needed a little more voltage for the cpu or something?
I have a ryzen 5 1600 coming after the weekend so i am looking forward to some overclocking fun. Any pointers or hints that could help me in my journey?
Thanks everyone for helping me with my issue, it is now solved great teamwork dudes or dudettes

Yes, keep your room temperatures cool, keep good airflow in the case, do dustout monthly too.

 

[PHaS3]

OK so a few days with no overclocked cpu or ram and the problem no longer occurs.

I'm glad that what I suggested in the first reply eventually resolved the issue

As @eidairaman1 says, keeps temps down, keep it clean. Use this experience to understand instability in overclocks.

Good luck with the new CPU.

 

[Bill_Bright]

Does this mean i just needed a little more voltage for the cpu or something?

It likely means you were providing the wrong voltage for optimal performance.

Overclocking it not a set and forget thing. It is a balance between voltage, temperature and stability. And remember, overclocking does NOT automatically mean a better gaming experience.

 

[PerfectWave]

i had kernel power problem with w10 version before the 1803 with my fx8350 after the update the pc went stable again. The pc randomly shut down for no reason at all.

 

[Sufran]

I assume you've applied no overclock to the card, since you replaced it. Have you installed the latest AMD drivers? If not, start there.

If that doesn't help, use DDU to completely remove the AMD drivers, and reinstall them fresh again.

If that doesn't help, try set your CPU back down to stock (as well as anything else you've OCd, e.g. RAM).

I'm glad that what I suggested in the first reply eventually resolved the issue

As @eidairaman1 says, keeps temps down, keep it clean. Use this experience to understand instability in overclocks.

Good luck with the new CPU.

Mate, I tried all of this. I also tried re installing the old gpu drivers. And still the problem occurs. Still my display goes off whenever I play games but other than that, the pc is still on. I also may add that when my display goes off, when i try ko hit numlock button the led lights won't go on or off.
Cable i used is DVI-D single link. My monitor ACER KA220HQ only supports VGA and DVI. I don't have the option to use HDMI.
I also updated my windows (1903)
I also updated my motherboard bios.
Is my power supply a problem?

Ryzen 3 2200g (no oc)
2x4gb ram 3200mz (but i only runs at 2400mhz)
Gigabyte Ab350m motherboard
Gigabyte Rx 580 4gb (no oc)
Rave 500w 80 plus

 

[fullinfusion]

Mate, I tried all of this. I also tried re installing the old gpu drivers. And still the problem occurs. Still my display goes off whenever I play games but other than that, the pc is still on. I also may add that when my display goes off, when i try ko hit numlock button the led lights won't go on or off.
Cable i used is DVI-D single link. My monitor ACER KA220HQ only supports VGA and DVI. I don't have the option to use HDMI.
I also updated my windows (1903)
I also updated my motherboard bios.
Is my power supply a problem?

Ryzen 3 2200g (no oc)
2x4gb ram 3200mz (but i only runs at 2400mhz)
Gigabyte Ab350m motherboard
Gigabyte Rx 580 4gb (no oc)
Rave 500w 80 plus

Did it do it on non 1903 Windows ?

 

[Sufran]

Did it do it on non 1903 Windows ?

I previously have non 1903 windows before upgrading. Still have the same issue.

 

[Bill_Bright]

Is Sufran hijacking the OP's thread?

 

[Sufran]

Is Sufran hijacking the OP's thread?

mate, what r u talking? enlighten me. I'm just asking here

 

[Tatty_Two]

mate, what r u talking? enlighten me. I'm just asking here

I think he is suggesting that you start your own thread with your own issues rather than jumping in on already solved issues that may be different to yours.