Question Games crashing (sometimes PC restarts on its own)

Feb 24, 2020
80
3
45
RX 5700 XT Sapphire oc special edition
Ryzen 9 3900X
Aorus x570 Master
Corsair RM750X

I dont even know where to start... Yesterday i was playing Jurassic World Evolution when all of the sudden my screen went black. I instantly thought amd driver issues still present. Then i realised that it wasnt a black crash screen but my pc had actually restarted itself. Hmm odd i thought but that there had to be an explaination. My first thought was that Windows update had a scheduled restart that i had missed because i was too into the game. Nope, everything was up to date and nothing was downloaded recently. Then i booted up Jurassic world again and it happend again, the pc restarted itself. Oh boy here we go again i thought (been having issues ever since i built the pc a couple of weeks ago, it did stabilize for a while but now its just a mess). The 3rd time however it didnt restart but the game crashed. Then i tried forza horizon 4 and it crashed there too.
It was late and I really couldnt be bothered, i'll look at it tomorrow i thought as was clicking my way to shut down. Then i went to the bathroom and when i came back the pc wasnt off, it had restarted... Mildly frustrated i went over and clicked shut down once more and watched what happens. It restarted again... At this point i was getting angry and went for the shut down button on the case, held it down for probably 10 seconds and nothing. Tried again for 20 seconds and still nothing, tried pressning it once and letting go nothing. Pc was refusing to shut down. How is it even possible i thought, how can it ignore the power button??
I was pissed at this point and went for the power supply switch. Bad idea i know but i was just so done man.
Turned it off and went to bed.

Today has been filled with crashes, one or two restarts but mostly crashes. I can shut it down now though so... Yippie?...
I tried DDU and reistalling driver and no change.
Whats interesting is that each time it restarts or a game crashes my AMD Radeon Software settings get reset. I tried not touching anything, undervolting, overclocking, automatic, default it makes no difference, still crashes.
My temps are fine, when gaming highest gpu temp is 73c when using default automatic. My custom fan curve keeps it at 64-65c.
CPU temp is also fine, steady around 50-55 sometimes spikes to 60ish. So really nothing out of the ordinary.

Sooooo.... Anyone got any idea whats going on? My best guess is that my pc is possessed by a demon.

Bruh you have no idea how much i regret going amd. Granted i dont know what the problem is but thats the problem! Is it the "amd driver issues" or is it deeper than that i literally have no way of knowing. If it is driver issues i wont really know unless i test every single little thing hardware wise.
Atleast with Nvidia i would know at this point if something is faulty or not.
 
Last edited:
Seems like it may be a fault PSU or dying motherboard. Do you have a spare PSU or a PSU you can borrow to test your system with? Does it crash or restart when under load (i.e. during games, intensive programs etc)?

Try a different wall outlet for your PC. Then try your system with another GPU if you can borrow one. If it works without your 5700xt then it should be the card's fault. If it still persists, then it could be a motherboard or PSU fault.
 
  • Like
Reactions: Vedixsz
Feb 24, 2020
80
3
45
I
Seems like it may be a fault PSU or dying motherboard. Do you have a spare PSU or a PSU you can borrow to test your system with? Does it crash or restart when under load (i.e. during games, intensive programs etc)?

Try a different wall outlet for your PC. Then try your system with another GPU if you can borrow one. If it works without your 5700xt then it should be the card's fault. If it still persists, then it could be a motherboard or PSU fault.

If only it was that easy, i have no way of testing my components outside of the pc they are in.
Everything is brand new though and everything is working, there is nothing in there thats not doing its thing. The motherboard has those check lights that it goes through whenever the pc is turned on indicating that everything is working properly.

Edit: Yes it only crashes and/or restarts during games so i guess under load if under load is playing a game for a system like mine. I can do other things just fine.

Its funny you mention trying a different wall outlet, its one of the ideas that popped into my head just today, im at work right now so cant try it until i get home. The other idea i had is trying two gpu cables, right now im only using one but it has two heads at the part that goes into the gpu. Im pretty sure thats how you are supposed to do it but i dont know for sure... (?) The gpu has two connectors and both are 6+2 if i remember correctly.
I will also unplug and replug everything perhaps even reslot the gpu. If all this doesnt fix it i will probably try stress testing the components to try and figure out which component is causing it.

I have been searching around and there are plenty off very similar sounding issues to mine. Overheating and either faulty or too weak PSU seem to be the most common causes. Mine is for sure not too weak with 750w so thats why i got the wall outlet idea.
I almost wish overheating was the problem because then you know the problem instantly.

I will update this thread when i try some of these potential solutions and/or i have further insight into the matter.
 
Last edited:
Feb 24, 2020
80
3
45
Wait, are you sure your GPU is plugged in properly? Do check that, because if they aren't that may answer why your thing crashes and restarts. If a component is not getting enough power, the system is designed to crash or reboot to protect itself.

Im not sure but i watched a pc build guide when i was building my pc and it looked like he connected his with one cable that had two ends just like mine. I tried finding the definitive answer to this but it seems to be one of those things that everyone forgets to explain. I will try with two cables and see what happens. If it solves the problem its going to look really ugly with the two ends just dangling under the GPU but oh well. Might get cables with just one head then. IF it solves my problem.
 
Using two power cables used to be the go-to back when GPUs were power drawing monsters. They also worked if the PSU you had was a multirail. Your PSU is of pretty good quality so I would not worry about that. Try 2 cables, but I am more concerned if the cables are inserted properly on both ends, i.e. they are properly inserted into the PSU and GPU.
 
  • Like
Reactions: Vedixsz
Feb 24, 2020
80
3
45
Using two power cables used to be the go-to back when GPUs were power drawing monsters. They also worked if the PSU you had was a multirail. Your PSU is of pretty good quality so I would not worry about that. Try 2 cables, but I am more concerned if the cables are inserted properly on both ends, i.e. they are properly inserted into the PSU and GPU.

Alright so i just got home and i plugged it into a different wall outlet. Booted up Jurassic World Evolution and i was able to play for 30 minutes without crashes or restarts. I dont want to get my hopes up but this is promising.
Another thing is the game started up into fullscreen on its own which it wasnt doing before, it was minimizing itself for some reason.
It makes sense for this to be a power issue because like you said the pc shuts down to protect itself.
Hopefully i have found the solution but only time will tell. I dont have time to play right now but i have tomorrow free to play whole day so we shall see.
 
Feb 24, 2020
80
3
45
Good luck. Let us know if you see anything happen.
I'm very interested too. Tell us how the story ends.

I found the time to play another hour and a half and i was able to do so without crashes or restarts. So thats now a total of two hours gameplay without issues. Im almost certain that i would have had problems had i not tried a different wall outlet.
Tomorrow will be the real test though as i will be playing the whole day essentially and i will most likely run a couple stress tests just to make sure that it can handle being pushed without restarting itself.

If everything goes well i will be updating you at the end of the long gaming session when i know for sure that the issue is solved.
 
Good luck. Remember, changing the wall outlet might have saved you today but if it is a dying PSU it is better to replace it asap before it explodes and takes the rest of your system with it.

If it is not a dying PSU, and all you had to was change the wall outlets, ensure your GPU cables are plugged in thoroughly.
 
Feb 24, 2020
80
3
45
Good luck. Remember, changing the wall outlet might have saved you today but if it is a dying PSU it is better to replace it asap before it explodes and takes the rest of your system with it.

If it is not a dying PSU, and all you had to was change the wall outlets, ensure your GPU cables are plugged in thoroughly.

Alright so i got an update.
I was gaming just fine, ran all kinds of games. Ran The Witcher 3 on ultra with hairworks maxed out, worked no problem.
Then i searched around for benchmarks and found two, Furmark and Heaven.
I did Furmark first and... it crashed. The program stopped responding. I tried it again and this time it restarted the pc and i got a "blue screen of death".
Stop Code: VIDEO_TDR_FAILURE
What failed: atikmpag.sys

Here is what i found relating to Furmark crashing:
"If FurMark crashes it means that your video card wasn’t able to handle the load. If you watch FurMark during the test, you will notice that the image looks odd before the crash. Small dots start to appear across the image because the video card is becoming too hot and the data being sent out from the card is corrupted.

A common reason for FurMark to crash is a graphics card overclock which is too aggressive. If you have overclocked your video card, you’ll need to reduce its settings in order for your video card to remain stable.
If you have not overclocked your video card, the crash might be happening because your video card’s cooling is inadequate. Check the card to make sure that the fan is working and that dust isn’t clogging it up.
With all of that said, your setup might be absolutely fine. Remember, FurMark is an abnormal test. Your graphics card may never be pushed to these limits in a real-world setting."

Perhaps worth noting is that i dont see any "small dots", it just freezes and crashes. Indicating that the crash isnt due to my GPU getting too hot i guess?...
It crashes within two minutes of starting the test and my GPU temp gets as high as 76c. Is it crazy that it gets that high in such a short timeframe? Furmark is the only thing that gets the temp this high this fast. Games cant even get it over 70c.
I have not overclocked my GPU, if anything i have underclocked and undervolted it. Not by much but still.

Now, from what i have read Furmark is like a super intensive test that pushes your system harder than any game will ever do. Sooo im not sure, i still feel like it shouldnt crash if its a thing that people run?

Anyway then i ran Heaven benchmark and it ran fine, no crash. Went trough all 26 scenes without issues.
Then i played some more games without issues.

Edit: I did find that some people are saying that Furmark is a bad test, that it has issues but they dont specify what.

Edit 2: I downloaded MSI Kombustor which from what i understand is similar to Furmark in terms of how much stress it puts on the GPU. I ran it for 15 minutes without issues.
However, one thing that makes zero sense to me is that the program itself showed that my temp got all the way up to 91c while Afterburner overlay showed a max temp of 68c... soo im just like... wut?

I did do a couple of changes so its not exactly the same. I got rid of the underclock and undervolt and let the card run at the settings it came in. which is 2095 frequency and 1200 voltage. Another thing is i didnt touch the power tuning, when i ran Furmark i had it at max (which is +50%) and now i had it at 0%.
I was told setting it at +50% only means that you are allowing the card to draw more power if it needs to and not that its hording all that power no matter what. So i really have no idea but im just going to leave it as is.
 
Last edited:
Feb 24, 2020
80
3
45
Maybe i shouldnt worry really... im finding a lot of people saying that you shouldnt even run Furmark and Kombustor because they put way to much stress on the GPU and could even kill it.
If its running "regular" benchmarks like Heaven and 3dmark (which i did also try and it runs fine) without issues then maybe thats adequate im thinking.
 
Feb 24, 2020
80
3
45
Did another Kombustor test and this time i left the fan speeds on auto, all i did was allowed the fans to go up to 100% if they need to. The highest they got was 80% and the temp never crossed 68c.
Guess thats the way to go from now on.
I might do another Furmark but not right now, probably tomorrow or the day after that because i think the pc needs a break. I have been abusing it all day.

As for Afterburner i think this thing has been lying to me all this time. I monitored my temps with HWMonitor this time and the GPU temp was exactly the same as what Kombustor was showing.
 
Feb 24, 2020
80
3
45
So i thought i had it, i dont have it. Its still crashing and now the GPU has a slight whine sound (that can only be heard over the fans by getting really close to the card) to it whenever i boot up a game. Im not sure if the sound was always there and i just noticed it now that i have the panel open but there is a whine sound nevertheless. I think its time to send the card back bois. I will just request a new one and if that one doesnt work flawlessly out of the box then god help them. Would never even consider AMD again.
 
Sorry for the delayed reply, I did not have a "watch" set on this thread. Whine is pretty much a wildcard these days. For most cards it means nothing and that is fine, but for some, it could be an inherent issue. 76C is not too bad. My 2080ti hits 80C and that is a bit worrisome.

Furmark could be buggy for you which is why it crashed but it still could point to a PSU issue. However, I do not want you spending money unnecessarily on a new PSU if it is a GPU problem. Give the GPU replacement a shot and let us know how it goes. If it is still failing, I guess try a new PSU. Just remember, use DDU to uninstall all AMD drivers before putting in the new GPU, then redownload them from the site to be safe.
 
Feb 24, 2020
80
3
45
Sorry for the delayed reply, I did not have a "watch" set on this thread. Whine is pretty much a wildcard these days. For most cards it means nothing and that is fine, but for some, it could be an inherent issue. 76C is not too bad. My 2080ti hits 80C and that is a bit worrisome.

Furmark could be buggy for you which is why it crashed but it still could point to a PSU issue. However, I do not want you spending money unnecessarily on a new PSU if it is a GPU problem. Give the GPU replacement a shot and let us know how it goes. If it is still failing, I guess try a new PSU. Just remember, use DDU to uninstall all AMD drivers before putting in the new GPU, then redownload them from the site to be safe.

Alrighty no problem mate!
Btw i was able to run Furmark it just stopped crashing for whatever reason... Man i wish computers made sense. It could actually just be that Sapphire has overestimated the card by a crap ton because if i "underclock" it the crashing actually stops completely.
It kind of annoys me that they ship the card at 2095 MHz when its unstable as all hell. It cant hold that for <Mod Edit>.
Im thinking they do that so that they can advertise a boost clock of 2035 MHz. It actually boosts beyond that in some games. Not only does it crash at those speeds it also makes your screen flicker a bit at times when playing games. Completely unsustainable.

Between not knowing if the whine sound was there from the start and not knowing if the thing is even working properly i will just get it replaced to be safe.
However im not sure what i should get though... should i just try my luck at another 5700 xt or throw some more money in there and get a 2070 Super...

Heres the thing, the rx 5700 xt is from what i understand PLUG AND PLAY. Which means that it should work without issues out of the box, granted it might get a bit hot without a proper fan curve running at 1200 mV but it SHOULD work. So if i do get another one and it crashes then i can probably assume that its the PSU right? I really want it to be this simple, it should be this simple.
This stuff is expensive man there should be some consistency i feel.
 
Last edited by a moderator:
The 5700xt had some driver issues early on into its release. I think it was only Sapphire which was not suffering from the issues. Try another 5700xt. If it does not work, I am not sure where else to go either. It is either a PSU problem or driver issues but I am pretty sure you used DDU to wipe drivers and reinstall them so I am not sure either.

The 2070 Super is a good card. If you return the later 5700xt and get a refund perhaps you can put it towards a 2070 Super and a PSU.
 
Feb 24, 2020
80
3
45
The 5700xt had some driver issues early on into its release. I think it was only Sapphire which was not suffering from the issues. Try another 5700xt. If it does not work, I am not sure where else to go either. It is either a PSU problem or driver issues but I am pretty sure you used DDU to wipe drivers and reinstall them so I am not sure either.

The 2070 Super is a good card. If you return the later 5700xt and get a refund perhaps you can put it towards a 2070 Super and a PSU.
Alright so then its likely that there is something wrong with my card, if Sapphire avoided the driver issues.
Yeah i will send the card back and get another 5700 xt. If it doesnt work flawlessly out of the box then i will order a new PSU to see what that does. If at that point im still having problems then i seriously dont know...

Im probably going to be sending the card back tomorrow so the whole process will probably take a couple days at least, when i have acquired the new card i will update you.

Until then!
 
  • Like
Reactions: ArchitSahu
Feb 24, 2020
80
3
45
Quick update for you, there was indeed something wrong with my card. I dropped the card off yesterday (didnt send it back there was actually a shop 30 mins away from me) and today they got back to me saying that the technician discovered that the card was faulty. They ordered a new one to the shop itself that will hopefully be there by monday.
Really relieved that the card was broken, seriously thank god i dont have to troubleshot further. Unless there are multiple components that are broken which would just be such a bummer man. Probably not though i mean what would the odds of that be...

Going to have to use my crappy laptop for a couple of days but at least im on my way to having a fully functioning pc.

I forgot to ask what it was that was wrong with the card specifically but when i go to get my new card i will ask and let you know.

Will update one last time (hopefully) when i get the new card and know that it was the only thing causing issues.
 
Last edited:
  • Like
Reactions: ArchitSahu
Feb 24, 2020
80
3
45
So i got the "new" card and its crashed a few times... Its acting pretty much exactly the same as the one i had. I honestly dont know if they even got me a new card, maybe i just dont trust people but it wouldnt surprise me if they slipped me another crappy one or just gave me back the one i had and hoped for the best. Which is for me to get stuck with a crappy card or add money for a more expensive one which is what i would be doing by getting the 2070s.
The thing is when i slottet the card in i didnt have to install drivers, it had the drivers my old card had.
Im also pretty sure the box it came in was missing a tape that my first box had.
I have ordered a new PSU and cables, im going to see what that does but if its the same then im just going to returen the 5700 xt and get a 2070s. If that doesnt leave me with a proper pc then im just going to return the whole thing and move on with my life. Im losing hair over this mate its actually stressing me out.
 
Last edited:
So i got the "new" card and its crashed a few times... Its acting pretty much exactly the same as the one i had. I honestly dont know if they even got me a new card, maybe i just dont trust people but it wouldnt surprise me if they slipped me another crappy one or just gave me back the one i had and hoped for the best. Which is for me to get stuck with a crappy card or add money for a more expensive one which is what i would be doing by getting the 2070s.
The thing is when i slottet the card in i didnt have to install drivers, it had the drivers my old card had.
Im also pretty sure the box it came in was missing a tape that my first box had.
I have ordered a new PSU and cables, im going to see what that does but if its the same then im just going to returen the 5700 xt and get a 2070s. If that doesnt leave me with a proper pc then im just going to return the whole thing and move on with my life. Im losing hair over this mate its actually stressing me out.

Sometimes we wish that PC's just displayed what the error is whenever it fails, but unfortunately we do not have that level of technology yet.

I feel your pain, give the PSU a shot. You are right in saying that you do not have to get drivers, but perhaps try DDU and see if that fixes anything. If beyond this, the PSU and the new GPU, the only other errors possible would be your motherboard, CPU, RAM or storage and replacing those would essentially be like getting a whole new PC. Let us know if anything changes for you.
 
  • Like
Reactions: Vedixsz
Feb 24, 2020
80
3
45
Sometimes we wish that PC's just displayed what the error is whenever it fails, but unfortunately we do not have that level of technology yet.

I feel your pain, give the PSU a shot. You are right in saying that you do not have to get drivers, but perhaps try DDU and see if that fixes anything. If beyond this, the PSU and the new GPU, the only other errors possible would be your motherboard, CPU, RAM or storage and replacing those would essentially be like getting a whole new PC. Let us know if anything changes for you.

Been back at the store and got the CPU tested as well as the GPU once more and both came out clean. Nothing wrong with them.
Since then i have removed everything and put the pc back together and now i dont seem to be having issues at all. The motherboard displays code AO which from what i have been able to find out means that everything is working correctly. I wasnt getting code AO before when i was having crashes and restarts.

What i will do when i have the time (probably in two days) is take the entire pc to the store and have it all tested just for peace of mind, to ensure that i can go ahead and throw the boxes away and not end up regretting it in a couple of weeks.
 
Been back at the store and got the CPU tested as well as the GPU once more and both came out clean. Nothing wrong with them.
Since then i have removed everything and put the pc back together and now i dont seem to be having issues at all. The motherboard displays code AO which from what i have been able to find out means that everything is working correctly. I wasnt getting code AO before when i was having crashes and restarts.

What i will do when i have the time (probably in two days) is take the entire pc to the store and have it all tested just for peace of mind, to ensure that i can go ahead and throw the boxes away and not end up regretting it in a couple of weeks.
That is great to hear. Good luck!
 

Latest posts