Question GPU failing? Or CPU? Confused!

LarryYadao

Distinguished
Apr 4, 2013
10
0
18,510
Hello everyone.

So i've come across a situation and not sure where to pinpoint it. Hopefully you guys can help. I'm thinking it's my GPU, but im not entirely sure.

First off, here are my PC specs,

AM Ryzen 7 1700
AMD Radeon R9 390
16 gigs of ram


About a month ago, my primary 24 inch Samsung monitor went out so until I got a new one, I was using a 17 inch wide screen. I currently run a dual monitor set up and have never had problems in all the years doing so.

Well, my friend gave me a LG 24 inch 1080p monitor and hooked it in, and everything was fine. Nice new monitor, and no rush to get anything new. Decided to see how games looked so I played Halo infinite with no issues.

Well, a few days later, I jumped on FFXIV (Which is a lot less of a graphic hungry game than Halo) and I played it for a couple of minutes and then game and screen froze and the screen turned black, then audio froze on the last note played in the background, and PC froze entirely where I had to hard reset my PC. Ok, no big deal, maybe hit a glitch. Tried playing again.. Same thing.. Ok.. Well soon after the same problem kept repeating, I got frustrated. The only thing I changed was the monitor so whats going on? So, first thing I did was change the monitor to 720p. Maybe the high res is messing it up? Idk. Anyways, I jumped back into FFXIV and though resolution wasn't too good because it was at 720, game didn't freeze this time! Ok, so that's odd.. Well, luck have it, the game servers went down so I couldn't play anymore anyways to even test if that solved the problem entirely.

So I then decided to jump into Halo since I had no problems playing with it before... And of course, when it loaded up, as soon as the main screen/menu popped up, the problem happened again.. Screen goes black, few seconds later, PC freezes, have to hard reset.

I told myself "Maybe its the monitor?" So, I switched to my old 17 inch monitor and then boom, same thing happened. Wth is going on?

For the record as well, my temps at my CPU and GPU were perfectly fine. Nothing overheating. GPU usually running at 70c and CPU roughly the same. I even had the game moved from one harddrive to another just to make sure the other harddrive wasn't going out.

I can browse the internet fine, I can watch youtube videos fine at 1080p, but as soon as I start a game, this happens.

Now, I WILL say that I have had this similar issue while recording music where I have my electric guitar plugged into an a computer recording interface. Ill be midway into the recording and my PC would do the same thing. It didn't happen often, but it did from time to time.

So basically, this is where i'm at. I think it's my graphics card, but i'm open for suggestions.

Thanks in advance everyone!

PS: I have updated ALL my drivers, windows drivers, and even ran ccleaner to clean anything up that needed it.
 
I would actually consider the PSU as a likely contender for the issue. Recording your guitar would/should not load up the GPU, but it could put some load onto the PSU possibly. I have had a bad PSU cause what appeared to be GPU faults before, as when you load up a GPU, you are also demanding more from the PSU, it can be hard to tell them apart initially. What PSU do you have? Make/Model/Wattage.
 
  • Like
Reactions: LarryYadao
Wow I didn't even think of that. Because it doesn't make sense if the same thing would happen when recording music and think it was the gpu.

My power supply is a Corsair CX series 750 watt modular power supply. I purchased it in 2019.
 
Just curious what speed your ram is at?

I had a 1600x with 3200mhz CL14 ram and it ran pretty well but would occasionally crash.

I think 2667mhz is the "max" recommended by amd.

I eventually upgraded to a 3700x and the memory errors stopped.
 
Well that's not the problem then... Hard drive in good health? Maybe try windows "check disk" or do a defrag if you don't have an SSD...

Maybe a Windows restore back to a known good time...

Outside maybe download and run cinebench to load up the processor and see if it crashes.

You certainly have a proper psu, but maybe it's bad... Tough to say as the 1700 isn't exactly a power hog, in fact it is a pretty low draw.
 
I actually tried to do a system restore but unfortunately it wasn't on so it never made one..

I have one ssd and one standard. I actually moved Halo infinite from the standard to the ssd just to see if the harddrive it was on was crashing. It didn't work. No matter what harddrive I had it on, the same problem happened.

I'm leaning either a bad psu or bad GPU. If it's the psu, it'll be cheaper than a GPU, but we will see lol.
 
Your PSU is decent, so not an immediate need to replace it. It is the cheaper of the two options though. One thing I recommend before you buy a new one though. Do you have a friend, or know someone with a PC that you could install your GPU in? If you put your GPU into a working PC, and the issue starts happening on the other PC as well, then you have figured out that the GPU is the issue. Alternately, if your friends PC also has a GPU that you could put in your system, see if their working GPU still works properly in your system. This way you can figure out if the problem is the GPU, or something else.
 
Ok so update...

I decided to switch out my graphics card to my old gtx 960. Which only has 2 gigs of VRAM instead of the 8 gigs my 390 has..

Well, after updating the drivers, I played BOTH FFXIV and Halo Infinite (even with Halo Infinite minimum requirements of a 4 gig card) and the games ran well with no issues. Temps were all good as well.. So in the end, turns out my graphics card went out. There were a few times my 390 ran around 80c over the past 3 months (Not consistently but happened more often than normal while gaming) so perhaps those were early signs of the card going out.

I appreciate you guys helping me out through this process. You guys always are great.

On the subject, I'd like to get an opinion. So the graphics card im using now, as stated, is a 2 gig card. Halo's minimum requirement is a 4 gig card. Even when I turn everything down graphically on Halo, it states i'd still be using roughly 2.5 vram despite my card being only 2 gigs. HOWEVER, when I DO play it, even playing online, lower res (Lower it to 720p) the game runs at 55-60 fps AND the temp of the card still sits at only 62c-65c. So with that being said, that shouldn't be an issue right? I know PC's nowadays have fail safes if the card gets TOO hot or something is wrong, but despite the card being under the minimum requirements, if the temp is good and i'm getting great fps, I should be ok in regards to damage done to my PC right?

Again, appreciate the help!
 
Minimum requirements are set as a guideline to make the game play nicely (usually). You also have a better system around the GPU, as your CPU and RAM is modern, which might help with using the 960. That same 960 in a "period correct" build may struggle more. That being said, if the card is running stable, and at good temperatures, then there is no reason to be worried about it. GPUs are made to be used at high/max capacity, so overloading a GPU, if the temps are good, means your FPS and/or gaming experience will get worse, but you wont necessarily hurt the GPU. There are cases where you can hurt the GPU, but it is not a huge concern.

As for the 390 being the problem, also consider power draw on the PSU. That R9 390 will draw, under load, 150W+ more power than the GTX 960 will. It is possible your PSU, as it is semi-decent, but not a great PSU, can provide enough power to run the 960 with no issues, but the higher power draw of the 390 pushes it in failure territory. I would still recommend running the 390 in another system to see if there is still a problem. If the 390 runs perfectly in another system, then I would say the PSU is failing at high load, but if the problem follows the GPU, then the PSU is safe, for now.
 
Minimum requirements are set as a guideline to make the game play nicely (usually). You also have a better system around the GPU, as your CPU and RAM is modern, which might help with using the 960. That same 960 in a "period correct" build may struggle more. That being said, if the card is running stable, and at good temperatures, then there is no reason to be worried about it. GPUs are made to be used at high/max capacity, so overloading a GPU, if the temps are good, means your FPS and/or gaming experience will get worse, but you wont necessarily hurt the GPU. There are cases where you can hurt the GPU, but it is not a huge concern.

As for the 390 being the problem, also consider power draw on the PSU. That R9 390 will draw, under load, 150W+ more power than the GTX 960 will. It is possible your PSU, as it is semi-decent, but not a great PSU, can provide enough power to run the 960 with no issues, but the higher power draw of the 390 pushes it in failure territory. I would still recommend running the 390 in another system to see if there is still a problem. If the 390 runs perfectly in another system, then I would say the PSU is failing at high load, but if the problem follows the GPU, then the PSU is safe, for now.


One thing I wanted to note is that my 390 when it was working, ran low settings Halo at 30fps (even at 720p) AND was running 75c and at times, running at 80c. (I cleaned it one time because it was running at around 83-85c) My current 960 runs Halo at low settings at 60fps and hardly even goes over 70c. (Usually sits at 68c while playing Halo) This even with being a lower VRAM graphics card. Not sure if that helps in regards to leaning more towards a GPU issue or not, but I would say if it wasn't for the lower VRAM, i'd dare say the 960 runs Halo a lot better than the 390.
 
That does indicate something was going wrong with the 390. That 390 should perform better than the 960 by a decent amount. The 390 was the competitor for the GTX 970, so it should be a step above the 960. Even if they matched performance, that would be concerning, but if the 390 is slower than the 960, there was something wrong.