Solved GPU overheating issues.

[TakeMeAsIAm]

So, i got an amd radeon r7 250 with 2gb gddr5 and 128 bits. When i first bought the card i mounted it on some old motherboard and everything was ok. I've recently upgraded my pc, and the upgrade was made of a new motherboard, asrock 970 pro3 2.0, new rams and new cpu, amd fx8350. When my pc was 100 finished and usable i decided to play some counterstrike. From 34 degrees in desktop, my gpu temperature jumped up 15 degrees in seconds. In like 30 mins, my pc force shut downed from overheating at around 80 or more degrees With the old motherboard I never had this problems and I played cs go on the gpu with the old motherboars as well. My pc is positioned in a relatively tight place, with not alot of space for the air to move, but the front and the back of the pc have full acces to air
Seeing that I didnt have any sort of overheating problems when I had the old motherboard, can anyone help me? The pc crashed from overheating even when I played roblox, game that requires not more than a 256 mb gpu. I also wanna say that the motherboard is fitting at the limit in the pc case, and that my powersource is 500 w. Any problems causing the overheat and/or solutions.

 

[TakeMeAsIAm]

And I also wanna say that the cpu didnt get hotter than 55, so it is surely from the gpu

 

[alphaniner]

It's normal for a big jump in GPU temp when you first start something that actually stresses it. Your PSU wattage should be sufficient AFAICT.

How do you know overtemp was the cause for the shutdown? Where did that temp (80° or more degrees) come from? Does your GPU heatsink exhaust outside or inside the chassis?

 

[RoasterMen]

What's your power supply? Are both fans from the front and back working?

 

[TakeMeAsIAm]

The heatsink is inside, the overheatint and shut down comes from everything I play and all of the fans are working. The power supply is 500w, and I know that over temp is the cause of the shutdown because it only shuts down when I play

 

[alphaniner]

I know that over temp is the cause of the shutdown because it only shuts down when I play

Based on the information you've shared I don't see how you can be so certain, especially if the CPU temp never exceeded 55°. GPU should automatically throttle itself if it gets too hot, not cause shutdown. I can't find exact specification for your card, but see this thread about the R7 240.

Aside from that, check the System Event Log for clues. Run Memtest and CPU stress test if you haven't already. Run a GPU stress test in windowed mode while monitoring temps and note the GPU, CPU, and System temps when (if) shutdown occurs. Also try to keep an eye on voltages.

And again, what was sensor reported the "80 or more degrees" you mentioned in the first post?

 

[TakeMeAsIAm]

I dont understand your question. What you mean by what was sensor? And i cant know the exact temperature, but after the first shutdown, i kept msi opened and also put my hand on the pc case to see how hot it is. I saw it on msi at around 76 and checked the temp with my hand, and shortly after, the pc force shut downed
I put my hand on the case again and it was way hotter than it was at 76. I have to correct myself, because those werent force shutdowns, but force restarts, and let me explain
After the first overheating my pc force restarted, but i didnt get even to the splash screen
I mean, it reached the first screen that is made by the motherboard and restarted in few seconds. It kept doing so after i manually closed the pc from its button, after letting the restarts go for like 5 times.

 

[TakeMeAsIAm]

If you need any aditional info or details, please ask me

 

[alphaniner]

By sensor I mean temperature sensor. The motherboard itself has a sensor like GPU and CPU. Anyway, since you now mentioned 'msi' I'm guessing you mean MSI Afterburner and the temperatures you mentioned (76, "80 or more degrees") were from your GPU?

Did you overclock your GPU, set any custom profiles, etc?

Is the GPU fan running at 100% when temps get high?

Does it make a difference if your case is open when gaming?

Did you look in event log to see if there were any Warning or Error level events around the time of the forced restarts?

 

[TakeMeAsIAm]

At the first time i booted my pc in xboost mode, mode given by the motherboard itself before the splash screen, but anyway that mode only overclocks the ram
Anyway i turned it off. I didnt mess with the clock speeds or anything, and only when i saw that the temperature is getting so high i put the fan speed on 100. It was 21 until then. Tomorrow i will be able to see the temperature with the case opened, and i think that i will be able today to check the event log. I only type Even Log in the start search thing, or? By the way, i let "repair your systen" or something like that work for like 10 mins after the first shutdown, but it didnt seem to find anything and i stopped it

 

[TakeMeAsIAm]

And yes, the temperature was from the gpu. The cpu didnt get hotter than 55 degrees

 

[RoasterMen]

Can you tell us what brand of power supply are you using?

 

[TakeMeAsIAm]

I think that its a no name sad quality one, but i can say that it is from 2015, has 500 w and that i didnt have any problems with it before. I will be able to tell you the exact brand, if it has one in like an hour

 

[TakeMeAsIAm]

So, i opened even viewer and there were errors, warnings and critical errors. The critical errors were called "kernel-power" and when i clicked it was described as "the system rebooted blah blah blah" so clearly that was the problems. There were also errors like ATI..... And thier describing was like ATI client service error

 

[TakeMeAsIAm]

Searched about kernel power and "An underpowered or failing power supply may cause this behavior. For example, if you added RAM or additional devices or hard disks when this problem began, the power supply may cause the problem.". It's so clear at this point that the supply is causing the problems, right?

 

[TakeMeAsIAm]

Well, on other site there was said that overheating can cause the problem as well. So, gpu or power source? If i run the gpu stress test and it stops at a certian temperature, then can we know if it was caused by the gpu or by the overused power supply?

 

[TakeMeAsIAm]

Well, now im 100% sure that its from the psu because on some websites it says that 75 degrees when gaming is actually cool and and gpus can handle 100 celsius. At this point i think that the only thing i can do is a new power supply, right?

 

[RoasterMen]

Yes, Get a better one. There are many good quality cheap power supplies out there (One with 80 Plus certification)