Sign in with
Sign up | Sign in
Your question

GPU or CPU overheating

Tags:
  • Nvidia
  • GPUs
  • CPUs
  • Graphics
Last response: in Graphics & Displays
Share
February 19, 2013 4:11:42 PM

Hello,

After searching, reading, searching some more and reading some more I finally decided it was time to accept that here is something I cannot find a solution to myself.

My issue is as following:
Either my GPU or CPU overheats when I'm playing some games, when this happens, the PC crashes... Sound is stuttering and the fans are still blowing.
I then press the restart button on the computer and the system starts up as it's supposed to.

I've been trying to monitor my GPU heating and it seems to hit between 60-65 celcius when playing games like Skyrim, Diablo 3, SWTOR, BO2 and Crysis 3 MP Beta (those are the current games I remember).

So this issue started about 2 months ago. After getting too annoyed I choose to format the computer, reinstall Windows 7. Clean install. This is about 3 weeks ago now...

So today it started again. Crashing that is. A friend of mine told me that ASUS has some kind of mechanism which may cause the crash, but I don't really know why it's happening now.

The GPU and Motherboard is no more than 1 year old. Also my RAM and HDD was bought a year ago.

My CPU has been transferred from another computer, but I never experience any issues with it and still don't (I've also tested heating on the CPU, and it seems fine).

I tried to take the fan from the side (on the case), and put it in front to get a better airflow, but it doesn't seem to be strong enough to cool it properly (although it helps a bit).
- Also I cleaned all my fans, GPU and CPU coolingblocks as well.

And as mentioned. This started recently, so this shouldn't have anything to do with the airflow although it isn't the best.
My room-temperature should be and is alot colder than summertime (where these issues weren't to be found), yet I see my GPU reach 65 celcius.

I don't know if my GPU reaches anymore as I never seem to notice anything before it happens.

PC Specs:
AMD Phenom II X4 955 3.2GHz CPU
12GB DDR3 1333MHz RAM (Kingston)
Nvidia GeForce GTX550 Ti 3GB (Club3D)
2TB SATA HDD (Western Digital I think)
Windows 7 x64
Newest GPU drivers (updated today).

I hope someone out there may be able to help me as I am a gamer, livestreamer and YouTuber. But I cannot record my games until this works properly again.


Thank you in advance, and sorry for the long post :) .

More about : gpu cpu overheating

a b à CPUs
February 19, 2013 4:13:29 PM

to test your cpu for overheating get prime 95 for stress testing and core temp to see your temperatures

for gpu get furmark
February 19, 2013 4:15:39 PM

alvine said:
to test your cpu for overheating get prime 95 for stress testing and core temp to see your temperatures

for gpu get furmark


Thank you for your reponse.

I have already tested both as mentioned. And they don't seem to overheat. But I will give it a go again.
For how long should I run these stresstests? :) 
Related resources
a b à CPUs
February 19, 2013 4:19:02 PM

prime 95 should be run for an hour at least
for furmark 30 mins...
you forgot to mention your power supply specifications...it might be giving up on you
a b à CPUs
February 19, 2013 4:19:09 PM

Can you list your MOBO?
February 19, 2013 4:22:22 PM

alvine said:
prime 95 should be run for an hour at least
for furmark 30 mins...
you forgot to mention your power supply specifications...it might be giving up on you


Ok, I'll test that.
PSU = Energon EPS-750W


chugot9218 said:
Can you list your MOBO?

ASUS M4N75TD
a b à CPUs
February 19, 2013 4:25:16 PM

also you can check your ram for errors with memtest86...you download the iso, burn it to a disk and boot from it..if you get errors it means there are issues with your ram
a b à CPUs
February 19, 2013 4:25:23 PM

That PSU should probably be replaced, it's an off brand and although it may not be causing this issue it could cause even worse ones in the future. Also, I noticed your MOBO supports Dual Channel and you appear to be running 3 sticks, I would try to replicate the issue using only two sticks.
February 19, 2013 4:28:21 PM

chugot9218 said:
That PSU should probably be replaced, it's an off brand and although it may not be causing this issue it could cause even worse ones in the future. Also, I noticed your MOBO supports Dual Channel and you appear to be running 3 sticks, I would try to replicate the issue using only two sticks.


I'm sorry, but you lost me there.
My MOBO supports Dual Channel, and I appear to be running 3 sticks?
Do you mean RAM slots? - If this is the case. I could try to run it with 8GB instead of 12GB (remove the 4GB RAM Block)

Regarding the PSU. I had this one before with no issues and it never caused me any issues. But you are right, I may need to replace it at some point, but it wont be right now. Thanks for the heads up though! :) 
a b à CPUs
February 19, 2013 4:35:45 PM

Yes, you have 4 RAM slots and because it is built for dual channel, you should be using either 1, 2, or 4 sticks (although 1 does not take advantage of Dual channel it should still operate fine). Having 3 sticks in a dual channel board can degrade performance and cause errors, apparently triple channel RAM is the same as dual channel, just tested in triplets instead, so that shouldn't make a difference, but you should be using dual channel if your board supports it.
February 19, 2013 4:41:48 PM

chugot9218 said:
Yes, you have 4 RAM slots and because it is built for dual channel, you should be using either 1, 2, or 4 sticks (although 1 does not take advantage of Dual channel it should still operate fine). Having 3 sticks in a dual channel board can degrade performance and cause errors, apparently triple channel RAM is the same as dual channel, just tested in triplets instead, so that shouldn't make a difference, but you should be using dual channel if your board supports it.


Okay, I will take out the 4GB RAM block. It should be more than enough with 8GB anyways. Hopefully that will increase performance slightly during gaming (I didn't notice any performance upgrade when I upgraded from the 4GB to 12GB). Hopefully this will make a difference.

Currently I'm CPU stresstesting, and during Test 3, it reaches 70 degrees celcius. This shouldn't happen. Right?
My CPU limit it 90 degrees celcius. I would like to point out that the stock fan I am using apparently is *** (according to numerous posts on the web).
February 19, 2013 4:45:52 PM

PSUs tend to lose some of their efficiency a little bit over time. This is true especially for above 500W cheap brand PSUs like the one you're using. So this could potentially be the cause.

I don't really see how the 3 DIMMs of RAM could be causing the crashes if you've had them working fine previously (although of course they can become dysfunctional in use over time as well), but that has nothing to do with memory bandwidth lanes. To test your RAM you should run either Memtest86 or Prime 95 Blend using as much RAM as you can allocate for 24 hours or more. If you get no errors, your memory is fine.
February 19, 2013 4:50:17 PM

hizodge said:
PSUs tend to lose some of their efficiency a little bit over time. This is true especially for above 500W cheap brand PSUs like the one you're using. So this could potentially be the cause.

I don't really see how the 3 DIMMs of RAM could be causing the crashes if you've had them working fine previously (although of course they can become dysfunctional in use over time as well), but that has nothing to do with memory bandwidth lanes. To test your RAM you should run either Memtest86 or Prime 95 Blend using as much RAM as you can allocate for 24 hours or more. If you get no errors, your memory is fine.


If it turns out to me my PSU, the issue wouldn't directly be the PSU. It would rather be loss of power from the plug as I have quite a few things connected at this point. I may need to re-arrange my cables and use a secondary plug for some of my cords.

The PSU is no more than about 6-7 months old and shouldn't be causing this issue (as this issue started about 2 months ago).

Thanks for the input though. Wether I like it or not I will have to get a new PSU eventually. But currently I don't have the spare cash for it, unfortunately.

I just remembered that I bought the RAM a few months back as well, so they might actually be the issue.
February 19, 2013 5:15:30 PM

Ok, so it turns out that when I ran furmark. It was running for 2-5 minutes, and the heat just kept raising. How can I fix this? I never had this issue before, and suddenly it overheats?

I would like to point out that nothing else but the extra 8GB RAM has changed (and of course a format or two since I bought it).
a b à CPUs
February 19, 2013 5:33:10 PM

I'm gunna cast my vote and say a hand full of things need to be addressed.

1) 70C is on the hot side. Maybe you can try to re-seat ur CPU cooler with a good thermal compound (i like Arctic Silver 5). Or get an aftermarket one like the cooler master 212 hyper (cheap and one of the best!)

2) Test the RAM 1 at a time with memtest and prime95 - if you get no issues, than I think you're prly fine on the RAM

3) Your system isn't all that entirely power hungry, but I'm sure that it'll be reaching it's limits in all systems (except RAM) while playing the new games that you have been. That being said, your minimum power supply for that GPU is 400W, and you have a crappy 500W PSU that may not even be delivering 350W at peak efficiency. I suggest that you get yourself a new, quality PSU (500W-600W). I recommend that it also be 80+ certified and from a good brand, like Corsair or Seasonic (do ur research on good brands/actual parts manufacturers).
a b à CPUs
February 19, 2013 5:37:07 PM

eXino said:
Ok, so it turns out that when I ran furmark. It was running for 2-5 minutes, and the heat just kept raising. How can I fix this? I never had this issue before, and suddenly it overheats?

I would like to point out that nothing else but the extra 8GB RAM has changed (and of course a format or two since I bought it).


How hot? It is perfectly normal for GPU's like yours to operate at up to 80C at full load. Any higher though, and you should prly be concerned.
February 19, 2013 5:39:39 PM

You said you bought more RAM or you replaced your RAM entirely? You can't really just 'add' memory, even if it's exactly the same brand and type. Whenever you buy memory you should get a RAM kit; a set of DIMMs that have been tested and verified to work together. Add anything that hasn't been tested with those exact DIMMs you have, and it could cause potential instability in your system.
February 19, 2013 5:42:44 PM

eXino said:
Ok, so it turns out that when I ran furmark. It was running for 2-5 minutes, and the heat just kept raising. How can I fix this? I never had this issue before, and suddenly it overheats?


There may not be enough air flow in your case. If you GPU has one of those coolers that just has a fan blowing inside the case, and not a slot cooler blowing it directly out the back, it is just recycling the hot air in your case, and your exhaust fans may not be getting the hot air out quickly enough.
February 19, 2013 5:54:45 PM

Why do you think the issue is overheating? ~70C is fine for a GPU under load. What temp is your CPU running at under load? Anything under ~85C isn't great, but its safe. You can check your CPU temp by running CPUz.

Furmark will heat up your graphics card a lot more than any game will. Its actually possible to harm perfectly fine cards running that software.

Overheating isn't likely to cause crashes anyway. The CPU and GPU are designed to throttle down at high temps, and you'd notice stuttering and things in your games before you'd experience crashing.
February 19, 2013 6:43:36 PM

helz IT said:
Why do you think the issue is overheating? ~70C is fine for a GPU under load. What temp is your CPU running at under load? Anything under ~85C isn't great, but its safe. You can check your CPU temp by running CPUz.

Furmark will heat up your graphics card a lot more than any game will. Its actually possible to harm perfectly fine cards running that software.

Overheating isn't likely to cause crashes anyway. The CPU and GPU are designed to throttle down at high temps, and you'd notice stuttering and things in your games before you'd experience crashing.


My GPU reached 63C and crashed. Tested it with Furmark. So apparently it's not overheating?
My CPU reached 75C while running the Prime95 Blend mode for 28 minutes.

I haven't yet experienced any stuttering nor lag during gameplay. It just crashes as if it overloads perhaps?

The PC crash is only happening when stress testing the GPU with Furmark.


@ittimjones,

My PSU is 750W, not 500W? It's should deliver enough power to run my current system.
a b à CPUs
February 19, 2013 6:45:44 PM

Poorly constructed PSU's tend to supply a much lower wattage than advertised, that is a bad PSU whether it appears to be working fine or not.
February 19, 2013 6:50:07 PM

chugot9218 said:
Poorly constructed PSU's tend to supply a much lower wattage than advertised, that is a bad PSU whether it appears to be working fine or not.


Quote from myself ealier:
Regarding the PSU. I had this one before with no issues and it never caused me any issues. But you are right, I may need to replace it at some point, but it wont be right now. Thanks for the heads up though! :) 

So I know that it needs to be changed, but it is not the PSU. How can the PSU cause a crash when running Furmark and only when running Furmark (or other graphical softwares, like games).
February 19, 2013 8:13:21 PM

eXino said:

So I know that it needs to be changed, but it is not the PSU. How can the PSU cause a crash when running Furmark and only when running Furmark (or other graphical softwares, like games).


The GPU and CPU require more power to run at full load than they do at idle. That's why their temp rises when you run demanding programs (more power means higher temps). If your PSU is unable to supply the required power, it can cause your PC to crash. Not saying for sure its your PSU, but its a likely culprit.

February 19, 2013 8:27:47 PM

Do you guys have any ideas & tips for which PSU I should get? At a fairly low price that is.
February 19, 2013 8:46:21 PM

If you haven't, you might want to download and install the latest drivers for your mb/peripherals. Windows might be using it's own drivers for some things when the manufacturer drivers are better.
a b à CPUs
February 20, 2013 12:46:11 PM

You do not understand, it is listed as 700w, that does not mean it is providing it. A well built 400w PSU is practically guaranteed to output the rated 400w. The wattage is not the only important spec of a power supply.
February 20, 2013 1:24:38 PM

chugot9218 said:
You do not understand, it is listed as 700w, that does not mean it is providing it. A well built 400w PSU is practically guaranteed to output the rated 400w. The wattage is not the only important spec of a power supply.


But I do understand. But my PSU is listed as 750W and not 500W or 700W as you mention.
The fact that it doesn't deliver that about of power is basically irrelevant as it no doubt delivers more than the 400W. My setup (as far as I know) needs more than 400W to run properly.

Keep in mind that I run a quadcore CPU and not a dualcore.

I know my PSU is *** and properly doesn't deliver the amount of power it says it does. Yet I haven't had ANY issues with this one or the one I had before (Exact same model).

Yes, It's a cheap PSU (kinda), but wether or not you like it, I doubt it delivers less than 400W.

But my main question was... Why does my GPU crash when getting overloaded as it apparently "only" hits 70C tops (I rarely hit that mark). Overall it's about 60C while gaming unless it's very graphical and or very openworld games like swtor, skyrim etc.
a b à CPUs
February 20, 2013 1:32:18 PM

And a common response seems to be that the system is trying to draw too much power for the PSU to handle, causing it to crash.
February 20, 2013 1:50:27 PM

chugot9218 said:
And a common response seems to be that the system is trying to draw too much power for the PSU to handle, causing it to crash.


Ok, but why didn't this happen until 2 months ago?
I think it may be related, but not directly. I think I have too much power connected to one plug meaning the PSU doesn't get the proper amount of power from the switch/plug in the wall.
a b à CPUs
February 20, 2013 1:59:13 PM

Components can fail or degrade at any time, especially cheaply made ones. Tom's did a pretty good article about low quality PSU's recently, there could be a part going bad in there which is causing the problem.
February 20, 2013 2:08:49 PM

chugot9218 said:
Components can fail or degrade at any time, especially cheaply made ones. Tom's did a pretty good article about low quality PSU's recently, there could be a part going bad in there which is causing the problem.


Okay, well I might just test with a proper PSU. But as I walk talking to the shop inwhich I bought the GPU, they mentioned that the issue wasn't likely to appear from the PSU, but the GPU itself. They wanted me to sent it in to be tested using Furmark by themselves any other test softwares.

Before I do that, I do want to test with another PSU though to make sure I don't send away my GPU for about 4 weeks :) .

Again, thank you for your inputs.
If I have any further questions or find a proper solution, I'll post it here for others to see :) .
!