Very weird temperature issue (GPU)

bamboozled23

Reputable
May 7, 2015
12
0
4,510
So, I started having this weird problem since 5 months ago. The GPU when watching videos in fullscreen, playing games and viewing streams reaches 99% usage and the temp shows 83 degrees Celsius. When this happens, the fan starts speeding up (which should be normal). The thing is, when I quit any of those activities the usage goes immediately at 0% and the temp at 60s, which is really, really weird (it is not possible for the temp to drop that drastically in only a second). To me it seems that something is causing the card to be confused as to what is actually happening. I am desperate.

I have tried clean installing drivers, cleaned just now the dust (which was minimal), applied paste both in the GPU's heat sink and CPU's, opened the case just so it can have more room to breath and recycle the air, but none of these things worked.

I am using a R7770 GHz Edition (the 1GB GDDR5).

Please help me! This is my first time posting in the forum, so feel free to ask me anything extra.

Edit: I think the usage is not something bad, so the real problem lies in the temperature..
 
Solution
The HD7770s are about 4 years old. They never really ran cool. nVidia cards always have.

A new power supply won't help. If it keeps under 85C then that's reasonable.

More ram won't help it keep cool either.

I'd leave it till you can save and get something better.

mishima1992

Reputable
Sep 15, 2014
7
0
4,520
it's normal when you play games, gpu load reach 100%
but i dont know about your "normal" gpu temp
my r7 260x only 60-70 degrees when playing online games. 40 degrees at idle
 

bamboozled23

Reputable
May 7, 2015
12
0
4,510
Yes, I know the usage is normal.

About the temps : my idle is 30ish to 40ish, But as I explained when I launch a game, or fullscreen video, or stream, it skyrockets almost instantly (after 2 minutes or so) to 83 degrees and the fan starts. As soon as I exit whatever I am running at the time (be it game, full screen video or stream) the temp drops immediately (after max 5 seconds) to 60ish. That dramatic drop is not possible. I am thinking that something is going wrong and the GPU thinks the temp is 80ish while in fact it is 60ish.

Am I explaining it enough? I don't really know how else I can express what is actually happening..
 

Damn_Rookie

Reputable
Feb 21, 2014
791
0
5,660
It may be surprising, but the temperature really can drop that much that quickly. Just did a quick test on mine, running a stress test for about 1 minute (letting the temperature reach 70 Celsius), and then turning it off. The temperature dropped from 70 to 55 in the space of a few seconds, with the drop to 60 being almost instant. After that the temperature drops more gradually, but that initial drop is very fast.
 

bamboozled23

Reputable
May 7, 2015
12
0
4,510


What would be the best way to show you my build? Dxdiag?

About the the way I measure temps, first time I saw this happening through CCC. After that I installed Afterburner and it showed the same things, so I guess it's not a problem of how I measure them? If you want to recommend any other way to do this, I am all ears.

Thanks for replying.
 

bamboozled23

Reputable
May 7, 2015
12
0
4,510


Alright. Let's say it does indeed do so. What about this starting to happen the last 5 months or that the GPU gets stressed to 83 degrees by just watching 360p full screen videos or streams? (I am not mentioning games because many games could be able to stress a GPU that much).

Thank you for replying :)
 


Just list your components.

What temps do you get from Afterburner - for your GPU?

 

bamboozled23

Reputable
May 7, 2015
12
0
4,510


Windows 8.1 Pro 64bit
CPU : Intel Core i3 530 @ 2.93 GHz
RAM : 4GB
GPU : MSI R7770 series

Right now, ive been using the pc for like 3 and a half hours and I am idling at 50 and rising. Also I do not know if it should be happening or not, but I am seeing random spikes to the memory and GPU clock accompanied with small temp rises. Also I have 0mV apparently. (by the time I finished writing this text the temp has gotten to 60)
 
Components list usually looks like -
CPU: Intel Core i5-4690K 3.5GHz Quad-Core Processor
CPU Cooler: Cooler Master Hyper 212 EVO 82.9 CFM Sleeve Bearing CPU Cooler
Motherboard: ASRock Z97 EXTREME4 ATX LGA1150 Motherboard
Memory: G.Skill Sniper Series 8GB (2 x 4GB) DDR3-1866 Memory
Storage: Seagate 2TB 3.5" 7200RPM Hybrid Internal Hard Drive
Video Card: MSI GeForce GTX 970 4GB Video Card ($424.95 @ Vuugo)
Case: Corsair Graphite Series 230T Black ATX Mid Tower Case
Power Supply: XFX 550W 80+ Bronze Certified ATX Power Supply
Optical Drive: LG GH24NS95 DVD/CD Writer
Operating System: Microsoft Windows 7 Home Premium SP1 (OEM) (64-bit)
 

bamboozled23

Reputable
May 7, 2015
12
0
4,510
Im sorry about the list. I am pretty new and I did not know what would be enough, in the original post I have specified the exact specs of the GPU. Also, I am not really sure what manufacturer my PSU is from, but it is 450W.

Gaming temps (also fullscreen video temps, stream temps) ~83 degrees (which is also the point the fan starts).

Since the temps are not weird, how come I didn't experience the loud fan the past years but only started experiencing it now? That is why I am confused, if it was acting that way from the beginning I would be okay with it, but to me this activity of the GPU was not a common thing until recently.
 

bamboozled23

Reputable
May 7, 2015
12
0
4,510


After I increased the fan speed to 80% (maximum speed for minimum noise), the temp has gotten to 40ish idle. When I went in game the temp got to 90ish and the fan did not start to cool it down. I had to do it manually. 50 degrees is too much, just, too much..

Edit : Underclocking the card a bit seems to have a positive effect and I set up the fan curve to speed the fan up after 80 degrees. But the question still remains. Why did my card, in the first place, start showing such high temps? (5 months ago it did not ever reach 80 degrees. no matter the game, no matter how much I played or whatever I did)
 

bamboozled23

Reputable
May 7, 2015
12
0
4,510
Hm, I only bought the card 3 years ago, plus I reapplied paste on the heatsink.

So, a PSU with more W would do a better job? Maybe adding 4 extra MB of RAM would help it at all? (at least temporarily, until I have the budget to get a new motherboard and CPU + GPU?)
 
The HD7770s are about 4 years old. They never really ran cool. nVidia cards always have.

A new power supply won't help. If it keeps under 85C then that's reasonable.

More ram won't help it keep cool either.

I'd leave it till you can save and get something better.
 
Solution