EVGA GTX 980 SC high idle temps?

K

killerc4

n00b
Joined
Jan 31, 2008
Messages
63
I just got my 980 in the mail and set it up for benchmark testing. The first thing I noticed is the idle temp is way higher than I read in reviews. Most reviews say 34c-ish and I'm getting 52c. Room temprature is 21c.

https://dl.dropboxusercontent.com/u/7637606/temps.jpg

When I run a benchmark, GPU usage goes to 100% and temp quickly caps at 80c and the PerfCap starts to throttle due to temp.

Is this normal?
 
That is definitely not normal. Are the clocks dropping down to 135/324 when the card is idle - your screenshot shows 1000/1750? Maybe there's a program running in the background keeping the GPU under load where it won't drop down to idle clocks and voltage. For reference with a custom fan profile in Afterburner my card typically idles between 34-36c @ Fan Speed 30%.
 
I closed all my extra programs and didn't notice a difference in clock speeds. I disabled my extra monitors and that seems to be what is keeping my clocks high. I run a 3 monitor setup and when I enable my third monitor the clocks go up. With 1 or 2 monitors however I was dropping down to 135/324 but my idle temp was 42c with 29% fan. A little high but much better.
 
Is that third monitor connected via DisplayPort by chance? Can you try the HDMI port? I had a similar issue a while back with the GTX 680.
 
Nvidia cards will not downclock once you have a 3rd monitor attached in my experience.
 
Kaos_Drem said:
Nvidia cards will not downclock once you have a 3rd monitor attached in my experience.
Click to expand...

My 970 clocks down to 135 with 2x DVI and 1x HDMI connections. I don't have any DP cables or I would give that a shot to see if DP is the problem.
 
i have 1 dp 1 hdmi and a dvi atm. waiting on a dvi to dp adapter to get rid of the hdmi
 
What drivers are you using? 344.16 should be the latest one with some fixes.
 
I'm on 344.16. Its down clocking now and I've let it idle for awhile but still getting 42c with a 30% fan. If you run gpu-z and a benchmark do you get the Perfcap Thermal reason throttle and/or does your card even reach 80c under load?
 
When I first installed my 980, it idled around 40c which I thought was high. Ran a few benches Fire,Heaven,Valley etc and after around an hour or so I noticed it dropped down to 30c idle. I figured maybe the tim had bedded?? JMO.
 
Idle temp is 52C? Might want to check the paste on that card.
 
No it is not normal.The problem is easy. It's called EVGarbage. You read the reviews which showed the EVGA card has terrible temps and noise, yet you still bought it? you deserve it buddy.
 
Seroman said:
No it is not normal.The problem is easy. It's called EVGarbage. You read the reviews which showed the EVGA card has terrible temps and noise, yet you still bought it? you deserve it buddy.
Click to expand...
]

uselessinformation-300x298.png
 
Seroman said:
No it is not normal.The problem is easy. It's called EVGarbage. You read the reviews which showed the EVGA card has terrible temps and noise, yet you still bought it? you deserve it buddy.
Click to expand...

It's a reference card, with the only difference being clockspeed (all nvidia reference cards are the same, just different brand). Jesus, some people need to do their reseach before posting useless info.

n=1 said:
Idle temp is 52C? Might want to check the paste on that card.
Click to expand...

Agree.
 
I made this account just to put this in this thread,Quite a few other people are having this issue as well my card idles at 40c and I'm pretty sure its because the voltage is locked at 1.05v all the time and the clockspeed is cranked even when the gpu usage is at 0%.download gpu-z and see if your card is doing the same maybe we got a bad batch of cards or something.
 
Software/driver issue. I've had that happen with many cards, especially when running multiple monitors (144hz for example).

Uninstall evga precision and afterburner (do not save profiles), and download nvidia inspector. Should give you an option for multi monitor power saving.
 
I have a replacement on the way. So I guess ill see if there's a difference with the new card. I'll try out all the suggestions in the meantime. Thanks guys
 
I have this exact same card. I'll have to remember to check my idle temps for you.
 
Seroman said:
No it is not normal.The problem is easy. It's called EVGarbage. You read the reviews which showed the EVGA card has terrible temps and noise, yet you still bought it? you deserve it buddy.
Click to expand...

You need a reading lesson. His card is reference, every manufacturer sells the same one.
 
mls1995 said:
I have this exact same card. I'll have to remember to check my idle temps for you.
Click to expand...

OK so I just checked my idle temps and I'm getting 30 C on the nose in my Corsair Air 540 case. So obviously 52 C is bullcrap.
 
OK so by my count you have 3 people (including me) stating their card idles around 30 C. So that's what you should expect give or take
 
gamingarena23 said:
My both cards run at 32c top and 30c bottom idle.
Click to expand...

Are your cards SLI'd in the slots directly next to each other? Just interested. That's the only way I can do SLI with my mobo so I just thought I'd ask in case I go SLI some day.
 
All the reviewers all idle at 30c also but I'm at 5 or so for the amount of people idleing 45-50c a few people from linustechtips are having issues also.
 
Well it's good that you're getting a replacement, but 42C idle is still on the high side. If I had to guess the card had a sloppy paste job.
 
There must be something running in the background. Does it do this right after you boot into Windows and haven't opened programs yet?
 
I get this issue if I browse to a website that uses a lot of Flash. For example, some websites display PDF documents using Adobe Flash Player and the GPU will show almost exactly the same numbers as robthatguy's screenshot. (Could be GPU acceleration being enabled in Flash)
 
I have two EGVA 970 SC blower model in SLI. My idle temps were low 50s and core clock was always running around 1100. But when I plugged in a second monitor the core clock went down to 300s and idle temps went down to 39-40. My first moniter is a Rog Swift monitor connected via Displayport. The second moniter is a 120hz ASUS connected via DVI-I.
 
I'm back on my 660ti and idle sit at 32c so no backround programs I'm just rma'ing my card also.Its a pain the majority of us paid for overnight air shipping and we need to rma :(.I tried everything and when upgrading from a 660ti to a 980 and expecting an overclock it sucks to have this happen.I emailed Zotac twice and got no real response and Newegg blocked my review on the card.Will be the last time ordering from them they've gone downhill quite quick the past year.
 
The idle temps on my EVGA GTX980 SC cards in SLI is 35C for the top card and 39C for the bottom card.
 
I ran into this exact same issue, but was able to fix it for myself. I exhausted every single available option with the exception of OS re-install and RMA. My card itself has been running fantastically and has an additional 80mhz on top of the factory OC and 400mhz more on the memory as well.

The change that ended up fixing my issue was a change of refresh rate. At 144hz my GPU would idle at 1ghz, memory would be maxed out and voltage would be up, but after lowering to 120hz or lower (currently at 120hz) the card began to downclock/lower voltage correctly. Obviously you should also make sure that the power management option in the control panel is set to the default of "Adaptive" and not "prefer maximum performance", but its worth mentioning to make sure, as those are the only two settings that raised my idle clock speed and voltage.

I am not in an open air scenario like most reviews would be, and I do have a GTX680 being used as physx right below my GTX980, so my temps are not going to be quite as low as others, but i sit right around 35 - 36 (always sub 40) after finishing a session of a game. It is closer to 30 on first boot before any 3D usage. I also have a custom fan profile that keeps my fan at 40% minimum and 80% maximum (depending on temp of course), but that same profile is used regardless of my settings.

TLDR: 144hz caused my problem, changing it to 120hz or below fixed it. Also make sure to leave "power management" in the control panel to "Adaptive" no other fix worked after trying every plausible. Hopefully future drivers will correct this issue. Cheers!

EDIT: Another suggestion in order to get 144hz in games would be to go ahead and set your desktop refresh rate to 120hz so you can manage to downclock/lower voltage when not in game, and then in the "manage 3d settings" / "global settings" area of the nvidia control panel, set preferred refresh rate to "highest available". That way when in desktop you will be at 120hz and idle correctly, then when in game you will be at 144hz and your card will move up speed like it is suppose to.
 
Last edited:
***The nVidia rep is going to look at this thread, so be sure to report if 144hz is causing your issue***

I am separating this reply as it is pertinent, and will update as the information comes in.

I have a dialog going with support over at nVidia. I wrote into them explaining the issue, as I wanted them to be aware of what 144hz was causing the clockspeed/voltage to do.

Initially the support rep believes (off the top of his head) that this is normal behavior, and was so with Fermi/Kepler as well, but he isnt certain and is going to follow up.

I think we all mostly know this to not be the case, but it has been so long since I have monitored my 680 that I honestly do not recall. I however, have my doubts that 144hz requires significantly more horsepower to push than 120hz. I would put it on the cable more than anything else to be honest (as far as hz capability that is).

EDIT: Another suggestion in order to get 144hz in games would be to go ahead and set your desktop refresh rate to 120hz so you can manage to downclock/lower voltage when not in game, and then in the "manage 3d settings" / "global settings" area of the nvidia control panel, set preferred refresh rate to "highest available". That way when in desktop you will be at 120hz and idle correctly, then when in game you will be at 144hz and your card will move up speed like it is suppose to.
 
Last edited:
Blazemore said:
My idle core is 1000mhz with using 2 dvi and hdmi ports. Temp 45- 50c
Click to expand...

What monitor and refresh rate are you running at?

All,

Could you please edit your post or add a new one stating what monitor/monitors your running, what refresh rates, and what cables your using to connect with? As well as, if this issue is affecting you or not. Lastly, state if changing the refresh rate had any success for you, so we can try and help nvidia know if there is indeed a problem with some monitors/cables/refresh rates/drivers, etc... Thanks!
 
Last edited:
my GPUs idles at 135MHz 850mv but it idles high too.
44c the first GPU, 39 the second one.
 
I got here because I had idle temps in the 50+C range. Turns out I had turned on KBOOST in the EVGA PrecisionX software. That forces full speed on the chip regardless of what's happening on the machine. I turned it off and idle temp was 29-30.
 
You must log in or register to reply here.
Back
Top