New Computer, new EVGA GeForce GTX 970. Can't stay stable in Witcher 3(25Euro/Dollar reward)

Flagg · Jan 26, 2016

Update 27/1 II
Ok with lots of knocking on wood and all that, system seems to be running stable now. I did lots of weird, small things(unplugging everything, switching to DVI-l, removing any and all extra electrick.. things plugged in same hub with PC and monitor. Also updated Bios. Yesterday evening, I was...very alarmed to notice there was clear static electricity happening on wires of my Headphones. Relocated them from back/motherboard to front of the PC. Currently everything seems alright. Not using PrecisionX either. Comp not sounding like it is a helcopter either, anymore. If things stay like this, I just might be happy!!

Thank you very much for one and all. Gonna have to wait and see.. Maybe 45 mins stable now..

UPDATE 27/1, Old issue/post still current I fear

This is how my Precision-X looks like when idle and on Dekstop: http://imgur.com/KNwAXdG

With great measure of knocking on the wood and stuff, I may have gotten things running stable. Seems keeping PrecisionX up and running, and working HARD did the trick. Now my fans are working so hard I recon I have a quieter vacuum cleaner than a PC.
Here is precision-X Overlay and curve after 30 mins of Witcher. http://imgur.com/IIdElpq
..Hm, since Precision X has " auto" switched on, does it even care of my curves running some default instead perhaps?

*** OLD POST BELOW, STILL a Current issue***

Hello,
Brand new computer with brand new parts. Using Windows 10. Using current official Geforce experience drivers for the GPU.

After few minutes of Witcher 3, my monitor(and display card, I assume..) are turned off. Only way to bring them back alive is forced power off/back on of the computer. Game itself doesn't seem to crash; sounds all keep playing all right even after the screen goes dark.

- -
SPECS:
Motherboard: ASUS Z170 Pro Gaming
CPU Intel Core i7-6700
GPU: EVGA GeForce GTX 970 Superclocked ACX 2.0 4 GB
Memory: Kingston HyperX Fury Black 2 x 8 GB
Power: EVGA SuperNOVA 750 G2 (80Plus Gold)
Cooling: : SilentiumPC Fera 2 HE1224 v2
- -

I assume these symptoms are caused by some issue with my power or display card. I spend six years of gaming without any difficulty with my previous computer so I'm completely lost in anything having to do with cooling or custom fan management.

I'll pay 25 Euro/Dollars for a detailed, good answer
that A) assumes I know nothing about finer bits of customizing and operating cooling, fan speeds and such
and
B) fixes my problem and results in me being happy owner of a stable healthy system that runs W3 without killing my GPU or CPU or anything else.
Pref Paypal for this

It feels very important for me to fix the disease, not the symptom here. Ability to run W3 isn't some be all end all here, but it feels very important to understand WHY I am unable to keep my system stable during it and sort the cause out.

According to HWMonitor, my GPU temperatures are approaching high 70's in C slightly before the GPU goes off. I've never seen it at 80c though. Elsewhere, I got told this is fairly high in-game reading for the card. I'd tweak my settings via EVGA precision X but I have no idea how to do so in optimal fashion in situation where the default settings of my card don't appear to work. Moment I turn Precision X on, it causes my fans go wild. Sound sounds healthy and everything, but they begin spinning far faster than they would when I do not have Precision X on. Is this normal? If I try W3 with Precision X and my fan curve on, fans go absolutely INSANE. Again, sound feels..healthy, they just do crazy RMPs.

Here is a link to my HWMonitor stats when system is in desktop, normal browsing. EVGA Precision X is not on. I'll send more pics very soon:

http://imgur.com/sNwxEqP

UPDATE: After some 30 mins in Star Wars, The Old Republic ( MMO from 2012) with ultra settings, system is stable, and GPU temp foes between 70c and 73c.

The GPU comes with factory-overclocking. I have done no overclocking myself. CPU is not overclocked either. Would be very grateful of any and all help here...these are sad issues to have with pretty good, brand new computer I could barely afford:/

blazorthon · Jan 26, 2016

That's odd. If the game keeps playing, then I'd think that the graphics card is not at fault and it is a problem with the monitor or the power supplied to the monitor. Do you have another monitor to test this theory?

Temps approaching the high 70's C are a little high, granted the monitor turning off is not a common symptom of an overheating GPU. That usually either causes the card to throttle performance (game continues playing, but with lower performance) or emergency shutdown (the game would not continue playing).

EDIT: I should note that a some crashes can keep the audio going in which case the graphics card may be at fault. However, that wouldn't be caused by the GPU overheating. It is possible that another part of the graphics card, say the VRM or memory, could overheat and cause weird things like this to happen because those often have insufficient cooling and no temp sensors to detect when they get too hot.

That also makes them hard to diagnose with certainty. At best, you could only then narrow it down to the graphics card being at fault if the same thing happens when the card is tested in another computer and then you have to RMA the card or something like that. Do you have another computer to test the card in?

Other possible causes are pretty much any other component having a problem, most likely (but still less likely than a graphics or monitor problem) would be the motherboard, RAM, and power supply. None of the components you listed are low quality, but even high quality parts fail sometime so we can check them too if the above things test out ok or you can't test them.

For example, you could try running Memtest on each memory module (you have to test each module one at a time when doing this) and you could try using a different power supply if you have one handy.

king3pj · Jan 26, 2016

I have 2 of the exact model EVGA 970 you are using in my system. The EVGA SC ACX 2.0 970.

I have no clue as to why you are experiencing the problem you are reporting and even if I did I would not accept the money.

I just wanted to chime in to report that my temperatures also reach the mid to high 70s on one of my cards so it may be normal for this model. My other card also reaches the 70s but is usally 4 or 5 degrees cooler.

I run SLI so it makes sense that my top card would be a few degrees warmer but even my bottom card is pretty close to the temperatures you are experiencing.

Edit: If you want to bring the temperature down a bit to see if that is what is causing the problem download EVGA's PrecisionX software and switch to the "aggressive" fan curve.

This will make your card run significantly louder but it will also make it run cooler.

Flagg · Jan 26, 2016

Thank you very much for your answers. I can only hope there is some solution to this somehow. Unfortunately I don't have a computer available where I could stick this card in, or rather it is so damn old that I'm not sure I could get any functionality or reliable results. I couldn't run W3 on it for example. I'd be really happy for fan settings that keep my temp below 70s and don't make my fans sound like they are going completely insane.

I find it very unlikely that it'd be problem with monitor. It doesn't do and hasn't done anything..strange under normal conditions.

What are the largest downsides of super fast fan speeds? Or rather, how large are the down sides obvious for all;p Besides noise, how quickly will high speeds kill fans? Does it actually end up being notable dent in electricity bill? I realize there are no accurate answers to these questions.

blazorthon · Jan 26, 2016

Spinning the fans up faster shouldn't hurt their failure rate much and these small fans don't use much power. The only real drawback is noise, but maybe there is a point where you can spin them up only somewhat faster to lower temps somewhat without it sounding like a jet parked in your backyard.

This link my help you with the fan curve:
http://forums.evga.com/Precision-X-Recommended-Fan-Curve-m1932635.aspx

king3pj · Jan 26, 2016

I don't know. I am much happier letting my 970s run in the mid 70s than the loud noise that comes with the "aggressive" fan curve preset.

I don't think that running the fans at higher RPMs will damage the card or have a large impact on your electric bill. I suppose the fans could wear out faster though.

Keep in mind that these cards are designed to run in the 70s though. They are prioritizing noise over temperature as long as you are within the safe range.That's why the GPU doesn't crank up its fans to full speed by default. They don't do that unless your card reaches much higher temperatures.

When you reach 80 degrees the card just stops running the boost clock to bring it's temperature down instead of cranking up the fans to max.

Also the fans don't run at all until the card reaches 60 degrees. That means that while gaming you will almost always be somewhere above 60 and below 80. Only games that aren't graphically intensive will be able to stay below 60 without the fans running.

You can experiment with creating your own fan curve though. Just go to custom and drag the sliders up and down. If you want the fans to run at 20% when your card reaches 40 degrees you can do that. Just gradually go up the scale and then compare your noise to temperature levels while in game and find settings that work for you.

sleepybp · Jan 26, 2016

That is definitely a let down to have crashing on a high quality new build such as yours.
For the process of elimination I would do the following: and you can always undo some of these later after stable

1.downclock the video card gpu clocks and memory speeds back to default gtx 970 speeds ( i know this is factory overclock but hey try it)
2. Increase the fan speed manually for the video card to a higher setting
3. Remove one stick of system Ram and test Witcher 3 game for stability ( also verify what voltage the ram is getting in motherboard bios and compare to ram specs on box)
4. Re-test Wither 3 game with only the other single stick of system ram for stability
5. Monitor cpu core temperatures during either gaming or witcher 3 tests, I know they were low at idle but double check where they are at on a good load
6. If it is still unstable I would recommend shipping back the video card for replacement

Good luck and dont pay anyone a reward on here we like to help

Flagg · Jan 26, 2016

Many thanks for the kind and well thought replies all. I didn't expect there to be some easy solution for this, but was of course hoping for one anyway. You know, some " ahh, this switch right here in depths of PrecisionX or Nvidia settings, pull it! ..or some such. I suppose I'll try to re-create the situation in some environment other than Witcher 3. Assuming something is very wrong with my card, or motherboard or something, would it be risky to use some Stress test diagnostics tool?

Flagg · Jan 26, 2016

Unfortunately I was able to produce the same thing in Star Citizen Alpha as well.

Skype call with my friend stays on. Game (sounds like) it stays on. Screen turns off. Trying to take screen off electric network and plugging it back won't solve anything. Only hard reboot of comp helps.

SkyMembrane · Jan 26, 2016

Flagg :

Without Precision X being active, EVGA cards don't activate any fans until they reach 60C. The reason why they go mental when you open the program might be because of the Fan curve or it not being set to auto and the fans are running at a high RPM. I'm using the same card and I noticed that when the program opens, the fans reach a high RPM for like half-a-second and then settle back down, which might also be what you're experiencing.

Your temperatures aren't too high, technically, though you could probably set a different fan curve and keep them lower. Your room's ambient temperature and the case's overall air flow also have an impact on your GPU's temps.

I would recommend that you use Geforce Experience recommended settings and try to run The Witcher 3 again. I'm using the same card and I don't seem to have any stability issues with those settings, so a comparison would be good to make. Have you used the installation disk that came with the card? I would guess so, but I think there were a couple of important things to install from it. It's been a while since I installed it.

Leave Precision X on at all time with this card, because EVGA has weird software settings.

You could also try and set Windows 10 to performance mode instead of balanced, which would pull as much power out of your component as they can give.

Don't touch the overclock, because if the card is broken, OC'ing it will void the warranty.

king3pj · Jan 26, 2016

You don't need to leave Pricision X open at all times. There is a setting to make your custom fan curve load up when Windows loads.

You also don't need to install anything from the disc. I have 2 of the exact same model in SLI and all I did was install the latest drivers from GeForce experience. There is nothing weird about the EVGA software settings as far as I have seen.

SkyMembrane · Jan 26, 2016

king3pj :

I've been using the same card, and while it does have the option to load when windows loads, it's been extremely inconsistent and sometimes just stops working until Precision X is loaded again. Obviously something between your situation and his is different, and if he doesn't try anything, how could he find what the issue is? I didn't run into any problems with my card either, but I'm not using the same set-up. Of course if you have any better ideas, go ahead and give him something. The temperature is obviously something to watch, but that doesn't tell him why it's so high, even when the fans start. I don't go over 65 Celsius when playing The Witcher 3 with a single card, and it's OC'd at +150Mhz on the core and 700Mhz on the memory. It shouldn't be running colder than his stock version, even if the ambient temperature is lower.

Setting up his fan curve is the first thing he should do, while leaving Precision X open, when his problem is fixed, then he can go and tweak the settings so he doesn't leave it open at all time. From personal experience, it's more stable to leave Precision X open, but that's a complete other story.

As far as the software goes, EVGA focusing on silence until the card reaches 60 Celsius and starts to make way too much noise is ridiculous. The card can run better with a custom fan curve that starts at 15% and goes up gradually, and 15% isn't loud enough to hear if you have any other fans running.

king3pj · Jan 26, 2016

I have tried to help in previous posts. I don't know what the issue is.

As for the part about the fans not spinning until the card reaches 60 degrees, that isn't really an EVGA design. Almost all GTX 900 series cards have that feature regardless of manufacturer.

17seconds · Jan 26, 2016

Don't run two GPU monitoring applications at the same time, i.e. HWMonitor & PrecisionX.

I'd be checking your power supply to see if that's the problem, after that, make sure the card and the power cables are securely connected.

Mid-70c's is very normal. In fact, the GTX 970 actually targets 80c as it's factory preset target temperature. It will adjust Turbo Boost clocks, fan speeds, voltages, etc. to target and maintain exactly that temperature. The fact that you have a very effective ACX cooler is the reason you aren't quite hitting 80c, but that is how the card is designed.

http://www.techpowerup.com/reviews/EVGA/GTX_970_SC_ACX_Cooler/29.html

Flagg · Jan 27, 2016

Thank you for all the continued attention to this issue folks. I'm trying to read up on all the suggestions/help/theories provided. I un/replugged everything external, changed to using the DVI-I port and did few supersitious things like plugging off my headphones.

This is how my Precision-X looks like when idle and on Dekstop: http://imgur.com/KNwAXdG

Is it safe to keep HWM and Precision-X installed and even run them during same boot of computer, as long as one doesn't run them at same time?

I've heard updating Bios(motherboard) is always kinda iffy and risky business. Manufacturer provides a ton of different small and larger updates to mine, from stuff like hotfix to product registration. How and where do I see if relevant bits are up to date?

With great measure of knocking on the wood and stuff, I may have gotten things running stable. Seems keeping PrecisionX up and running, and working HARD did the trick. Now my fans are working so hard I recon I have a quieter vacuum cleaner than a PC.
Here is precision-X Overlay and curve after 30 mins of Witcher. http://imgur.com/IIdElpq
..Hm, since Precision X has " auto" switched on, does it even care of the curvers I'm using, is it instead running some default version of it's own perhaps?
I keep wondering if this is just treating sympthom, not the disease. I reallize the card should be perfectly stable and operational with temps like 76-80 or so. I am not exactly calm and content with idea of having to run some exceptional and " unantural" fanning solution to keep everything working alright.

rcald2000 · Jan 27, 2016

Question #1: What are your CPU temps during load?

Question #2: Have you tried documenting your precise display settings? If it were me, I'd unplug the monitor power for 30 second, reconnect and then do a factory reset from the display OSD.

Question #3: What is the monitor, computer, laser print (if any) plugged into? normal surge suppressor, battery backup, what? I just want to be sure that this isn't an issue with the display itself going into power save money via Windows or the monitor itself.

Flagg · Jan 27, 2016

^

#1 CPU Temps haven't given me an alarm. Maybe 60c at most, usually bit less.
#2 Nope I'm unsure what exactly you mean. WHich settings exactly are you after?
#3 Using a Belkin surge protector, this isn't the exact model but very close in ballpark: http://www.oodeys.com/resource/image/BELKIN-SurgeMaster-Superior-27646.jpg

Yesterday I actually removed everything except computer and monitor from the vicinity, plugging them in elsewhere instead.

Paul NZ · Jan 27, 2016

Have you updated the BIOS on this since you built it?? That's up to 1105 as of last week.

Because some of the Skylake CPU's have a bug, it'll freeze under load.

The CPU that was affected is the 6700K. But this bug may also affect other Skylake CPU's

So, it may pay to update the BIOS, if you haven't yet.. To see if that fixes it

If you want to test if this is the prob follow this

Download Prime95 28.7

rcald2000 · Jan 27, 2016

#1: 60 C wouldn't cause an issue, so we can rule that out as a cause of instability.

#3: You ruled this out as an issue, since you plugged it in elsewhere.

#2: Sorry I wasn't clear. I was asking if you've tried resetting the monitor from it's on screen menu. I'm wondering if the screen is turning off due to some power save setting. Resetting the display will remove your customized color and brightness settings, hence my remark to write them down.

leifus · Jan 27, 2016

This sound alot like what i am experiancing with my 780 SLI, but i have overclocked them. When that happens to me i still have sound and screen turns black. the reason why this happens to me were cuz Nvidia drivers crash. to fix this problem i had to bump my voltage on my GPU and then it went stable.

king3pj · Jan 27, 2016

Another option is to RMA the card and get a replacement from EVGA.

New Computer, new EVGA GeForce GTX 970. Can't stay stable in Witcher 3(25Euro/Dollar reward)

Honorable

Glorious

Distinguished

Honorable

Glorious

Distinguished

Distinguished

Honorable

Honorable

Honorable

Distinguished

Honorable

Distinguished

Champion

Honorable

Splendid

Honorable

Polypheme

Splendid

Honorable

Distinguished

Share this page