Did Elite in VR kill my GTX 1080?

Darkh4 · Oct 18, 2016

So I just got my Oculus CV1 two weeks ago, and after getting everything set up properly, I've been loving Elite in VR.

But today as I was transporting my Type 9 to the star system that I've been hanging out in lately, the display just abruptly cut out. No display at all, in either the head set or the monitor. Had to do a soft reset of my PC to get it back. Now, my display cuts out whenever I try to even launch the game in VR. Everything works fine when I boot up other games, like Deux Ex:Mankind Divided, or the Witcher 3. The Oculus Dreamdeck also runs with no problems. And I'm not seeing any temperature issues in the GPU.

But now Elite can only be opened in 2D monitor mode without crashing my graphics card, and even then my card starts making weird whining sounds that seem unrelated to the fan speed, and the sounds change pitch depending on which options screen I have open. This does not happen running anything other than Elite.

I'm running the 373.06 Nividia drivers, with the Oculus App Version 1.9.0. The rest of my rig is as follows:

Asus Turbo GTX 1080
Asus Pro Gaming/Aura Z170
16 GB DDR4 Corsair Memory
Intel I7 6700K Skylake
Samsung 1 TB SSD
Corsair RM750x PSU
Windows 10 Home

Is my 1080 all done running Elite in VR, or is there something I'm overlooking?

Thanks, CMDRs!

RobLocksley · Oct 18, 2016

I see it says Asus Turbo, overclocked from factory? Try underclock it; set it up like a vanilla 1080 and see if you can use VR.

Got a Titan Pascal that I have watercooled, I OC it to about 2100 with turbo and if I get a problem with my computer (when I OC the CPU too much) I take it down as my powersuply is getting a bit old (850 signature Astec).
So it could be your PSU too if you are pushing it...

The whining is probably a cap vibrating working hard and has been linked to Ramdac in some cases, so yes it can be your card is faulty.

Hope you get it sorted.

KellyR · Oct 18, 2016

How old is your system? Or if the graphics card isn't original equipment, how old it it?

Darkh4 · Oct 19, 2016

The card does have an automatic overclock, but it's not very extreme. I think by default it is set to 1734 or so. And up to this point it has been fine. I will definitely try setting it lower when I get done with work, though I don't know why it would work fine for a couple weeks with E : D, then suddenly quit.

The system is about 4 weeks old, though graphics card was pulled out of a previous build and is about 2 months old. The PSU is four weeks old like the rest of the system, and I would think is adequate to the task.

It's bizarre that it only seems to have trouble with Elite. Other highly-demanding games like the Witcher 3 set to Ultra settings work just fine. No whining sound or anything. And other VR apps work properly as well. But Elite in VR crashes my video output, and Elite in normal monitor mode causes the whining sound.

Other than lowering the overclock, I'm not really sure what my next step is going to be....

Thanks for the suggestions, CMDRs! o7

[HJ]-RedRaven · Oct 19, 2016

I'm at a bit of a loss to explain what the whining sound might be. If its not a fan bearing, which can happen even on new cards. Some electronic components do emit a high-pitched hiss or whine - (not everyone can hear it); check all the other fans too, tighten screws and re-seat cards if necessary.

My GTX1080 changes noise a little when I enter stations, as it switches to ASW at 45fps. It cools a bit and the fans slow down slightly. I only noticed it the other day as the wife was talking to me as I was landing so I was holding the Rift up to see where I was going but could hear the GPU fans change pitch.

It maaaay be worth re-installing ED if its just that program having an issue.
Or reinstalling the graphics drivers, even if they are up to date. Use the Guru3D driver cleaner.

Good luck - let us know how you go!

Darkh4 · Oct 19, 2016

I think RobLocksley is probably on the right track about the vibrating capacitor. I can crank the fan right up manually in GPU Tweak, and the sound is not produced. And it only happens in Elite, instantly changing pitch when I open different game option menus. It's an abrupt change, not like the gentle rise and fall of a fan. It never gets to produce the sound when the Rift is enabled, because the display cuts out within a second of the first splash screen opening.

The weird sound and sudden failure makes it seem like a hardware issue, but the fact that all the problems only manifest with one game is suspicious. I find it difficult to believe that Elite's splash screen puts a heavier load on my GPU than actually playing The Witcher 3 or Deus Ex:MD at max settings.

I will try the reinstall of my graphics drivers and the game, in that order. My internet connection is abysmally slow (3 Mbps), so reinstalling the game will probably take a couple days.

Thanks again!

Monoceros · Oct 19, 2016

Managed to do this myself and was due to pushing things to far, overclocked the GPU/CPU and running the game at max settings and debug tool set to 2. Thought I had killed the rift for a few secs, need a full power reset on the PC and back to default clock settings and game back to low VR via EDProfiler. Slowly built it back up, tweaking various game settings and GPU/CPU overclock settings and can now run well.

Darkh4 · Oct 19, 2016

I have just returned from work, and to my surprise everything is working again! Elite loads and runs on the Rift, and there's no whining sound. But I didn't change a thing. The only thing that happened is that my rig sat powered off for 12 hours. That almost makes it seem like a thermal shutdown, but my GPU temp rarely goes over 70 degrees C under heavy load in Elite with VR. It's not even supposed to start throttling until around 80.

I don't know, maybe I was pushing it too hard, like Monoceros suggests. But it had been running fine in it's current configuration for days with no indication of a problem. I had the debug tool set to 1.25, Super Sampling at 1.0, most settings on high or ultra, but with some strategic ones set to medium or off. I was getting a steady 90 FPS in space, with ASW kicking in at stations and planets.

I guess I'll have to start dialing things back a little more...

Monoceros · Oct 19, 2016

Sure the power off is what sorted it, did you try a full power cycle of just a soft restart? I found I had to power off for 30sec or so to return to normal.

Think mine was more to do with overclocking CPU in conjunction with to aggressive pixel depth settings, most of my game settings are now set as high as they are able, supersmapling at 0.85 and the debug tool set at 2.0 the GPU is OC at 109%. Also recently enabled ASW though not sure it made a great deal of difference. Have the STRIX GTX1080 O8G card.

Xebec · Oct 19, 2016

Darkh4 said:
I have just returned from work, and to my surprise everything is working again! Elite loads and runs on the Rift, and there's no whining sound. But I didn't change a thing. The only thing that happened is that my rig sat powered off for 12 hours. That almost makes it seem like a thermal shutdown, but my GPU temp rarely goes over 70 degrees C under heavy load in Elite with VR. It's not even supposed to start throttling until around 80.

I don't know, maybe I was pushing it too hard, like Monoceros suggests. But it had been running fine in it's current configuration for days with no indication of a problem. I had the debug tool set to 1.25, Super Sampling at 1.0, most settings on high or ultra, but with some strategic ones set to medium or off. I was getting a steady 90 FPS in space, with ASW kicking in at stations and planets.

I guess I'll have to start dialing things back a little more...

If you want to keep full performance, I recommend getting something like MSI Afterburner and setting up a custom fan profile that's a little more aggressive than the factory settings.

My girlfriend's gaming rig uses a GTX 970 (OC model), and one game that we play (ARK Survival) occasionally crashes after 1-2 hours of play; no other game has this problem. After some monitoring, it would crash even with temps in the mid-70s -- definitely below the 80C (or 95C like AMD GPUs) throttling point. I configured the fan to kick in more aggressively around 65, 70, and 75C and now she doesn't have the crashing problem any more.

Keep in mind there are other parts of the GPU that work very hard that you don't have temperature info for -- the VRMs, the memory chips, etc and the fan profile may help those things run just a bit cooler too. MSI Afterburner works with any GPU.

Darkh4 · Oct 19, 2016

Yeah, I didn't do a full power off and wait, because I didn't think it could be thermal problem. The temperatures hadn't gotten that high according to the monitor. Plus it only happened with Elite. To be honest, this happen just an hour before I had to leave for work, so I only had time to try some other games and then try some things with Elite to make it work. I rebooted several times, but never did a full power off.

My card came with a program called GPU Tweak II that serves as it's setting and monitoring program. I do use it to set a more aggressive fan profile, because the factory setting just isn't adequate. Guess I'll have to get even more aggressive with the profile.

Heavenly-Hammer · Oct 19, 2016

ED didn't kill your 1080, your garbage 1080 killed itself. RMA it.

Darkh4 · Oct 19, 2016

Heavenly-Hammer said:
ED didn't kill your 1080, your garbage 1080 killed itself. RMA it.

Heheh, I wasn't implying that Elite itself broke my card. I was asking if the combination of Elite and VR could have unintentionally caused some sort of driver conflict that was crashing the card. I guess I could have phrased that better. What I get for posting in the middle of 12-hour nightshifts. Games don't break hardware. Overtaxing and excessive overclocking breaks hardware. Anyway, it's working for now. But in the event of its demise, would you care to recommend a replacement?

shadragon · Oct 20, 2016

(Deleted) Wrong forum. Sorry.

Darkh4 · Oct 20, 2016

Turns out it's not fixed, and the issue has spread to other games. A clean graphics driver reinstall seems to clear up the problem until I run Elite on the Rift again. After about 20 minutes of Elite on the Rift, my graphics card shuts down, and refuses to run other games either. My GPU temp was a comfortable 54 degrees C when it shut down, so I don't think it's a thermal issue.

I'm beginning to think Heavenly-Hammer may be right about the card failing, but why would a driver reinstall correct the issue? Is it possible that a 750 watt power supply just isn't enough to drive a 1080 and a Rift at the same time? Or that the registry edit I made to enable ASW is causing some kind of conflict with the new Oculus App version and it's somehow corrupting my graphics drivers?

Edit: It has started shutting down at the desktop now. RMA it is. I've had to reinstall my old GTX 780 for now. No Rift for a while....:|

Beavis Tingleball · Oct 20, 2016

your 780 should run vr well enough to hold you over.. I would also suggest buying another brand of gfx card if you can get a full refund. MSI has an amazing component build on their cards and offers one of the best warranty policies there is. my MSI 1080 armor is silent at 100% fan and holds a super stable 2100+ on a manual OC.

xondk · Oct 20, 2016

Darkh4 said:
Turns out it's not fixed, and the issue has spread to other games. A clean graphics driver reinstall seems to clear up the problem until I run Elite on the Rift again. After about 20 minutes of Elite on the Rift, my graphics card shuts down, and refuses to run other games either. My GPU temp was a comfortable 54 degrees C when it shut down, so I don't think it's a thermal issue.

I'm beginning to think Heavenly-Hammer may be right about the card failing, but why would a driver reinstall correct the issue? Is it possible that a 750 watt power supply just isn't enough to drive a 1080 and a Rift at the same time? Or that the registry edit I made to enable ASW is causing some kind of conflict with the new Oculus App version and it's somehow corrupting my graphics drivers?

Edit: It has started shutting down at the desktop now. RMA it is. I've had to reinstall my old GTX 780 for now. No Rift for a while....:|

The issue is not Elite, it is likely your overclocking that has burned out the card, if so, your RMA will get denied :/

[HJ]-RedRaven · Oct 21, 2016

xondk said:
The issue is not Elite, it is likely your overclocking that has burned out the card, if so, your RMA will get denied :/

Doubtful; overclocking can stress cards, but if there's a failing capacitor or other bit of hardware, it'd fail anyway even if normally clocked.

nVidia knows you're going to stress the cards (just by playing games on them), and overclocking is not only allowed, but actively encouraged. The chips and cards are designed to handle overclocking, but bad components can slip in past the gatekeeper.

Hope the RMA goes ok.

At least you have the 780GTX to tide you over again - it will run VR, and ASW will help it a fair bit too. I ran my 780GTX for a few weeks before my 1080 arrived. It will do the job for now!

X-Terminator · Oct 21, 2016

Just wondering if you're running any Fan Control program and a Max Boost switch to engage the factory SuperClock settings?
My EVGA 970 uses X-Precision and it needs a Custom Fan Curve especially well set just for ED which is higher than other games for fan speed else it will in fact crash. It has a KBoost button on it that engages SuperClock. If not on then it doesn't run at its best obviously. If my Custom Fan Curve is not on then the fans are not even on unless it got over 80degrees but it crashes before this so it must be enabled.

Darkh4 · Oct 21, 2016

The saga continues, turns out it's not the card. I put my 780 in the rig and the same thing happened when I tried to launch The Witcher 3. Even in a different PCI slot. I'm now operating on the theory that the power supply is the issue. It is a Corsair RM750x that is less than two months old, but I'm thinking it may not have been up to the task of driving a GTX 1080 under heavy load while also powering a Rift, a Rift sensor, a HOTAS, a second storage HDD, and a wireless usb adapter.

A 600 watt PSU is supposed to be enough for a 1080 in a typical rig, so I thought a 750 watt supply should be fine, but I'm not sure this counts as a typical rig with the Rift. The RM750x as also one of Corsair's new-style PSUs that relies on passive cooling until the temperature reaches a certain level, and it's possible the fan isn't coming on soon enough, which allowed it to cook under heavy load.

A failing power supply would tend to explain why allowing the system to cool down causes everything to work normally again, and why I can play other, non-VR games indefinitely as long as I don't stress the system with a heavy load first. It might also explain the whining/buzzing sound.

The overclock on the 1080 is from the factory, at their settings, so I doubt they would have denied the RMA for it. The only thing I changed was the fan profile to be more aggressive, when I noticed that the card was reaching 80 degrees. With the new profile it wasn't even reaching 60 degrees.

Unfortunately, the GTX 780 only has one HDMI output, so it's not really ideal for VR. I appreciate the thought, though.

I've put in an order for a 1000 watt PSU with monitoring software that allows a fan profile setup, so I'll see how that goes. It's overkill, but I'm not taking any chances with frying my brand new rig.

In the meantime it is working fine on the 1080, as long I don't stress things too badly. I've moved the GPU power cable to the other PCI-E output on the PSU in the hope that it's undamaged, but I don't intend to put it to the test. So no VR for now.