GTX 980 ti SLI running two 4k Monitors crashing

[DocJuan24]

Specs: eVGA GTX 980 ti SLI, 2x Samsung 28e590D monitors, eVGA SuperNOVA 1200 P2 Power Supply, AMD FX 9370 water cooled, GTX 980 ti's running SLI.

One monitor I could run a Twitch stream and play games like GTA V maxed etc no issues with the Twitch stream minimized in the background.

However I purchased my second 4k monitor this week and hooked it up tonight and tried running GTA V on one of my monitors and shifted the Twitch stream over to the second monitor. (I'm mirroring the desktop so not extending into some crazy resolution.) both monitors are running 3840x2160, anyway everything is working fine for the first 5 minutes of game play, after that the 2nd monitor starts flickering black and moments later my PC freezes and goes to a blue screen.

My suspicion (IS that 980 ti's can't handle dual 4k.) I just want to confirm them, or is one of them faulty etc, could it be that I have one hooked up via DP, and one HDMI?

Is there a way to get this configuration to work? Is this actually a CPU issue? These are things I want to rule out before deciding on purchasing Titans. Can the Titans even run 2 4k monitors?

Thanks!

Specs if needed.


CPU: AMD FX-9370 Black edition 4.4 GHz
CPU Cooler: Corsair H110i GTX Liquid CPU Cooler
Motherboard: ASUS ROG Crosshair V Formula-Z
Memory: 1 x G.SKILL Ripjaws X Series 16GB (2 x 8GB) 240-Pin DDR3
Storage: SAMSUNG MZ 500GB Solid State Drive
Video Card: Dual Nvidia GeForce GTX 980 Ti Video Card
Case: Corsair 760T White ATX Full Tower Case
Power Supply: eVGA SuperNOVA 1200 P2
Operating System: Microsoft Windows 10 (64-bit)
Keyboard: Razer Ananasi
Mouse: Razer Deathadder CHROMA
Monitor: 2x Samsung 28" 4k monitors @ 60hz
-edited to update correct power supply-

 

[Darkbreeze]

980TI SLI "CAN'T" handle two 4k monitors. It can barely handle one and maintain 60fps. Sometimes, depending on the demand of the title, not even that. But I doubt that alone is the cause of your crashing.

More likely is the fact that you're running two 980 TI's with the extremely power hungry FX-9370 which AMD recommends using an 850w PSU with, minimum, and recommends a 1000w PSU for in most cases. Granted, you're running a high quality PSU, but with two 980TI's, and a 9370, you're very close if not over the capability of your PSU.

It's also entirely possible that the addition of a second 4k load is creating a thermal condition. I'd monitor the thermal values of the graphics cards, CPU and the motherboard in conditions similar to when it's failing. Looks like you have that 9370 underclocked. Is there a reason for that? Were you having problems with CPU thermals or VRM throttling previously? Considering you're running a liquid cooler, it's highly likely to have a VRM throttling condition or instability, which MIGHT be responsible, or something similarly related, since without an air cooler on the CPU you are probably getting very little cooling across the motherboard VRMs that would be present with an air cooler.

It's also possible you simply have a bad card. And I just noticed that you have two different power supplies listed. EVGA P 1200 and Corsair 860w, so which one do you actually have?

I'd install HWinfo, not any other utility, run "sensors only" and take a peek at what all the thermals as well as the system voltages are doing when running in a similar scenario as what you've described above. If possible, take screenshots and post them here. Probably it will take three screenshots to capture all sensors.

Also, if you have turbo core enabled in the bios, I'd turn it off.

 

[DocJuan24]

980TI SLI "CAN'T" handle two 4k monitors. It can barely handle one and maintain 60fps. Sometimes, depending on the demand of the title, not even that. But I doubt that alone is the cause of your crashing.

More likely is the fact that you're running two 980 TI's with the extremely power hungry FX-9370 which AMD recommends using an 850w PSU with, minimum, and recommends a 1000w PSU for in most cases. Granted, you're running a high quality PSU, but with two 980TI's, and a 9370, you're very close if not over the capability of your PSU.

It's also entirely possible that the addition of a second 4k load is creating a thermal condition. I'd monitor the thermal values of the graphics cards, CPU and the motherboard in conditions similar to when it's failing. Looks like you have that 9370 underclocked. Is there a reason for that? Were you having problems with CPU thermals or VRM throttling previously? Considering you're running a liquid cooler, it's highly likely to have a VRM throttling condition or instability, which MIGHT be responsible, or something similarly related, since without an air cooler on the CPU you are probably getting very little cooling across the motherboard VRMs that would be present with an air cooler.

It's also possible you simply have a bad card. And I just noticed that you have two different power supplies listed. EVGA P 1200 and Corsair 860w, so which one do you actually have?

I'd install HWinfo, not any other utility, run "sensors only" and take a peek at what all the thermals as well as the system voltages are doing when running in a similar scenario as what you've described above. If possible, take screenshots and post them here. Probably it will take three screenshots to capture all sensors.

Also, if you have turbo core enabled in the bios, I'd turn it off.

It's the Evga p2 1200 watt psu sorry copied someone else's spec list and didn't delete that one for some reason! So after the 2nd crash I was watching the temps on both gtx cards and they were running 44-48 C. The CPU I didn't take a look at. What I don't get is I previously had the one 4K monitor and a 1440p monitor hooked up and the system had no issues with me running a game on the 4K screen and running a Twitch stream on the 2nd 1440p monitor.

As for airflow I have this baby in a basement that rarely gets above 60f. I've got 4 120mm fans blowing directly on the mobo or pulling air in, 1 240mm fan pushing air out the front of the case, 2 120mm fans pushing air through the top/radiator and one 140mm fan pushing air out.

As for VRM the CPU as far as I can tell has never given me issues. I've never had it go above 40c even when running intensive programs. Of course this was using something like speed fan to monitor it. I might have the speed on the processor wrong, I've not OCed or touched the voltages on anything on the system.

I'll get HWInfo installed and drop some screenshots tonight.

I just want to note, I have an i7-6700k standing by, I'm wondering if replacing the processor might help?

I just want to point out I'm not looking to run a game in 4K on one monitor and a second game in 4K on the 2nd monitor. I just want to run a game on the 1st monitor in 4k and run my chat programs or have word open or YouTube/Twitch open on the 2nd screen. Is it possible that Twitch is so CPU intensive that it's overloading my CPU? Or is it strictly a GPU issue? Guess HWInfo will be able to give us an idea of what's going on?
-edit to answer questions-

 

[RobCrezz]

980TI SLI "CAN'T" handle two 4k monitors. It can barely handle one and maintain 60fps. Sometimes, depending on the demand of the title, not even that. But I doubt that alone is the cause of your crashing.

??

Why cant they run two 4k monitors? Hes only gaming on one...

 

[Darkbreeze]

I didn't catch that part. I was under the impression he was gaming on both screens. Even without gaming on both screens though, 4k x2 is still a lot of pixels to render considering 980 TI struggles to maintain 60fps on one 4k display if it's any kind of demanding game.

 

[RobCrezz]

I didn't catch that part. I was under the impression he was gaming on both screens. Even without gaming on both screens though, 4k x2 is still a lot of pixels to render considering 980 TI struggles to maintain 60fps on one 4k display if it's any kind of demanding game.

It's not rendering 2x 4k though..

 

[Darkbreeze]

Yes, I really think the 6700k is about six hundred times better option than any FX processor, especially the 9370 or 9590, which are not so affectionately looked upon to be something of a joke in most tech circles. Just really terrible marketing scam by AMD to factory overclock an FX chip to obscene levels with no consideration of power or heat. All but a handful of boards with experience VRM throttling or instability with these chips due to the amount of voltage needed to run them at stock speeds unless you downclock and undervolt them, which is what a lot of people end up doing.

Do you also have a motherboard for that 6700k? The 6700k will easily run circles around the 9370 in pretty much everything aside from a handful of video encoding and content generation applications.

 

[Darkbreeze]

I didn't catch that part. I was under the impression he was gaming on both screens. Even without gaming on both screens though, 4k x2 is still a lot of pixels to render considering 980 TI struggles to maintain 60fps on one 4k display if it's any kind of demanding game.

It's not rendering 2x 4k though..

So one display is turned off then and not showing anything? Regardless, it's improbable that the issue is related to this anyhow.

 

[DocJuan24]

I didn't catch that part. I was under the impression he was gaming on both screens. Even without gaming on both screens though, 4k x2 is still a lot of pixels to render considering 980 TI struggles to maintain 60fps on one 4k display if it's any kind of demanding game.

It's not rendering 2x 4k though..

So one display is turned off then and not showing anything? Regardless, it's improbable that the issue is related to this anyhow.

Here's what happens. For the purpose of this post picture yourself seated at your desk, to your left is 4K monitor 1 to your right is 4K monitor 2. I'll open my browser in monitor 1(all of the stuff I open opens here.) I'll DRAG the browse from monitor 1 to monitor 2 and load a Twitch stream.

Meanwhile monitor 1 has nothing running on it besides the desktop. So I'll load GTA V on monitor 1. The Twitch stream and browser are still running on monitor 2.

GTA V loads on monitor 1. I'll begin playing it suddenly the screen of monitor 2 will go black. It will flicker and come back on. The sound from the video and the video itself will go out when the screen on monitor 2 goes out and when it comes back and shows an image the video and sound from the Twitch stream will resume.

Meanwhile on monitor 1 GTA V is running perfectly. No stutter. No fps drop nothing. Flawless gameplay. Eventually. After monitor 2 shuts off and comes back on again, BOTH monitor 1 and monitor 2 go black, and eventually just a blue screen appears on Monitor 1.

This also may or may not be relied, but if I alt tab outside of GTA V? At ALL? Like to change. The music on YouTube on monitor 2? GTA V on the first monitor will freeze and lock me out from using the main monitor. I have to restart the computer. However I can still do stuff on monitor 2. As for the 6700K? I have an Asus Maximus viii hero alpha, but man was I really hoping to wait until Pascal dropped before I jumped into that build.

 

[DocJuan24]

I didn't catch that part. I was under the impression he was gaming on both screens. Even without gaming on both screens though, 4k x2 is still a lot of pixels to render considering 980 TI struggles to maintain 60fps on one 4k display if it's any kind of demanding game.

It's not rendering 2x 4k though..

So one display is turned off then and not showing anything? Regardless, it's improbable that the issue is related to this anyhow.

Hmm....Interesting thing. So I went to take a screen shot of HWInfo for you and it's registering my display as 7680x2160. It's taking a picture of BOTH screens.

So IF this is indeed the issue and I'm somehow effectively DOUBLING the resolution. (Forcing my GPU to run 7-8k vs 4k) Is there a way to set it up so that my display isn't "extended?" and use the 2nd monitor for other things?

 

[Darkbreeze]

Screenshots always take snapshots of all active screens when they're in extended mode. Extended IS what you want, if you don't want whatever is on the primary screen to simply be duplicated onto the second screen. There is apparently now configuration settings for a two display eyefinity, but you would have to intentionally configure that so it wouldn't be the case here.

Anytime you have two active displays though, and both are connected to the same GPU card, you ARE going to be displaying double the pixels. You'd have to be, otherwise there wouldn't be anything on one of the screens. Also, can you please edit the system specs you posted in your first post to reflect your ACTUAL system, rather than the wrong one you copy/pasted. It's confusing.

Just take the screenshot, and then edit the second screen out with whatever graphics or photo editing program you have so that only the HWinfo window is left and post that. I always have to do that whenever I take screenshots since I have three displays going all the time.

 

[RobCrezz]

I didn't catch that part. I was under the impression he was gaming on both screens. Even without gaming on both screens though, 4k x2 is still a lot of pixels to render considering 980 TI struggles to maintain 60fps on one 4k display if it's any kind of demanding game.

It's not rendering 2x 4k though..

So one display is turned off then and not showing anything? Regardless, it's improbable that the issue is related to this anyhow.

Well, 2d rendering, which takes very little power.

Im just making a point, that running 4k in windows, doesn't require a powerful GPU.

 

[Darkbreeze]

I agree, in most cases. There are processes, like gaming of course, but also a few others, that tax the GPU quite heavily. But if he's gaming on one display and only the desktop or some browser tabs in the other, that aren't also displaying some kind of taxing process, it should be ok. If he's gaming on one display and doing something else taxing on the other, it might be contributing.

I still think it's possible that it's related to the very high voltage demands of 980TI SLI, WITH an FX-9370, although that P2 should be able to handle it fairly easily. Maybe instability due to VRM throttling since he's using water cooling rather than air and is getting no direct residual airflow on the motherboard in that area.

Just for grins, I'd try taking off the side panel, pointing some kind of fan in the area of the VRMs/CPU, and see if the problem still exists. Might even be that you simply have one GPU card with an issue. And as I said before, if you're using a power strip, ditch it. Plug directly into the wall. I've seen a whole lot of mystery issues related to problems with cheap ass power strips.

 

[azzazel_99]

I'm willing to bet he is at his psu's limit. Dump the garbage amd cpu and put in the skylake. You can also try to game in 4k and on each card individually to make sure the new card is bad but i highly doubt it is. Much more likely with 2 980ti's and that poor excuse for a cpu you're using is hitting the power wall.

 

[Darkbreeze]

With an EVGA P2 1200w, not very likely, unless there is an issue with the supply itself. 980TI and 220w CPU will easily run on a good 1000w unit. 1200w P2 should be fine, again, unless there is an issue with it or something else. Pulling anything near peak amps through a cheap power strip or incompatible UPS WOULD cause issues though.

 

[DocJuan24]

With an EVGA P2 1200w, not very likely, unless there is an issue with the supply itself. 980TI and 220w CPU will easily run on a good 1000w unit. 1200w P2 should be fine, again, unless there is an issue with it or something else. Pulling anything near peak amps through a cheap power strip or incompatible UPS WOULD cause issues though.

Here we go. Got some screen shots for you.

First with nothing running besides Twitch.
1st half

2nd half

HWInfo taken with Twitch and GTA V running.
1st half

2nd half

HWInfo Taken with Twitch and GTA V running > 10 mins of gameplay at this point 2nd monitor is flickering in and out. Otherwise game is running smoothly.
1st half

2nd half

HWInfo taken with Twitch and GTA V running > 25 mins of total gameplay 2nd monitor is flickering in and out every 15 seconds, but game is running smoothly. So far no hard crashes or BSOD.
1st half

2nd half

It's worth noting that I had to reset the PC a couple of times while taking these. I did some research and verified that it isn't uncommon for the game to "freeze." when alt tabbing out of it. In that case I just made sure I stopped the stop watch on my phone and reset it when I restarted the game. So the 1st set of screenshots is 0 time spent in GTA V.
The 2nd set is 5 minutes. The 3rd set is 15 mins of timed gameplay and the 4th set was the longest I ran the game at a total of 30 minutes.

 

[Darkbreeze]

If it were ME, I'd RMA the motherboard. There is no possible way those thermal sensors are reading correctly. Try doing the same thing with twitch and GTA V, but use AMD Overdrive and see what the thermal margin says. It's impossible for any 8 core FX chip at 1.35v or higher, with 90%+ load on all 8 cores to only be at 25-30°C. There is a clear problem with thermal readings. It MIGHT just be HWinfo, so try Overdrive which is usually more accurate on AMD chipsets, and if we see a similar overall picture, then I'd say something is wrong with the board and your problem might be thermal and the VRMs or CPU are actually overheating but the system isn't reacting the way it should because the thermal sensors don't indicate a problem and are not causing the cooling to ramp up.

I'm uncertain, as I don't see anything else going on there aside from that, but clearly SOMETHING is wrong with that picture.

 

[azzazel_99]

his twitch only temps on the cpu are 10c also. I'd say try real temp, core temp or the overdrive. Thats def not right.

 

[DocJuan24]

If it were ME, I'd RMA the motherboard. There is no possible way those thermal sensors are reading correctly. Try doing the same thing with twitch and GTA V, but use AMD Overdrive and see what the thermal margin says. It's impossible for any 8 core FX chip at 1.35v or higher, with 90%+ load on all 8 cores to only be at 25-30°C. There is a clear problem with thermal readings. It MIGHT just be HWinfo, so try Overdrive which is usually more accurate on AMD chipsets, and if we see a similar overall picture, then I'd say something is wrong with the board and your problem might be thermal and the VRMs or CPU are actually overheating but the system isn't reacting the way it should because the thermal sensors don't indicate a problem and are not causing the cooling to ramp up.

I'm uncertain, as I don't see anything else going on there aside from that, but clearly SOMETHING is wrong with that picture.

Never mind! Let me get some screen shots.

about 5-7 mins of gameplay in GTA V. I'm run it for a little bit longer if needed. Unless you all are positive it's mobo or even CPU (shit CPU bogging down great GFX cards.)

 

[Darkbreeze]

I'd move to the 6700k, now, and maybe try to figure out the issue with the AMD configuration as a separate system.

That being said, that still looks off. I believe that a 47°C thermal margin would be about the same as what HWinfo was reporting, and I KNOW that cannot be correct. Even an FX-8320 or 8350, which are the same chip, at 1.35v with those kind of loads would be up into at least the high 50's or somewhere in the 60's.

Is that chip at stock settings or has it been underclocked via the bios or a windows utility, because I'm only seeing a max speed of 4.1Ghz in any of your screenshots and stock speed for the FX-9370 should be 4.4Ghz with a max turbo speed of 4.7Ghz, and you're getting nowhere near that, but even at 4.1Ghz at 1.35v, temps should be higher than we're seeing here even with great cooling.

So either it's being throttled to that speed, has been reconfigured/underclocked, or something else is going on.

 

[azzazel_99]

Temps look more normal in the upper 40's and looks like the cpu is maxed more or less. Swap out the cpu with the skylake and new motherboard then sell you current cpu and motherboard and soak up some of the cost. Then you'll be on your marry way.

 

[Darkbreeze]

That's not a 40 degree temp though. It's a 47 degree thermal margin. Which means it's showing about a 35 degree temp, and that still can't be right.

 

[DocJuan24]

These aren't normal temps even with the water cooler on it? Also keep in mind my basement floor is COLD. Not your standard 70f room temp. It's 55-60 degrees F on a good day and I've got those 4 120mm fans on the case shooting this cool air right on top of the mobo.

 

[azzazel_99]

If you're liquid cooled, downclocked, and the room temp is that cood i dont see why that cant be right.

 

[Darkbreeze]

Yes. You're both right, if that's the case. If there is a VERY low ambient temp, the CPU is downclocked and you're running a 240 or 280mm cooler, then temps can be much lower than normal, but they still wouldn't be 20-30°C with a high load. Those are more like idle temps, not temps under load and I don't care if you're running a custom loop they still wouldn't be that low under a major load.

Let's clarify, so there can be no mistake.

Please download Prime95 version 26.6, and ONLY version 26.6, and run small FFT for five to ten minutes then take a screenshot of both HWinfo (First half. Include CPU 0 and CPU package temp) AND Overdrive thermal margin, WITH Prime still running. It's possible that twitch and GTA-V are not fully loading the CPU. I just want to positively determine there are no thermal issues before moving on to other pastures.

http://windows-downloads-center.blogspot.com/2011/04/prime95-266.html