Warframe is the only game that crashes/restarts PC

[.rot.]

It's been countless months of erratic, unexplained crashing and now I have chose to return last month to check on the Railjack content; nothing has changed. Because of this game, I have had nights upon nights of tired, baggy-eyed researching in attempt to find solutions to this. I quite literally have even resorted to a fresh installation of Windows 10. I have disabled every known background application and ran the game that way, I have unplugged every USB device except for mouse and keyboard, I have unplugged all my internal drives save for my C drive which has warframe on it, I underclocked/undervolted my card (gtx 1080), I've done hardware stress tests- all the results coming back in my favor, I bought a new PSU, the list just goes on and on and on... My girlfriend has even told me to just play other games, there's just something wrong with warframe.

My specs:

windows 10
gtx 1080
4770k
16gb ram

[MillbrookWest]

A generic "my client crashes" doesn't really help DE find any solution. Because similarly, my client hasn't crashed in months.

When a crash happens, after booting back in, you can grab your EE.log and pass that through in a support ticket. DE might be able to see whats up if indeed there is an issue with the game.

Alternatively, you can use windows' "event viewer" and check to make sure there were no windows logs generated. Typically, the games you run are no different to the browser you have open now. They run in some space within windows. Which means if they crash, windows is still running. The game might cannibalize your input (so you can't get back to windows) but windows is still running and would log that the program had an error. I say this because an app crashing your entire system means its gone beyond windows, meaning a hardware fault is likely suspect.

If the 1080 is factory overclocked, try removing the overclock. Same for CPU. Run memtest to make sure the RAM sticks are fine. If they come back fine, it may the motherboard. Faulty VRM's etc.

[.rot.]

2 hours ago, MillbrookWest said:

A generic "my client crashes" doesn't really help DE find any solution. Because similarly, my client hasn't crashed in months.

When a crash happens, after booting back in, you can grab your EE.log and pass that through in a support ticket. DE might be able to see whats up if indeed there is an issue with the game.

Alternatively, you can use windows' "event viewer" and check to make sure there were no windows logs generated. Typically, the games you run are no different to the browser you have open now. They run in some space within windows. Which means if they crash, windows is still running. The game might cannibalize your input (so you can't get back to windows) but windows is still running and would log that the program had an error. I say this because an app crashing your entire system means its gone beyond windows, meaning a hardware fault is likely suspect.

If the 1080 is factory overclocked, try removing the overclock. Same for CPU. Run memtest to make sure the RAM sticks are fine. If they come back fine, it may the motherboard. Faulty VRM's etc.

I've already sent a support ticket spanning over the course of 5 days, in which I attempt to replicate the issue and then provide the logs each time a crash occurs. I have also used the event viewer to the fullest extent I can, only finding a generic kernel 41 errors occurring. I did indeed remove the overclock for my GPU as well as run stock for my CPU. Did a memtest as well for my RAM and they are fine. I've literally exhausted most of these recommendations as well as other methods from countless forum topics dating back to like 2015, making my eyes bleed reading the same thing over and over again.

[MillbrookWest]

4 minutes ago, .rot. said:

I've already sent a support ticket spanning over the course of 5 days, in which I attempt to replicate the issue and then provide the logs each time a crash occurs. I have also used the event viewer to the fullest extent I can, only finding a generic kernel 41 errors occurring. I did indeed remove the overclock for my GPU as well as run stock for my CPU. Did a memtest as well for my RAM and they are fine. I've literally exhausted most of these recommendations as well as other methods from countless forum topics dating back to like 2015, making my eyes bleed reading the same thing over and over again.

A quick google shows that Microsoft note this as a potential hardware failure as well (one of three scenarios, the other two being normal); The others being an Stop Error (which would give you a blue screen), or physically shutting down by holding the power button down. 

Quote

Typically, the symptoms described in this scenario indicate a hardware problem. To help isolate the problem, do the following:

• Disable overclocking. If the computer has overclocking enabled, disable it. Verify that the issue occurs when the system runs at the correct speed.
• Check the memory. Use a memory checker to determine the memory health and configuration. Verify that all memory chips run at the same speed and that every chip is configured correctly in the system.
• Check the power supply. Verify that the power supply has enough wattage to appropriately handle the installed devices. If you added memory, installed a newer processor, installed additional drives, or added external devices, such devices can require more energy than the current power supply can provide consistently. If the computer logged Event ID 41 because the power to the computer was interrupted, consider obtaining an uninterruptible power supply (UPS) such as a battery backup power supply.
• Check for overheating. Examine the internal temperature of the hardware and check for any overheating components.

If you perform these checks and still cannot isolate the problem, set the system to its default configuration and verify whether the issue still occurs.

What motherboard, and how old is the bios?

[.rot.]

Gigabyte G1.Sniper Z87 and I believe I have whatever the latest version is for the bios.

I will take note on the power part you listed, I have been considering purchasing a UPS, but that would be around $200. Also worth mentioning, I was sort of rash with the decision, but since PSU is generally the first thing we point to for event ID 41, I went and got a 1000w Corsair PSU. Problem still obviously persists.

As for temps, I monitor them like a hawk and everything stays within a healthy range.

[MillbrookWest]

7 minutes ago, .rot. said:

Gigabyte G1.Sniper Z87 and I believe I have whatever the latest version is for the bios.

I will take note on the power part you listed, I have been considering purchasing a UPS, but that would be around $200. Also worth mentioning, I was sort of rash with the decision, but since PSU is generally the first thing we point to for event ID 41, I went and got a 1000w Corsair PSU. Problem still obviously persists.

As for temps, I monitor them like a hawk and everything stays within a healthy range.

I pulled the quote from Microsoft's page, the efficacy of a UPS is entirely unknown to me in this case, but since the PSU is new, you might end up wasting money (that would be my 2c).

The MS page also states that you can try CAPS LOCK when the system hangs to verify if it is potentially a hardware lockup:

Quote

This scenario includes the following circumstances:

• To verify that a computer is unresponsive, press the CAPS LOCK key on the keyboard. If the CAPS LOCK light on the keyboard does not change when you press the CAPS LOCK key, the computer might be completely unresponsive (also known as a hard hang).

as per: "Scenario 3: The computer is unresponsive or randomly restarts, and Event ID 41 is not recorded or the Event ID 41 entry or lists error code values of zero"

In which case, they give the points above.

[.rot.]

Typically my PC will just restart when this problem occurs but there are times where my screen freezes (I suppose "hard hang") and I have to resort to a hard shut down, but I'll check and do some testing with that in mind. I notice some tile-sets can cause this issue to happen much more frequently, such as snow tile-sets and I know I will crash without a doubt during a Void Fissure in a snow tile-set, but Abaddon, Europa has been a spot where I can replicate this like 90% of the time.

Strange how when going through some motions in the support ticket, I was able to play like a 5 hour session including those snow tile-sets, but within the past 48 hours it has returned to its erratic, randomly crashing state. It's very tough to gauge this.

[MillbrookWest]

10 minutes ago, .rot. said:

Typically my PC will just restart when this problem occurs but there are times where my screen freezes (I suppose "hard hang") and I have to resort to a hard shut down, but I'll check and do some testing with that in mind. I notice some tile-sets can cause this issue to happen much more frequently, such as snow tile-sets and I know I will crash without a doubt during a Void Fissure in a snow tile-set, but Abaddon, Europa has been a spot where I can replicate this like 90% of the time.

Strange how when going through some motions in the support ticket, I was able to play like a 5 hour session including those snow tile-sets, but within the past 48 hours it has returned to its erratic, randomly crashing state. It's very tough to gauge this.

I'm not familiar with a scenario where Windows will restart in this manner. Typically, windows only unknown 'exit' is with a blue screen.

I had a quick google look at the board, two issues i saw were with 4xRAM sticks vs 2xRAM sticks (4 can lead to issues), and a BIOS setting for Gen2 PCI-E. Not sure if they'll help any tho.

EDIT: The random restart (not shut down) would also further suggest hardware isn't playing nice somewhere.

Edited January 14, 2020 by MillbrookWest

[.rot.]

Just attempted to play and my PC restarted within taking 3 steps in my orbiter. For the RAM, I have 2x8. 

[MillbrookWest]

Just now, .rot. said:

Just attempted to play and my PC restarted within taking 3 steps in my orbiter. For the RAM, I have 2x8. 

Do you have vsync enabled?

[.rot.]

Yes I do and I guess to shed a bit more light, I have a dual monitor setup my secondary monitor is a 32" 1080p 144hz, my main being a 32" 1440p 144hz.

[MillbrookWest]

Just now, .rot. said:

Yes I do and I guess to shed a bit more light, I have a dual monitor setup my secondary monitor is a 32" 1080p 144hz, my main being a 32" 1440p 144hz.

It could be a GPU issue of some sort. If you have a second card, you can try testing with that.

Alternatively, that PCI-E gen2 issue mentioned above would relate to the GPU. You could see how your mileage goes with that option.

[.rot.]

I'll give that a shot.

I just find it really strange that I am able to play hours and hours of Red Dead Redemption 2 and Warframe is the game that causes me these issues.

[.rot.]

Just to confirm, I notice the the PCI-E configuration was set to Auto, should I be changing it to Gen 2? I also noticed Gen 3 was an option. I did a quick google and I'm still not sure which would make a difference and what Auto was causing my card to be read as.

[MillbrookWest]

8 minutes ago, .rot. said:

I'll give that a shot.

I just find it really strange that I am able to play hours and hours of Red Dead Redemption 2 and Warframe is the game that causes me these issues.

It depends on how devs want to use the GPU tbh. Modern games tend to be very heavy on shaders and subsequently implement Deferred rendering (and long render pipelines), So most of the work gets done in the shader units. Warframe is still a forward rendered game as far as i'm aware, and quite lite. I can't recall off the top of my head which part is responsible for delivering frames (ROPs?) but Warframe at high frame rates would be pushing a lot of work through the entire GPU.

On my PC, warframe at 1440p @120Hz produces an audible coil whine from my 970. Other games might hit the core harder and never produce the same sound, but whenever the 970 is allowed to hit insane high framerates, i always get coil whine (regardless of game).

Just now, .rot. said:

Just to confirm, I notice the the PCI-E configuration was set to Auto, should I be changing it to Gen 2? I also noticed Gen 3 was an option. I did a quick google and I'm still not sure which would make a difference and what Auto was causing my card to be read as.

I would assume Auto would default to the hardware spec, so gen3. The post i found recommended dropping it to gen2. Again, not too sure what your mileage will be like with it tho.

[.rot.]

I tried the gen2 setting and the problem is still occurring. I'm truly beyond puzzled. But as expected, I crashed within 3 minutes of a Neo Void Fissure Excavation in the snow tile-set (of course). Prior to that, I did a quick Capture in Abaddon, Europa (snow tile-set) and flew around Orb Vallis for a few moments. I don't even know where to start if there are any settings I can edit within nvidia control panel, the launcher settings, the game, etc. I obviously don't want to resort to running it in a manner where it looks bad, considering I've been playing stuff like RDR2, Witcher 3, Control and the like.

[MillbrookWest]

12 minutes ago, .rot. said:

I tried the gen2 setting and the problem is still occurring. I'm truly beyond puzzled. But as expected, I crashed within 3 minutes of a Neo Void Fissure Excavation in the snow tile-set (of course). Prior to that, I did a quick Capture in Abaddon, Europa (snow tile-set) and flew around Orb Vallis for a few moments. I don't even know where to start if there are any settings I can edit within nvidia control panel, the launcher settings, the game, etc. I obviously don't want to resort to running it in a manner where it looks bad, considering I've been playing stuff like RDR2, Witcher 3, Control and the like.

You could try running the card with a frame rate limit - Nvidia's new drivers have a frame limiter in the control panel. It's not ideal since it's a high refresh monitor, but it might help isolate the problem since high frame rates pull more power through the card, potentially tripping something. 72 would still keep similar frame pacing for a 144Hz monitor.

Outside that, going back to Microsoft's recommendation, 2 monitors aren't considered "default" so you could try with just a single monitor.The issue could be with nvidia's drivers since the two screens aren't the same.

Edited January 14, 2020 by MillbrookWest

[.rot.]

I've actually tried that (single monitor) in order to rule that out and simply use my 1080p monitor to see how it would run, the crash still occurred. I also limited the frames, even down to 60fps to no avail.

[MillbrookWest]

8 minutes ago, .rot. said:

I've actually tried that (single monitor) in order to rule that out and simply use my 1080p monitor to see how it would run, the crash still occurred. I also limited the frames, even down to 60fps to no avail.

As i mentioned above, the fact that it restarts (and not just shut down) tends to be somewhat indicative of hardware/driver issues. Not that they usually cause trouble, but are the drivers for the board also up to date? (might as well ask)

It sounds like it may be something with the GPU. Have you tried giving Kombustor a run? It's got a bunch of stress tests that test different parts of the gpu, you may find something that trips up.

If you have a second system, you could try swapping the GPU's and seeing what the system behavior of both systems is. Or if not, try swapping out the card for a different one.

 

[.rot.]

Just did a 37min run in the Neo Void Fissure excavation map in a snow tile-set after turning these settings off: Runtime Tessellation, Local Reflections, Volumetric Lighting, High Shader Quality, Dynamic Lighting. I'm wondering which could be the culprit here, but for now I'm going to keep playing and see how I fair. Any recommendations on this regard? Although maybe I was just lucky in this particular run and didn't crash which does occur with this strange issue, making it hard to narrow things down to.

The GPU drivers are up to date. I've never even heard of Kombustor, but I'll take note of that.

I sadly do not have a second system or any old GPU's available to me, but I am planning on upgrading to a new card if the new nvidia series releases this year. I'm also waiting for the new Intel CPU standard to come out so I can take the leap on a MoBo and some DDR4 ram as well. So all in all, I'd hope this issue be non-existent by the end of this year, but for now I must bide my time. 

[.rot.]

Nevermind, seemed to have crashed 5mins in to fighting a Teralyst

[sam686]

23 hours ago, .rot. said:

Typically my PC will just restart when this problem occurs but there are times where my screen freezes (I suppose "hard hang") and I have to resort to a hard shut down

There might be something wrong with graphics card, possibly the GPU but not entirely sure. Try using a different graphics card or integrated graphics. Does a different graphics card or integrated graphics stops all your crashes?

Which stress tests you used? For CPU stress test, I like Prime95 as it stops on errors. I don't know of a good GPU stress test, as most GPU graphics stress test like Furmark only use a tiny 0.1 GB of graphics memory and no error checks. I have seen Warframe use up 5 GB of my graphics memory in Plains of Eidolon and Orb Vallis. A faulty spot on graphics memory might crash computer on games that use large amount of video memory.

 

I have Asus B450-F, Ryzen 2400g with Vega 11, iGPU as primary, HybridGraphics enabled. The rendering is on the RX 580, output to monitor plugged into integrated graphics. I have seen my computer hard hangs with frozen screen while using AMD RX 580 and was overheating with default clocks. Pushing the reset button to reboot, makes my AMD RX 580 disappear from device manager and game runs at a lower FPS on slower GPU. Have to power cycle to recover Rx 580. I lowered the AMD RX 580 clocks to avoid overheating problems, but have to remember to lower clocks again on each graphics driver updates. My computer case have a reset button that can plug into motherboard header connections.

My user benchmark results: https://www.userbenchmark.com/UserRun/21360374

[taiiat]

so, you have tried running a CPU stress Test, a GPU stress Test, and then running both at once?
also, which stress Tests did you choose for each?

 

what other games do you play, so that we can determine if those are good comparisons with Warframe?

[99Moon_Dancer]

my guess is the Board. I work a lot with Hardware and there are issues with Gigabyte boards after a timeframe of 4-6 years

u can try Setting the powerplan to High / Ultimate Performance (via Settings -> System -> Power and Sleep -> Additional power settings) that fixes the issues in most cases

the GA-Z77X-UD3H for example needs for example extra voltage on the cpu (you can try that to 50 to 100mv offset ) after running 4-6 years fine some of them wont even post without it

 

Edited January 17, 2020 by Injected_Pie

[(PSN)silvia95guy]

I am with you OP!!! I've been trying to figure this issue out as well for months on a newly built system! I have a few reddit threads I've started on the topic, as well as have found hosts of many more people with the exact same issue, and only with Warframe. One of the things I have in seen in common among them all is that they've all seemed to be a GTX 10 series card. I'm running a 1070Ti, that I bought from my buddy who had none of the issues described above. I built my system almost exactly the same as his. Both running the same Mobo (MSI B450 Tomahawk Max), GPU (1070Ti, him 2070 Super now), PSU (corsair cx750M), RAM (corsair vengeance 2933 2x 8gb) , and CPU (Ryzen 7 2700, he had 2700X, 3700X now). He ran Warframe and says he never had an issue. I can run it maybe 30 minutes if I'm lucky. Usually just 5-10 min. Only in mission or in POE. I can spend all the time I want in my orbiter. My CPU and GPU temps never get over 60. Voltages and power stay steady with almost zero fluctuations, and the logs indicate nothing unusual up to the point the system restarts. No event in event viewer, other than the critical error for the system reset. Complete mystery.

 I took my PC over there today so we could try and trouble shoot the issue.  I've already DDU'd the system multiple times, as well as a smorgasbord of other things (power management, performance settings, turning off physics, running DX10... you get the point.) So, we we strait to wiping the system clean, and doing a clean install of windows. All we downloaded was the most recent GPU driver, Steam, and Warframe. Didn't matter. We swapped out the PSU, and even gave the GPU it's own power supply. Nope. We tried everything we could think of, and more that we found online, and nothing worked. Super frustrating, especially after putting so much money into this game. I can play Warframe for short stints at a time, but inevitably I know it's going to reset my system, so I don't bother anymore. I can say it's given me a chance to enjoy other games. Metro Exodus has been a lot of fun, and beautiful too! Play for hours on ultra with no hiccups. I've benchmarked and stability tested my system for hours on end, so I'm pretty confident in my equipment, but Warframe... Idk?