Question Display Issues with GPU - Need Help Diagnosing

Jul 21, 2024
10
0
10
Hello everyone,

I'm facing a frustrating problem with my PC and I'm hoping someone here can help me out.

Issue:A couple of days ago, my screen went black while playing CS2, but the sound kept playing and all LEDs and fans were still running. After shutting down and restarting, I got a "No input detected" message.

Steps I’ve Taken:

  1. Switched to Integrated Graphics: I connected my HDMI to the motherboard's port (using the integrated graphics) while leaving the GPU in place. This led to frequent blackouts and restarts.
  2. Removed the GPU: With the GPU removed, the system worked fine.
  3. Driver Uninstallation and Reinstallation:
    • Used DDU to uninstall the NVIDIA drivers.
    • Performed a clean installation of NVIDIA drivers.
  4. Checked BIOS: When I rebooted into BIOS with the HDMI connected to the motherboard and the GPU still installed, the GPU wasn't detected.
  5. Reseated the GPU: I reseated the GPU and repeatedly tried uninstalling and reinstalling the drivers. Sometimes, the GPU was detected, and I could use the PC for a few hours before it crashed again.
Observed Problems:

  • Total blackout
  • Total green screen
  • Total bright purple screen
When the GPU was detected, I tried shaking it slightly to check for loose PCIe slots or improper seating, but there was no effect.

Current Situation:I'm currently using my PC with integrated graphics, and the GPU is removed.

System Specifications:​

  • CPU: i7 6700k
  • GPU: GTX 1070 DDR5 8gb
  • Motherboard: MSI Z170A
  • RAM: 32gb RAM 3600Mhz
  • PSU: beQuiet! Straight Power 10 cm 80+ 500W

Questions:​

  1. Is this more likely a GPU failure or a motherboard issue?
  2. Are there any additional troubleshooting steps I should take?
  3. How can I definitively determine if the issue lies with the GPU or the motherboard?
Any help or suggestions would be greatly appreciated. Thanks in advance!
 
Last edited:
  • CPU: [Your CPU Model]
  • GPU: [Your GPU Model]
  • Motherboard: [Your Motherboard Model]
  • RAM: [Your RAM Amount and Model]
  • PSU: [Your PSU Model and Wattage]
What the sys specs really are? 🙄

Is this more likely a GPU failure or a motherboard issue?
Most likely GPU gave up a ghost. If it were MoBo issue, you couldn't boot to OS even.

Are there any additional troubleshooting steps I should take?

How can I definitively determine if the issue lies with the GPU or the motherboard?
Try with 2nd, known to work GPU. With this, you can rule out MoBo PCI-E x16 slot being the issue.
 
What the sys specs really are? 🙄


Most likely GPU gave up a ghost. If it were MoBo issue, you couldn't boot to OS even.


Try with 2nd, known to work GPU. With this, you can rule out MoBo PCI-E x16 slot being the issue.
My bad, I have asked GPT to reformat my question. So, here's the specs:
  • CPU: i7 6700k
  • GPU: GTX 1070 DDR5 8gb
  • Motherboard: MSI Z170A
  • RAM: 32gb RAM 3600Mhz
  • PSU: beQuiet! Straight Power 10 cm 80+ 500W

The reason I suspected a MoBo issues is that the audio and graphics driver on my integrated graphics doesnt seem to work very well as well. There's audio stuttering here and there and the screen switches off and on frequently. With that said, I will get my hands on a new GPU to test the PCIE slot. A new PSU is on its way too, so that I can test the PSU too, because the one I am currently using is 7 years old. Maybe the PSU isn't providing sufficient wattage. If PSU ain't the issue, I will return it.
 
Motherboard: MSI Z170A
MSI Z170A Gaming M5?
This beauty? Specs: https://www.msi.com/Motherboard/Z170A-GAMING-M5/Specification

If so, then i have the very same MoBo in use with my PC. 😀 (If interested, Skylake build, full specs with pics in my sig.)

A new PSU is on its way too, so that I can test the PSU too, because the one I am currently using is 7 years old.
Mediocre quality PSU and due to the age, probably already past it's usefulness. So, do try with new PSU.
Btw, make and model of new PSU is?

The reason I suspected a MoBo issues is that the audio and graphics driver on my integrated graphics doesnt seem to work very well as well.
Hmm....

Does the same issues (iGPU audio and graphics) also happen in Safe Mode?

With that said, I will get my hands on a new GPU to test the PCIE slot.
Gaming M5 MoBo has 3x PCI-E x16 slots. Have you tried them all with your GTX 1070?
 
You will need to check if there is a firmware update for the card
You do realize that OP's GPU is ~8 years old, right?

Also, GTX 1070 did see few BIOS updates, but depending which SKU it is exactly, latest BIOS release was 6-8 years ago. So, it is highly unlikely that GPU BIOS is the issue.
If OP would have RTX 40-series GPU, then maybe, one can look if there is new BIOS for GPU. But for such an old GPU, BIOS flash won't do much good. Other than giving it a good chance to completely brick the GPU, IF the update should fail for whatever reason (just like when updating MoBo BIOS).


@microwave1025 , if you're willing, then all GTX 1070 BIOSes are listed here.
Link: https://www.techpowerup.com/vgabios...&version=&interface=&memType=&memSize=&since=

Make sure to select the proper model that matches yours.
But i, personally, do not suggest updating GPU BIOS. Too high risk involved.

All the GPUs i've ever had, none have seen BIOS update. All have worked fine until obsoletion (replacement with better GPU).

To see which version BIOS you have, use GPU-Z to see it,
link: https://www.techpowerup.com/gpuz/
 
MSI Z170A Gaming M5?
This beauty? Specs: https://www.msi.com/Motherboard/Z170A-GAMING-M5/Specification

If so, then i have the very same MoBo in use with my PC. 😀 (If interested, Skylake build, full specs with pics in my sig.)


Mediocre quality PSU and due to the age, probably already past it's usefulness. So, do try with new PSU.
Btw, make and model of new PSU is?


Hmm....

Does the same issues (iGPU audio and graphics) also happen in Safe Mode?


Gaming M5 MoBo has 3x PCI-E x16 slots. Have you tried them all with your GTX 1070?
Mine is MSI Z170A Gaming M3.

New PSU(for testing, a better will be acquired if it is indeed the PSU problem): be quiet! System Power 9 500W cm, 80+ Bronze

The issues doesn't arise in Safe Mode.

For my MoBo, it has 2x PCI-E x16 slots. I have tried both, and its the same issue.

However, I have gotten a new HDMI cable today. It worked wonderfully until it crashes (Gray Screen, Dark Blue spots on the crashing screen). It only crashes when I am gaming though, right now while browsing, it works fine.

Here are my GPU-Z infos

PSU is on the way. I have updated the BIOS as well, before I saw your messages. It's still fine now. My 2x16gb RAM sticks are not the issue.
 
Update after new PSU:

I have switched the PSU. The problem still persist. Cable ain't the issue, as the new HDMI cable doesn't affect anything after many tries. PSU ain't the issue either. Unfortunately, my GPU is probably the issue.

The weird thing is, the display still ran perfectly after a few hours of not touching anything on the PC. GPU is not overheating while gaming (low 70 degrees celsius). Display is well for 30 minutes or so until it finally crashes, usually while gaming. The GPU fans kept running even when the PC successfully booted using iGPU and not under load. On the 30 minutes I had before it crashed, the fans are not spinning , if not under load.

On the matter of iGPU, after switching the HDMI cable, the screen did not switch off and on as frequent as it did with the old HDMI cable.

What I will do:
1. Get DVI cable.
2. Get a GPU for testing.
 
Mine is MSI Z170A Gaming M3.
Oh, the lesser version of the MSI Z170 Gaming series. Still good MoBo.
When i was looking for MoBo back in 2016, i did contemplate buying the high-end Gaming M7 as well. While top-end Gaming M9 ACK was way off my budget. Settled with Gaming M5 in the end.
Nice reminiscence :) , but i digress.

New PSU(for testing, a better will be acquired if it is indeed the PSU problem): be quiet! System Power 9 500W cm, 80+ Bronze
New PSU isn't an upgrade over your old PSU, but instead a downgrade.

Like i said, Straight Power 10 is mediocre quality. But System Power 9, is actually older platform and low quality PSU.

For 2nd opinion about your Be Quiet! units, look them from PSU tier list,
link: https://forums.tomshardware.com/thr...er-list-rev-14-8-final-update-jul-21.3624094/

You'll find Straight Power 10 in Tier B.
And System Power 9 in Tier C.

For proper PSU to use, i'd suggest good/great quality unit. E.g Seasonic Focus/Vertex/PRIME, Corsair RMx/RMi/HXi/AXi, Super Flower Leadex Gold/Platinum/Titanium (or any Tier A PSU in that matter).
I'm personally running 2x PRIME and 1x Focus unit.

Since PSU switch was a downgrade, i wouldn't rule out PSU issue. Only when 2nd PSU would be of good or great quality, the blame would be lifted from PSU.

The issues doesn't arise in Safe Mode.
When iGPU works fine in Safe Mode, CPU and iGPU are sound, whereby issue lies somewhere with 3rd party application, which is disabled when running Safe Mode.

When you have GTX 1070 in the system, does it act up in Safe Mode as well? Or not?

The weird thing is, the display still ran perfectly after a few hours of not touching anything on the PC. GPU is not overheating while gaming (low 70 degrees celsius). Display is well for 30 minutes or so until it finally crashes, usually while gaming. The GPU fans kept running even when the PC successfully booted using iGPU and not under load. On the 30 minutes I had before it crashed, the fans are not spinning , if not under load.
I'm suspecting high GPU temps here. Latter is easy to test;
download and run HWinfo64, link: https://www.hwinfo.com/download/

Monitor temps both on idle and under full load.

For easier full load, bench your GPU with Unigine Superposition, link: https://benchmark.unigine.com/superposition
E.g 1080p Extreme preset is good torture test for GPU. Superposition also shows/records temps. Oh, HWinfo64 also has logging feature, to record and save all the telemetry it shows.

And while ~70C may not be much for new GPU, for your old timer, it could be what brakes the camel's back. Since as hardware ages, it's performance also drops.

On the matter of iGPU, after switching the HDMI cable, the screen did not switch off and on as frequent as it did with the old HDMI cable.
I too would suggest different display port. HDMI usually is iffy when using it with PC monitors. So, try using DP instead (DisplayPort).

1. Get DVI cable.
With DVI, you have to be careful, since there are 3 forms of DVI. DVI-A, DVI-D and DVI-I.

1st you need to look which DVI cable comes out of the monitor and then, which DVI port is on the GPU.
Diff between the DVI ports are:
DVI-A - supports only analog video signal (similar to VGA)
DVI-D - supports only digital video signal (similar to HDMI and DP)
DVI-I - supports both analog and digital video signals.

And here how the female ports differ between versions:

392px-DVI_Connector_Types.svg.png


Most likely, dual-link DVI port is used by your monitor.
 
Oh, the lesser version of the MSI Z170 Gaming series. Still good MoBo.
When i was looking for MoBo back in 2016, i did contemplate buying the high-end Gaming M7 as well. While top-end Gaming M9 ACK was way off my budget. Settled with Gaming M5 in the end.
Nice reminiscence :) , but i digress.


New PSU isn't an upgrade over your old PSU, but instead a downgrade.

Like i said, Straight Power 10 is mediocre quality. But System Power 9, is actually older platform and low quality PSU.

For 2nd opinion about your Be Quiet! units, look them from PSU tier list,
link: https://forums.tomshardware.com/thr...er-list-rev-14-8-final-update-jul-21.3624094/

You'll find Straight Power 10 in Tier B.
And System Power 9 in Tier C.

For proper PSU to use, i'd suggest good/great quality unit. E.g Seasonic Focus/Vertex/PRIME, Corsair RMx/RMi/HXi/AXi, Super Flower Leadex Gold/Platinum/Titanium (or any Tier A PSU in that matter).
I'm personally running 2x PRIME and 1x Focus unit.

Since PSU switch was a downgrade, i wouldn't rule out PSU issue. Only when 2nd PSU would be of good or great quality, the blame would be lifted from PSU.


When iGPU works fine in Safe Mode, CPU and iGPU are sound, whereby issue lies somewhere with 3rd party application, which is disabled when running Safe Mode.

When you have GTX 1070 in the system, does it act up in Safe Mode as well? Or not?


I'm suspecting high GPU temps here. Latter is easy to test;
download and run HWinfo64, link: https://www.hwinfo.com/download/

Monitor temps both on idle and under full load.

For easier full load, bench your GPU with Unigine Superposition, link: https://benchmark.unigine.com/superposition
E.g 1080p Extreme preset is good torture test for GPU. Superposition also shows/records temps. Oh, HWinfo64 also has logging feature, to record and save all the telemetry it shows.

And while ~70C may not be much for new GPU, for your old timer, it could be what brakes the camel's back. Since as hardware ages, it's performance also drops.


I too would suggest different display port. HDMI usually is iffy when using it with PC monitors. So, try using DP instead (DisplayPort).


With DVI, you have to be careful, since there are 3 forms of DVI. DVI-A, DVI-D and DVI-I.

1st you need to look which DVI cable comes out of the monitor and then, which DVI port is on the GPU.
Diff between the DVI ports are:
DVI-A - supports only analog video signal (similar to VGA)
DVI-D - supports only digital video signal (similar to HDMI and DP)
DVI-I - supports both analog and digital video signals.

And here how the female ports differ between versions:

392px-DVI_Connector_Types.svg.png


Most likely, dual-link DVI port is used by your monitor.
Before I say anything about my progress. I gotta thank you. You have been guiding me throughout the process, and I will appreciate it, even if it comes to nothing.

So, I have gotten my hands on a Corsair RM850x, and nothing changes. Game still crashes.

In safe mode, gtx 1070 works well, probably because its not on load. Oh, before getting into the game, I ran superposition, here's my result. It did not crash or anything, but it was at 69 degree celsius peak.

After getting into the game, the display crashes straight away. But it did post the next time I rebooted the PC. The process repeats for a total of 3 times, before no signal was detected and did not post. I have to wait around 30 minutes, before I got another post from the GPU. GPU fans were on full spin while it crashes and while I am trying to go through post with GPU and while I am using iGPU.

CS2 did not crash while I tried playing it with iGPU.

GPU temps:
Normal temps: around 45-50 degrees celsius
On load temps: around 70 degrees celsius


Meanwhile, my screen just turned blue for a few seconds while I am tying this. Only tabs open are gmail, calendar and this forum.

My monitor is quite old, therefore I will be forced to use DVI-D port. The cable and the test GPU will arrive tomorrow.

And it crashes again, I am on my iGPU now. GPU feels hot.
 
before getting into the game, I ran superposition, here's my result.
The result links to this very topic.

It did not crash or anything
That's strange, since IF GPU would be the issue, it would surely crash during the bench, especially at 1080p Extreme preset. Since that preset is GPU torture and puts the GPU on far higher load than any game could possibly do (well, except maybe Crysis).

One option what to do, would be letting the benchmark run continuously for ~1h, rather than doing single run. Sadly, for Superposition, the "Stress" test is reserved for payed version and not available for Free version.
However, Unigine Heaven,
link: https://benchmark.unigine.com/heaven

Does have Stress version as well. Namely, once you run the benchmark, it will continuously loop around, until you stop it manually.
Running (benching) GPU for ~1h is the endurance test, to see if GPU is stable on longer period of time.

I suggest that you do the endurance test with Unigine Heaven and look if your GPU can sustain the load for ~1h or not.

In safe mode, gtx 1070 works well, probably because its not on load.
This indicates software issue, rather than hardware. Since in Safe Mode, GPU drivers aren't loaded. Instead, Win own display drivers are loaded.

Another option is uninstalling GPU drivers + DDU, but then installing version or two older GPU drivers. Since latest drivers may not be most stable ones. It wouldn't hurt to try that one out too.

After getting into the game, the display crashes straight away.
Does it happen in one specific game? E.g. CS2? Or does the crashes happen in any game?

Since if it is isolated to one specific game, it could be game issue. Namely some corrupt files in the game. Game reinstall could fix that.
 
The result links to this very topic.


That's strange, since IF GPU would be the issue, it would surely crash during the bench, especially at 1080p Extreme preset. Since that preset is GPU torture and puts the GPU on far higher load than any game could possibly do (well, except maybe Crysis).

One option what to do, would be letting the benchmark run continuously for ~1h, rather than doing single run. Sadly, for Superposition, the "Stress" test is reserved for payed version and not available for Free version.
However, Unigine Heaven,
link: https://benchmark.unigine.com/heaven

Does have Stress version as well. Namely, once you run the benchmark, it will continuously loop around, until you stop it manually.
Running (benching) GPU for ~1h is the endurance test, to see if GPU is stable on longer period of time.

I suggest that you do the endurance test with Unigine Heaven and look if your GPU can sustain the load for ~1h or not.


This indicates software issue, rather than hardware. Since in Safe Mode, GPU drivers aren't loaded. Instead, Win own display drivers are loaded.

Another option is uninstalling GPU drivers + DDU, but then installing version or two older GPU drivers. Since latest drivers may not be most stable ones. It wouldn't hurt to try that one out too.


Does it happen in one specific game? E.g. CS2? Or does the crashes happen in any game?

Since if it is isolated to one specific game, it could be game issue. Namely some corrupt files in the game. Game reinstall could fix that.
I swapped to a new GPU, and it games well, works well without crashes. Unfortunately it is indeed my GPU that has been acting up.

Before the benchmark test, I deleted drivers in safe mode using DDU and installed a driver from May.

My 1070 did survive the unigine heaven benchmark test for an hour though. But it crashed straight away after I switched windows.

After successfully rebooting, I validated CS2's and BF1's files. Crashes right away in CS2 start screen.

Honestly this has been so draining. This is the first dekstop PC that I got to play the games I wanted to play. I thought it would last longer. Oh, forgot to mention, it's 2nd hand. But the seller told me everything works well and he has showed me. I have used it for 3 months plus now. I am still thinking whether the sending the 1070 to the repair shop is worth it, or I should just get a new RX 6600 for < €200. I could just get a PS5 and keep my PC with it's iGPU for general use. Oh well, gotta wait till Black Friday.
 
for what i read, it seems that the 1070 may have some low voltage problem on some phase, if the gpu works perfectly fine in normal things you could try to undervolt it.
it will not fix the problem but probably it would let you play some games.

you could undervolt with the MSI afterburner curve.
 
I thought it would last longer.
Given that GTX 1070 was launched Q2 2016, it is 8 years old now, and i'd say it has hold up exceptionally well. Of course, nothing lasts for forever.

Back in 2016, i bought my GTX 1060 3GB, which i upgraded in Q1 2020 to GTX 1660 Ti. After 4 years of usage of my Pascal architecture GPU, it sat 3 years, until i put it into use again, namely replacing the aging GTX 760 i had in my Haswell build. And now, it is back in service. :)

I am still thinking whether the sending the 1070 to the repair shop is worth it, or I should just get a new RX 6600 for < €200.
RX 6600 is a bit better than GTX 1070.

New GPU would be better, since it's new and also has warranty, just in case something should happen with it.
But repairing the old, there's no telling when another component on it is going to die. Since 8 years for a GPU is quite a lot. If it would be CPU, it would be completely different story, since CPUs easily last 10 years. Many even 20 years. That is, if you don't OC the chip.
 
for what i read, it seems that the 1070 may have some low voltage problem on some phase, if the gpu works perfectly fine in normal things you could try to undervolt it.
it will not fix the problem but probably it would let you play some games.

you could undervolt with the MSI afterburner curve.
Hey, after reading your suggestion. I undervolted the curve by following some tutorials on youtube. Undervolting didn't solve anything, but underclocking made my sessions last longer. For example, I could last a CS2 match with underclocking before the gpu stops working, and around 10 minutes of gaming before underclocking.
 
Given that GTX 1070 was launched Q2 2016, it is 8 years old now, and i'd say it has hold up exceptionally well. Of course, nothing lasts for forever.

Back in 2016, i bought my GTX 1060 3GB, which i upgraded in Q1 2020 to GTX 1660 Ti. After 4 years of usage of my Pascal architecture GPU, it sat 3 years, until i put it into use again, namely replacing the aging GTX 760 i had in my Haswell build. And now, it is back in service. :)


RX 6600 is a bit better than GTX 1070.

New GPU would be better, since it's new and also has warranty, just in case something should happen with it.
But repairing the old, there's no telling when another component on it is going to die. Since 8 years for a GPU is quite a lot. If it would be CPU, it would be completely different story, since CPUs easily last 10 years. Many even 20 years. That is, if you don't OC the chip.
Well yeah, I should had kept in mind that an 8 year old second hand GPU comes with its own risks. I was too stubborn to acknowledge that.