Graphics Card GTX 1070 possibly dead but I'm unsure. Monitor blank when PC is turned on, would appreciate any help possible!

[ItsmeJosh]

Hello! This is my first build and post so I will try to be as concise as I can! The parts were bought during the days leading up to and including black friday 2016 however I had just kept on to the parts until exams were finished. I'm writing this in hope that this issue would be solved without having to buy new parts, but it does seem unlikely, and if not how others would handle the problem.

Apologies if this is in the wrong thread!

Build..
CPU: Intel i5 6600k
CPU cooler: Cooler Master 412s
GPU: Palit GTX 1070 Super Jetstream
Motherboard: MSI Z170A Krait Gaming
Memory: 2x HyperX Savage 4GB - DDR4 (8GB)
PSU: Rosewill Capstone 750W Gold

Parts bought: November 2016
Build completed: 19/01/17
Build broke: 2/02/17
The system is running on Windows 10 Pro

So the current situation is;
When the Graphics card is installed in the PC the monitor would remain blank however I think it is receiving a signal since the monitor isn't giving the 'no signal' popup. I had also noticed the boot time, if it is running, is longer which i timed using the LEDs on my keyboard. They would flash before a few seconds of being off then have its normal stable light, all while the monitor remains blank. Even the initial MSI popup to enter BIOS at the start won't show.
When I remove the GPU, the startup to my desktop is normal and the monitor is displaying as it was before.
Everything else appears to be working; fans are spinning, LEDs are lit even the GPU LEDs.

The events leading up to this was that I was playing Witcher 3 for the first time, at ultra 1080p capped at 60fps with vsync, however after about 4 hours the game showed artifacts, froze and then the computer crashed. I of course tried to reload the game however crashes occurred within seconds of actual gameplay as in I loaded into my saved file before crashing. Crashes would range from game freezing then crashing to desktop to freezing the whole computer giving me the blue screen of death causing the computer to restart after gathering data. Although most common crash is just the computer freezing and I would be forced to restart using the power button on the PC. Also following some crashes to the desktop, artifacts would show up at the desktop. Attempts were made to fix the game (updating drivers, altering the video quality and framerates, increasing gpu fans) without having to install third party software.

This was during late last night so I had just slept and attempted again this afternoon only to see the monitor won't display anything.

So far my attempts in fixing this have been..
- Making sure all the wires are fitted
- Replacing the power supply wire to the GPU
- Uninstalling Nvidia drivers ( couldn't reinstall without GPU added back)
- Uninstalling softwares made on the day of the game crashes
- Moving the GPU to the second PCIE slot
- Switching monitors and output types (HDMI and DVI-D Dual Link)
- Disabling Intel HD graphics in Device manager
I have a feeling maybe something in the BIOS I can change but I would probably need the GPU installed.

So my thoughts..
At the moment, I am 80% it is the video card that had died, 15% it being the power supply, 5% Motherboard.
The video card is most likely the problem since it is stopping my monitor displaying anything and the artifacts were popping up. I also think it may be causing a mini crash of some sort when the computer is just booting up since it causes my keyboard LED to flicker then it turns off before the normal LED stable light. The motherboard also has an 'EZ LED debug' light for the CPU, RAM and GPU and that when the GPU is installed it causes it to light up on boot up however it does turn off afterwards, even though there is still no display.
The reason I think it may be the power supply is that perhaps it isn't supplying enough power to the GPU which may be causing the artifacts. The wire is tightly fitted, in fact I couldn't even comfortably remove the 6 pin wire from the PSU and had to use another wire to check that wasn't the problem. Also this PSU is my newest part since my old one was making some clicking noise which i just replaced immediately. I don't think the GPU broke by overheating, I kept quite an aggressive fan curve (80% by 50 degrees, 100% at 60 degrees) since I'm paranoid about that and my room is quite cool where the idle temperatures would be around 25-30 degrees for the CPU and GPU. I have not seen my GPU go above 50 degrees however it most likely have during the 4 hour gameplay but I don't think it would have gone to the point where it would break the GPU.
Motherboard may be a problem if the PCIE slots alone are somehow damaged however I tried both slots and still no display but I think it may still be possibility.
CPU and everything else should be okay since it's able to boot up normally and work smoothly without the GPU installed.

What I don't understand is how the PC could at least get to the desktop to not able to show up anything without use overnight. I really hope i'm just being stupid and missing to do something blindingly obvious. It should probably be noted that the graphics card was fine before this however I had not played any demanding games before this either or ran any test benchmarks. I was just playing games like League of Legends, Dota2, Mafia 3 ( free with mobo ), Brothers: A tale of two sons, Magicka, so nothing really to stress the GPU.

In summary;
GPU may be dead after playing graphically demanding game causing artifacts and crashing
PSU supplying power to GPU may be defective
[strike]Motherboard PCIE slots may be defective[/strike]

I would like other people's insights into this since I can't be 100% sure on what the defect is as I don't have anything to install the GPU in or another GPU to test into my build and if it's not the GPU and I end up getting a new one and it still doesn't work I'd have to question the other parts or maybe I was just unlucky or maybe the PSU somehow surged power into the Graphics Card.

As I've said it's my first time building and I was excited to finally be playing games so it really sucks that I've got a problem this early so thanks in advance to those who try to help

EDIT1: PCIE slots not likely broken, just came to realisation that it powers the fans and LED on my graphics card so data transfer part is most likely fine as well. So either the PSU killed my graphics card or the graphics card is just weirdly faulty, breaking down under a somewhat demanding environment.
EDIT 2: GPU LED colour settings are kept so I'm fairly sure the computer is booting up fine and PCIE slots 100% not defective. Really just a lack of display from the graphics card.

Edit 3: GPU does not seem to be quite dead!
I was about to call for an RMA but decided to test the GPU once more and to my surprise it managed to boot up with display on to the monitor?? However the resolution that it booted into was much lower than the native resolution of 1920x1080 and with the wrong aspect ratio.
Just before this, all I did was change a few things in the BIOS such as boot priority and disabling 'windows 7 installation' and a few more things under peripherals. Regardless if I accidentally tweaked something in BIOS to make it boot up with display or not, it at least gave hope that the graphics card still had a bit of life.
Changing the resolution to 1080p caused some artefacts across the desktop which only disappeared under lower resolutions. Initially lead me to think the GPU was faulty but now I think that it really is just not receiving enough power to go for those higher resolutions.

Following this I had decided to download the latest Nvidia Drivers for 1070 Windows 10 64-Bit; Version 378.49 released 2017.1.24. without GeForce Experience and clean installation. However after installation the black screen came back. Which lead me to believe perhaps the latest drivers were causing this. And so after rebooting without GPU and uninstalling I rebooted with GPU and I installed the 376.19 Version released 2016.12.5, without GeForce Experience and Nvidia HD Audio driver and clicked clean installation since I didn't really need it. I rebooted and managed to finally get a display with a driver installed however resolution was low, and no Nvidia control panel. Opening Thundermaster it was clear they couldn't detect my GPU as the values were all 0 for clock and speed and memory etc. I then went into Device Manager and updated Nvidia drivers through that by searching automatically online then I rebooted and here I am now with the latest driver installed. Weirdly booted completely normal with the GPU still installed although i'm scared that this may just be a one off and the black screen returns instantly after restarting.

Not sure if GPU is 100% healthy, perhaps PCIE connection is being interrupted by the GPU sagging. During some previous load ups, the motherboard sometimes takes a while to detect the GPU so maybe there's compatibility issues between the two. EZ debug lights still pops up during boot up but maybe it's just a normal delay in trying to detect the card
Not really sure what these findings confirmed. At least the card isn't fried as I initially though but this only means there are other problems. Hopefully this is all just bad drivers that caused screen tearing, artefacts, no display on monitor at boot up, crashing of games, blue screen of deaths and inability to detect the GPU quickly sometimes.

So now currently this computer is on its first successful boot up. Drivers installed are Nvidia Graphics and 3D Vision Driver 378.49, 3D Vision Controller Driver 369.04 and Nvidia PhysX system software 9.16.0318

Also funnily during this I learnt that more than one display outputs are put into the same screen, the computer would think there are more than 1 monitor available causing a black screen on load up but with just a mouse. Fixed by just blindly typing password to get into desktop and removing one of the display output wires. Hopefully next edit is the last with me saying ' pc has restarted successfully multiple times with display on monitor'

[strike]Edit 4: It's been over 24 hours since this problem initially started and basically spent the whole day reading and trying to figure out what's happened but the computer seems to be running fine again but who knows how long for. Not entirely sure what caused the malfunctions whether it's just the drivers, the GPU or a combination of many things but for now it's okay. If nothing else, this post serves as a log of one of my first breakdowns of owning a PC even though I don't understand why.[/strike]

 

[I3lue1]

This is the first time that I hear of a GPU "breaking". Most of them are broken from the factory, some stright up break after some time or overheating or etc. Yours broke up after 4H of gaming. A surge would fry it on the spot (no artefacts, just pc shutdown and black screen); gpu problems would mean you would find/ feel them within 30 mins of gaming; temps, eell usually cards reset the Pc or throttle when they have a temp issue.

My guess is that either the throttling didn't work, or the temp sensor was broken, or somehow the card got overclocked to a breaking point and stayed there (least likely).

Somehow find another Pc and try it in that one.

I hate to be the bearer of bad news but your card seems like it's dead. If so, the flip side would be that you can probably RMA it (you have warranty right?) and I'm pretty sure that the next one won't be an issue. Also I read somewhere that people had some problems with the msi 1070 aero (because of Micron VRAM instead of Samsung) and they fully fixed them with a videi card firmware update.

Good luck and come back to tell us what happened.

 

[ItsmeJosh]

This is the first time that I hear of a GPU "breaking". Most of them are broken from the factory, some stright up break after some time or overheating or etc. Yours broke up after 4H of gaming. A surge would fry it on the spot (no artefacts, just pc shutdown and black screen); gpu problems would mean you would find/ feel them within 30 mins of gaming; temps, eell usually cards reset the Pc or throttle when they have a temp issue.

My guess is that either the throttling didn't work, or the temp sensor was broken, or somehow the card got overclocked to a breaking point and stayed there (least likely).

Somehow find another Pc and try it in that one.

I hate to be the bearer of bad news but your card seems like it's dead. If so, the flip side would be that you can probably RMA it (you have warranty right?) and I'm pretty sure that the next one won't be an issue. Also I read somewhere that people had some problems with the msi 1070 aero (because of Micron VRAM instead of Samsung) and they fully fixed them with a videi card firmware update.

Good luck and come back to tell us what happened.

Thanks for the reply System miraculously running again, if it happens again and I can't do the same thing to 'hotfix' it then I will probably just try to send the card to be fixed. I think I have warranty for a year so hopefully no real loss!

 

[I3lue1]

So wait: it's magically running? This is impossible. Pcs don't "heal" like animals. (no sarcasm)

Artefacts in the desktop means some very serious issues. How do you define this as normal?

Can you do some benchmarks? Pasmmark, Aida64, Furmark, 3Dmark, etc and run one of them for at least 15-30 mins?

It COULD be your PSU, but most PC components (including video card and PSU) are plug and play; which means you put them in and they work.

All GTX 1070 are rated for 150W, which isn't that much. I had a situation in which my PSU wasn't good enough for my card and what happened was that during heavy gaming my PC would shutdown or restart.

I've never heard of your PSU, but if it's 80% Gold certified then it should easily handle the load (unless it's faulty).

Honestly I don't consider this issue fixed.

Edit: I just checked:
http://www.palit.com/palit/vgapro.php?id=2629&lang=en&pn=NE51070S15P2-1041J&tab=sp

This is rated for 170W which is still very acceptable. I had a gtx 780 ti (250W consumption) on a fully noname 1000W PSU; I bought the PSU for about 20-40$ from the PC store closest to me and I had absolutely no problems running games in high demand on it (Mirror's Edge Catalyst, Shadow Warrior 2) or doing any kind of benchmark.

Please get back to us with those benchmark results and how exactly is it working now (are you gaming on it? do you still have artefacts? etc).

 

[ItsmeJosh]

So wait: it's magically running? This is impossible. Pcs don't "heal" like animals. (no sarcasm)

Artefacts in the desktop means some very serious issues. How do you define this as normal?

Can you do some benchmarks? Pasmmark, Aida64, Furmark, 3Dmark, etc and run one of them for at least 15-30 mins?

It COULD be your PSU, but most PC components (including video card and PSU) are plug and play; which means you put them in and they work.

All GTX 1070 are rated for 150W, which isn't that much. I had a situation in which my PSU wasn't good enough for my card and what happened was that during heavy gaming my PC would shutdown or restart.

I've never heard of your PSU, but if it's 80% Gold certified then it should easily handle the load (unless it's faulty).

Honestly I don't consider this issue fixed.

Edit: I just checked:
http://www.palit.com/palit/vgapro.php?id=2629&lang=en&pn=NE51070S15P2-1041J&tab=sp

This is rated for 170W which is still very acceptable. I had a gtx 780 ti (250W consumption) on a fully noname 1000W PSU; I bought the PSU for about 20-40$ from the PC store closest to me and I had absolutely no problems running games in high demand on it (Mirror's Edge Catalyst, Shadow Warrior 2) or doing any kind of benchmark.

Please get back to us with those benchmark results and how exactly is it working now (are you gaming on it? do you still have artefacts? etc).

Hey thanks for replying back!

Problem came back this morning but seemed somewhat working last night with small problems, i'll post videos on youtube and link it here once it's done uploading on the situation, no speech or anything, just a simple video of my setup booting up with and without the graphics card installed. Coincidentally did my first benchmarks using novabench and furmark (first two that were free that I saw) on it just to see how it was but unfortunately do not have much of the data saved but i specifically remember on furmark that the max power consumption was just a bit over 100% TDP so I think it is receiving the power it needs and the PSU is fine I think. However artefacts showed roughly midway through furmark which remained even once the test was completed throughout the desktop and that's probably the GPU or drivers busting up again, i'll try to go back to the previous drivers and see if that works. I have random pictures taken from my phone of the situation so I'll have links to soon! Just recently came home but I'll be trying to figure this out again until I pass out for the night.


Thanks again!

edit1: I attempted witcher3 again one last night but it froze, had some artefacts then crashed to desktop after what feels like 10 seconds of moving around so as per usual then I just called it a night after that and this morning back to the same old no display on monitor business!
Uploads should be done within an hour of this update, they are short videos but I have ADSL internet and uploads speeds are very sub par.

 

[ItsmeJosh]

I've included some videos and images of the boot ups that I've described.
Just for clarity, a very brief timeline;

• 
  
  Played witcher3 until crashing 02/02/17 at about 1 am headed to sleep.
  
  Computer deemed broke on 02/02/17 at about midday however somehow managed to boot up with graphics card with some tweaking after some tweaking on 03/02/17 at about 1 am. Ran some benchmarks (furmark and novabench) with questionable results, played witcher3 for about 10 seconds before it crashed and called it a night at about 2 am.
  
  Checked again at 03/02/17 at about 2 pm and same problem as before.
  

Maybe give an uncertainty of an hour for each time provided.

Links:
Boot Up with GPU installed ( Video length: 1 minute 1 second)
[video="https://www.youtube.com/watch?v=1Q3n42ZfzPo"][/video]
notes:
Whilst the GPU lights were turned of the 'EZ debug LEDs' were on indicating VRAM was not detected or faulty (paraphrased from other posts) However debug lights turns off once the GPU LED lights turns on some moments later.
PCIE slot most likely okay, it was able to power on GPU LEDs instantly when power was turned on

Boot Up without GPU ( Video length: 37 second )
[video="https://www.youtube.com/watch?v=Qt_LWRfrvms"][/video]

Aftermath of Furmark Benchmark when GPU miraculously powered on the display ( Video length:1 minute 7 seconds)
[video="https://www.youtube.com/watch?v=c6JiTgkYOHY"][/video]
Lots of artefacts pretty much but resolved when rebooted which is how the boot up normally should be

Some scavenged benchmark results (novabench and furmark), foolishly didn't record them but here is what I have during the time the computer decided to work for the night

Novabench mark
http://imgur.com/fQyce32.png

FurMark benchtest using preset 1080 (FHD)
http://www.ozone3d.net/gpudb/score.php?which=369869&p=jj
http://i.imgur.com/jUlp7KX.jpg
Due to artefacts may be hard to read, here's what I can make out that seems important;
Max GPU Temp: 39 degrees Celsius
Min FPS: 48 Max FPS:105 Avg FPS:78

Not too sure yet what the normal scores are, during both tests the GPU whining fairly audibly in my opinion.

GPU-Z results from Furmark Test, I believe this was a separate test where I had chose the GPU stress test button instead of the preset. I didn't run it for very long. Roughly about 35 seconds ish
http://i.imgur.com/EiNDxMU.gif
What I can take from this is that the PSU is able to give the GPU enough power seen by the graphs when under load so I think I can finally start ruling that out as the main issue.

This is just a bonus pic from my phone of the aftermath of furmark ( 1080 preset )
http://i.imgur.com/9c031vY.jpg

Anyway what I think the problem may be is that the GPU is slightly faulty or that the latest drivers are faulty from Nvidia (Version 378.49 released 2017.1.24). I can rule out PSU from the furmark test and the PCIE slot being faulty from the motherboard since it's able to provide power right away but somehow the GPU and motherboard becomes disconnected but reconnects after some time? Maybe the two have compatibility issues.

I feel like i'm just repeating everything I've said before

 

[I3lue1]

I also have a 1070 and my score is 5900, min fps 98, max 100, average 99 and temperature of 72 degrees. So, the GPU is barely scraping by given the faulty memory.

Yes, you kinda are repeating yourself . I will do it also:
There's nothing that can be done. Just RMA the card. Trigger the RMA, wait for 1-2 weeks until you get a new one, happy gaming.

This is not a "my CPU fan broke and my PC resets from the heat issue". This is a faulty component issue on a PCB that was printed and manufactured by machines which make very tiny actions (which implies that you can't really go off de-soldering the memory and solder back new memory chips).

You could probably find a way to restrict your GPU from using the faulty memory, but that would take vast amounts of time, could still cause issues, and will leave you with a lot less memory to work with and make all GPU activity slower since it would constantly have to avoid that area. Gaming with such a video card would be like tying your legs together and then going on a race with people that didn't tie their legs: illogical and impractical. So, RMA.

 

[ItsmeJosh]

I think you're right about the memory problems!

I was able to repeat and reinstalled drivers. Rolled back to Version 376.33 released 2016.12.14. Initially blacked screened but reinserted graphics card and it booted up. Instantly tested using FurMark 1080 preset only to still show artefacts as it progressed and weird shapes. Attempts involved slowing down the desktop where even shut down took an extra bit of time. Other attempts involved the test not being completed as computer crashed and had to hard reset by holding down the power button. Some artefacts initially didn't show straight after the test but after a few seconds and ended up freezing the computer which forced me to hard reset.

Attempted FurMark with Integrated graphics card and as slow as it was, it didn't have any artefacts or weird shaping.

Noticed that the graphs on GPU-Z has the memory controller load just randomly dropping to 0 followed by the power consumption dropping to around 20% whilst the GPU load was still 100% for a 1-2 seconds then crashing.

Thanks for the help I will try to RMA the card now and update here on whatever happens!

Update: Called up and they're sending me a new card by tomorrow, hopefully it really is just the graphics card!

 

[I3lue1]

It's just the graphics card, don't worry.

 

[ItsmeJosh]

Got the replacement graphics card! Everything working better than before, benchmarked using furmark and completed it with no artefacts. Computer booting up normally with display on the monitor. Witcher 3 back to normal too although i'm playing at a step down on graphics just because I don't wanna over-strain the GPU on the long hours of gameplay. Thanks for the help

 

[I3lue1]

No worries; I'm glad that everything turned out fine and you got a replacement blazing fast. You can strain it as much as you want as long as you have good cooling on your GPU. A rule of thumb is to try keep it at MAXIMUM 75 degrees.

Enjoy your new machine .

 

[ecsstatick]

Thank goodness I found this thread. Literally everything you described here is happening to me after playing (guess it...) witcher 3. Although I can't even get my machine to boot up with artifacts. I did the breadboarding, etc, and put in an old 760 I had lying around and everything works. Just put in my RMA before reading this thread. Glad I did that. Hoping it's more a faulty card rather than anything else, which it seems to be. I don't think I have much lack of cooling, but maybe I should consider more? Who knows.

Thanks for the thread of information. I found it highly useful.

 

[I3lue1]

@ ecsstatick: it's hard to say what exactly is the issue it could be the card manufacturer (do you have the same card as ItsmeJosh?) or the game itself that somehow overrides the GPU failsafe mechanism against burning itself. It could be that you've blown some capacitors (which can be replaced by most PC service shops).

RMA-ing it was probably the best move since opening the card itself (even in a PC service shop) will void the warranty (unless said shop is the one from which you bought the card).

Enjoy your new toy once it gets back to you .