Question Rtx 3080 stopped working.

[LightingBlade360]

Rtx 3080 stopped working suddenly.Everything turns on but no signal to monitors and no light to keyboard and mouse. When I removed the card and I connected the monitors to integrated graphic it worked fine and I went ahead connected my old gpu to the same PCIe slot and it worked fine too.
Update: the GPU started working again. After playing and streaming for 2hours the screen went black and started showing the same symptoms again. What should I do?

 

[tecmo34]

What specific RTX 3080?

What are the temps on the card and CPU?

What is your PSU, make, model, age?

It could be the card or it could be the PSU is starting to die and the power requirements for the 3080 causing it to shut down.

 

[LightingBlade360]

P

What specific RTX 3080?

What are the temps on the card and CPU?

What is your PSU, make, model, age?

It could be the card or it could be the PSU is starting to die and the power requirements for the 3080 causing it to shut down.

Palit rtx 3080 g

What specific RTX 3080?

What are the temps on the card and CPU?

What is your PSU, make, model, age?

It could be the card or it could be the PSU is starting to die and the power requirements for the 3080 causing it to shut down.

Palit rtx 3080 game rock.
Temp were 60 on CPU and GPU while playing game
Brand new PSU Deepcool 850 gold+

 

[tecmo34]

We, as Tom's Hardware , didn't give that a very good review. It could be your PSU or the GPU, but I would lean toward the PSU first as a potential issue since no issues with no GPU or another GPU of lower power requirements.

Do you have an option to swap PSU's or try the card in another system?

 

[LightingBlade360]

We, as Tom's Hardware , didn't give that a very good review. It could be your PSU or the GPU, but I would lean toward the PSU first as a potential issue since no issues with no GPU or another GPU of lower power requirements.

Do you have an option to swap PSU's or try the card in another system?

I plugged my old GPU it worked fine.
And stress test the whole system without the GPU everything works fine.

 

[kurdtnz]

What is your old GPU? (that is working.)

 

[LightingBlade360]

What is your old GPU? (that is working.)

We, as Tom's Hardware , didn't give that a very good review. It could be your PSU or the GPU, but I would lean toward the PSU first as a potential issue since no issues with no GPU or another GPU of lower power requirements.

Do you have an option to swap PSU's or try the card in another system?

GTX 1070

 

[~cw]

I went from an EVGA GTX1070 SC to a 3080 FE and the power draw requirements are significantly higher. I also replaced my PSU (went for a Corsair RM750) to avoid any potential PSU performance issues due to its age. I've had issues in the past with PSUs becoming less able to deal with transient power spikes and very high current draw as they've aged, caps have dried out, etc.



If you reboot in Safe Mode, does the GPU reliably work for more than several hours? Obviously you can't test 3D acceleration, but you can at least run Furmark and do some CPU benchmarking to keep thermals up at gaming level temperatures.

If the GPU output still displays after 3/4 hours when it would have cut off in normal mode, it's one indication that its 3D acceleration features may not be working well. Could be hardware, but could just as well be driver problems.

From Safe Mode, try a full DDU driver removal then a fresh install of the latest drivers. Doing this cured some black screen issues I had a couple of years ago with a GTX 1070.

If that makes no difference, I would boot back into normal mode, then keep a close eye on the card's power draw and temperatures using the latest version of HWiNFO64 running while you use the PC. Be sure to enable logging first in HWiNFO, then you can review the logs with the developer's accompanying log viewer app (called "Generic Log Viewer" if you do crash and have to reboot. You may find that the PSU is shutting down power supply to the GPU if it's attempting to demand more current than it is capable of delivering in a short period of time (a transient spike).

You could also try installing MSI Afterburner and reducing the Power Limit to 80% or 90% to further see if it's related to unachievable power demands by the GPU on the PSU.

Other apps like OCCT and RealBench can apply simulated benchmark and stress test loads to the GPU as well as the CPU, so you can stress test your machine at similar levels to demanding gaming or desktop editing/rendering.


Do you also have integrated graphics which you could enable in Windows, so you could plug your monitor to that if the 3080 shuts down again? You could check if the entire PC is frozen or if it's just the card shutting down suddenly.

If after all that fault-finding comes up with nothing, I'd be tempted to RMA the GPU. It'll be a hassle if you leave it until the warranty period is up. Good luck...

 

[LightingBlade360]

I went from an EVGA GTX1070 SC to a 3080 FE and the power draw requirements are significantly higher. I also replaced my PSU (went for a Corsair RM750) to avoid any potential PSU performance issues due to its age. I've had issues in the past with PSUs becoming less able to deal with transient power spikes and very high current draw as they've aged, caps have dried out, etc.



If you reboot in Safe Mode, does the GPU reliably work for more than several hours? Obviously you can't test 3D acceleration, but you can at least run Furmark and do some CPU benchmarking to keep thermals up at gaming level temperatures.

If the GPU output still displays after 3/4 hours when it would have cut off in normal mode, it's one indication that its 3D acceleration features may not be working well. Could be hardware, but could just as well be driver problems.

From Safe Mode, try a full DDU driver removal then a fresh install of the latest drivers. Doing this cured some black screen issues I had a couple of years ago with a GTX 1070.

If that makes no difference, I would boot back into normal mode, then keep a close eye on the card's power draw and temperatures using the latest version of HWiNFO64 running while you use the PC. Be sure to enable logging first in HWiNFO, then you can review the logs with the developer's accompanying log viewer app (called "Generic Log Viewer" if you do crash and have to reboot. You may find that the PSU is shutting down power supply to the GPU if it's attempting to demand more current than it is capable of delivering in a short period of time (a transient spike).

You could also try installing MSI Afterburner and reducing the Power Limit to 80% or 90% to further see if it's related to unachievable power demands by the GPU on the PSU.

Other apps like OCCT and RealBench can apply simulated benchmark and stress test loads to the GPU as well as the CPU, so you can stress test your machine at similar levels to demanding gaming or desktop editing/rendering.


Do you also have integrated graphics which you could enable in Windows, so you could plug your monitor to that if the 3080 shuts down again? You could check if the entire PC is frozen or if it's just the card shutting down suddenly.

If after all that fault-finding comes up with nothing, I'd be tempted to RMA the GPU. It'll be a hassle if you leave it until the warranty period is up. Good luck...

Did all of that still nothing. As long as the GPU is connected to motherboard PC wouldn’t boot.

 

[LightingBlade360]

Did all of that still nothing. As long as the GPU is connected to motherboard PC wouldn’t boot.

I have been using the pc with my old CPU GTX 1070 for a week now. I haven’t faced any problems. Using same PCIe slot and same power connector.

 

[Satan-IR]

All this might unfortunately point to a faulty or dead graphics card (3080).

By trying the card in another known good system, as tecmo34 said above, you can somehow make certain if the card has problems.

 

[LightingBlade360]

All this might unfortunately point to a faulty or dead graphics card (3080).

By trying the card in another known good system, as tecmo34 said above, you can somehow make certain if the card has problems.

I RMA the card.

 

[LightingBlade360]

I RMA the card.

Update: retailer informed me that GPU was faulty and they will be sending me a new rtx 3080 and the replacement wouldn’t count has a warranty because the GPU was dead on arrival.

 

[zen.sends]

Update: retailer informed me that GPU was faulty and they will be sending me a new rtx 3080 and the replacement wouldn’t count has a warranty because the GPU was dead on arrival.

You mean they are not giving you a warranty on the new card they are sending you?

 

[spentshells]

That is great to hear!

 

[LightingBlade360]

You mean they are not giving you a warranty on the new card they are sending you?

No they are giving one year warranty

 

[zen.sends]

Oh that is such a relief