[SOLVED] 2080ti failing heaven basic benchmark jumps to 100% and crashes

[Wobble135]

motherboard: z390 elite aorus

cpu: i9 9900k(never overclocked)

gpu: evga 2080ti XC2 ultra(11gb)

psu: corsair rm1000i

ram: 32gb 4 sticks(4x8) ballistic 3200 ddr4

hp 25x: 144hz

In games clear makes randomly lines and spots as intensity increases it worsens. Eventually crashing most of the time. Will pass 2d benchmark but fails 3d benchmark (something about Direct 12x) cpu passes stress test perfectly, ram is good. Have tried uninstalling drivers and reinstalling also done the same with the OS. Would love to know if anyone knows what is wrong.

 

[Ralston18]

Try to find or otherwise identify potential causes.

Start with Reliability History (more user friendly) and Event Viewer (less user friendly) to look for error codes, warnings, and even informational events that precede or correspond with the crash times.

Errors can be clicked for more information or detail. That information or detail may or may not be useful.

Likewise use Task Manager and Resource Monitor (use both but only one at a time to observe system performance and activities that lead up to a crash.

Watch when the system is not crashing and then do whatever you do when the system begins crashing.

How old is the PSU? Heavy use for gaming, video editing, or even crypto mining?

Lastly you can use Process Explorer (free via Microsoft) to also see what all is going on.

https://docs.microsoft.com/en-us/sysinternals/downloads/process-explorer

 

[Eximo]

Sounds like GPU artifacting.

Step 1) Power down
Step 2) Remove the GPU
Step 3) Re-install GPU

You would be surprised how often it works. Heavy GPUs and lots of temperature cycles, they can work their way into an intermittent connection. If you have severe GPU sag, that is another concern. Might be worth taking the heatsink off in that situation, re-doing the thermal compound and putting it back together, that can relieve some of the bowing which makes parts of the PCIe slot lose connection.

If this GPU is new to you, return it if possible.

 

[Wobble135]

Try to find or otherwise identify potential causes.

Start with Reliability History (more user friendly) and Event Viewer (less user friendly) to look for error codes, warnings, and even informational events that precede or correspond with the crash times.

Errors can be clicked for more information or detail. That information or detail may or may not be useful.

Likewise use Task Manager and Resource Monitor (use both but only one at a time to observe system performance and activities that lead up to a crash.

Watch when the system is not crashing and then do whatever you do when the system begins crashing.

How old is the PSU? Heavy use for gaming, video editing, or even crypto mining?

Lastly you can use Process Explorer (free via Microsoft) to also see what all is going on.

https://docs.microsoft.com/en-us/sysinternals/downloads/process-explorer

The psu is like a year old not used for mining a decent amount of gaming but on a game that doesn’t take much of the gpu capacity.
Just ran a gpu test again . The results were 3d graphics test result: failed
“An error occurred during the DX12 3D test” this time also failed the 2D graphics test “error verifying data in GPU memory”
But still works perfectly when no under high load game or benchmark. Can play low end games with hardly any problems at 1080p 144hz. When looking at resource monitor when running the benchmark everything is fine except for the gpu it spikes to 100% crashes down. Can still use it to browse web do anything on pc works normal until I play a high end game or run benchmark. Even when running benchmark temp doesn’t go over 70C

 

[Wobble135]

Sounds like GPU artifacting.

Step 1) Power down
Step 2) Remove the GPU
Step 3) Re-install GPU

You would be surprised how often it works. Heavy GPUs and lots of temperature cycles, they can work their way into an intermittent connection. If you have severe GPU sag, that is another concern. Might be worth taking the heatsink off in that situation, re-doing the thermal compound and putting it back together, that can relieve some of the bowing which makes parts of the PCIe slot lose connection.

If this GPU is new to you, return it if possible.

I have tried that. Even tried it in different pci slot.It seems odd that the card is only pulling max 170W when running benchmark shouldn’t 2080i be closer to 300w

 

[Eximo]

What about GPU temperatures?

Highly possible there is a hotspot on the GPU causing a problem, or even overheating memory or VRMs.

 

[digitalgriffin]

motherboard: z390 elite aorus

cpu: i9 9900k(never overclocked)

gpu: evga 2080ti XC2 ultra(11gb)

psu: corsair rm1000i

ram: 32gb 4 sticks(4x8) ballistic 3200 ddr4

hp 25x: 144hz

In games clear makes randomly lines and spots as intensity increases it worsens. Eventually crashing most of the time. Will pass 2d benchmark but fails 3d benchmark (something about Direct 12x) cpu passes stress test perfectly, ram is good. Have tried uninstalling drivers and reinstalling also done the same with the OS. Would love to know if anyone knows what is wrong.

Is this a new or used GPU? If I were to guess, this is a reflashed used Mining card on the edge of dying and you got fleeced to offload his dead card.

Try downclocking memory and gpu core clocks and report back.

 

[Wobble135]

No

What about GPU temperatures?

Highly possible there is a hotspot on the GPU causing a problem, or even overheating memory or VRMs.

Not that I know of temp stats are cool

 

[Wobble135]

Is this a new or used GPU? If I were to guess, this is a reflashed used Mining card on the edge of dying and you got fleeced to offload his dead card.

Try downclocking memory and gpu core clocks and report back.

Yes I tried downclocking memory and gpu doesn’t do much. The event log says “failed to write gpu buffer object(-5) block 1/100”. “error writing to gpu memory”

 

[Eximo]

So is this a used card? Looking more and more like it needs a flash to the original vBIOS or there is something horribly wrong with it.

 

[digitalgriffin]

Yes I tried downclocking memory and gpu doesn’t do much. The event log says “failed to write gpu buffer object(-5) block 1/100”. “error writing to gpu memory”

Last thing you can do is go into safe mode (Hold down left shift key on restart menu selection trouble shoot options...) and run DDU (Display Driver Uninstall) then reboot and reinstall.

If that does nothing for you, and you aren't under warranty, I would risk disassembling the card and checking thermal pads and repaste it. Do this at your own risk.

Other than that, you are SOL I'm afraid.

 

[Wobble135]

So is this a used card? Looking more and more like it needs a flash to the original vBIOS or there is something horribly wrong with it.

Yes it is used. I have updated bio to most recent update. I got it from my uncle he had it for a year or so before(used for gaming) when I bought it from him he ran it through a benchmark on his pc in front of me and it did fine.

 

[punkncat]

If your uncle purchased this, follow up with him to see if warranty applies.
If not, try the deep clean, repaste, thermal pad replacement and pray.

Bad time to be out a card of that level.

 

[Wobble135]

Last thing you can do is go into safe mode (Hold down left shift key on restart menu selection trouble shoot options...) and run DDU (Display Driver Uninstall) then reboot and reinstall.

If that does nothing for you, and you aren't under warranty, I would risk disassembling the card and checking thermal pads and repaste it. Do this at your own risk.

Other than that, you are SOL I'm afraid.

One last thing I noticed the gpu and mem(gpu) clock jump up and down as its lagging out does this mean anything. Is there anyway it could be not getting enough power

 

[Ralston18]

If you have a multi-meter and know how to use it (or know someone who does) then you can do some testing on the PSU.

Not a full test because the PSU is not under load but any voltages out of tolerance make the PSU very suspect.

https://www.lifewire.com/how-to-manually-test-a-power-supply-with-a-multimeter-2626158

Remember that PSUs provide three different voltages ( 3, 5, and 12) to system components. A problem with any voltage will cause problems.