Question Hard lockups since acquiring a 3080Ti

[gamerk316]

About two weeks ago, I purchased a new GPU (model GV-N308TGAMING OC-12GD) to replace my old 1080Ti.

The new GPU was installed in the same PCI-E slot without issue, and both PCI-E 8-pin power connectors were installed without issues. Installed latest WHQL drivers and card appears fine at a glance.

However in the three weeks I've had the new GPU installed my system has had two separate instances of hard lockups while watching videos online.

Both times the system hung while browsing Youtube via Google Chrome with HW Acceleration of videos active. [Note: I've since turned off HW Acceleration for testing purposes.] As of yet, no issues have occurred while in game, although I haven't been gaming as much due to work hours.

My system has never had problems in the three years prior with my previous GPU, which makes me suspect some instability with either the GPU or the current GPU drivers. But I don't have any clear evidence of a physical defect with the GPU either. My question for now is more about confirming if there is a problem with the GPU/Driver, and how to go about proving it.

Any advice?

 

[Third-Eye]

About two weeks ago, I purchased a new GPU (model GV-N308TGAMING OC-12GD) to replace my old 1080Ti.

The new GPU was installed in the same PCI-E slot without issue, and both PCI-E 8-pin power connectors were installed without issues. Installed latest WHQL drivers and card appears fine at a glance.

However in the three weeks I've had the new GPU installed my system has had two separate instances of hard lockups while watching videos online.

Both times the system hung while browsing Youtube via Google Chrome with HW Acceleration of videos active. [Note: I've since turned off HW Acceleration for testing purposes.] As of yet, no issues have occurred while in game, although I haven't been gaming as much due to work hours.

My system has never had problems in the three years prior with my previous GPU, which makes me suspect some instability with either the GPU or the current GPU drivers. But I don't have any clear evidence of a physical defect with the GPU either. My question for now is more about confirming if there is a problem with the GPU/Driver, and how to go about proving it.

Any advice?

What model PSU do you have? Are you certain your power supply is sufficient? You should really have a 850watt at minimum for a 3080 TI.

The new GPU could be messing with your memory. You may want to run a memory test like memtest86 to see if it catches errors. If you get errors, drop the speed and see if it changes anything.

 

[gamerk316]

What model PSU do you have? Are you certain your power supply is sufficient? You should really have a 850watt at minimum for a 3080 TI.

The new GPU could be messing with your memory. You may want to run a memory test like memtest86 to see if it catches errors. If you get errors, drop the speed and see if it changes anything.

I have an 850W unit; I think its a Corsair offhand. I note both times lockups have occurred were watching Youtube, so the GPU wasn't running at max clocks. Also ran Warzone with everything maxed @4k like a champ. So I think I can rule out the PSU.

I ran a memory test late last night (1 pass) and it passed fine. Forget the brand offhand, but it's a 2x16GB config running it's XMP profile, so I wouldn't expect any issues there. I plan to run a longer pass sometime tonight.

What's annoying is, so far, I haven't observed any issues while in-game, which is making this hard to nail down. At least then I can just blame the GPU and get an RMA started. Might do that if the problem remains, but this feels like it could be something else that isn't playing nice with the GPU.

One final thing I did: Last night I removed MSI Afterburner. I was using it previously to tweak my 1080TIs fan profile since the fan ramp up by default was at about it's idling temperature, so I had manually adjusted the fan profile so the fan wasn't constantly cycling on/off. I'm hoping MSI Afterburner was just not playing nice, especially since I hadn't updated it for about two years.

Right now, I'm waiting and seeing if the problem occurs again if not, and am planning to go the RMA route if it does. I'm hoping the issue is either the GPU itself (which an RMA will address) or some SW problem that eventually "goes away" with updates.

 

[Third-Eye]

I have an 850W unit; I think its a Corsair offhand. I note both times lockups have occurred were watching Youtube, so the GPU wasn't running at max clocks. Also ran Warzone with everything maxed @4k like a champ. So I think I can rule out the PSU.

I ran a memory test late last night (1 pass) and it passed fine. Forget the brand offhand, but it's a 2x16GB config running it's XMP profile, so I wouldn't expect any issues there. I plan to run a longer pass sometime tonight.

What's annoying is, so far, I haven't observed any issues while in-game, which is making this hard to nail down. At least then I can just blame the GPU and get an RMA started. Might do that if the problem remains, but this feels like it could be something else that isn't playing nice with the GPU.

One final thing I did: Last night I removed MSI Afterburner. I was using it previously to tweak my 1080TIs fan profile since the fan ramp up by default was at about it's idling temperature, so I had manually adjusted the fan profile so the fan wasn't constantly cycling on/off. I'm hoping MSI Afterburner was just not playing nice, especially since I hadn't updated it for about two years.

Right now, I'm waiting and seeing if the problem occurs again if not, and am planning to go the RMA route if it does. I'm hoping the issue is either the GPU itself (which an RMA will address) or some SW problem that eventually "goes away" with updates.

Have you tried older drivers yet? Try drivers a couple months older. It could just be a problem with Windows that might be fixed with sfc /scannow in a terminal. a lot of issues are hard to track down the cause of when games run fine but desktop programs don't. You could possibly have a problem with your OS drive that you might find with a scan.

 

[gamerk316]

Have you tried older drivers yet? Try drivers a couple months older. It could just be a problem with Windows that might be fixed with sfc /scannow in a terminal. a lot of issues are hard to track down the cause of when games run fine but desktop programs don't. You could possibly have a problem with your OS drive that you might find with a scan.

Its possible, though I do stress everything was fine with the 1080Ti, but a SFC certainly wouldn't hurt.

As for drivers, I had initially installed the February drivers before moving to the March 22 ones after the first lockup. It's possible two consecutive NVIDIA drivers have a periodic issue though. I also note the last drivers on the Gigabyte product page are the ones from January; think its worth giving those a shot?

EDIT

Forgot to mention that I plan to reseat/reconnect the card once I'm off work tonight. This doesn't smell like a bad connection, but it's something I'll be doing at some point if the issue remains, so may as well get it done with.

 

[gamerk316]

Updated: Reseated the card; I noted when I examined it felt that the back end was moving a LOT more then it should. Now it's definitely securely in the slot. Hoping that's all the problem was.

I did see one thing odd: It looked like one of the pins might possibly be damaged; the next to last pin (from the back) doesn't seem to go all the way down the board; not sure if this is normal or not.

Photo at: https://www.dropbox.com/s/iaqhn7fexkw7831/unnamed.jpg?dl=0 (won't insert normally for some reason).

 

[kerberos_20]

Updated: Reseated the card; I noted when I examined it felt that the back end was moving a LOT more then it should. Now it's definitely securely in the slot. Hoping that's all the problem was.

I did see one thing odd: It looked like one of the pins might possibly be damaged; the next to last pin (from the back) doesn't seem to go all the way down the board; not sure if this is normal or not.

Photo at: https://www.dropbox.com/s/iaqhn7fexkw7831/unnamed.jpg?dl=0 (won't insert normally for some reason).

thats okay, all GPUs has it trimmed, its just sense pin

 

[Satan-IR]

Updated: Reseated the card; I noted when I examined it felt that the back end was moving a LOT more then it should. Now it's definitely securely in the slot. Hoping that's all the problem was.

I did see one thing odd: It looked like one of the pins might possibly be damaged; the next to last pin (from the back) doesn't seem to go all the way down the board; not sure if this is normal or not.

Photo at: https://www.dropbox.com/s/iaqhn7fexkw7831/unnamed.jpg?dl=0 (won't insert normally for some reason).

Nothing to worry about. They're made like that, as said above.

The PRSNT1# and PRSNT2# pins must be slightly shorter than the rest, to ensure that a card can senses if it is fully inserted in the slot.

 

[gamerk316]

Ok; that's what I figured.

Gonna monitor for reoccurrence for now. Crossing fingers that nothing is physically wrong with the PC and I don't have to go through an RMA.

 

[gamerk316]

Well, made it just over two weeks, and the problem came back. Pretty sure I'm going to start an RMA at this point, but posting one more time in case anyone has additional ideas.

Some thoughts:

1: I am way behind on the BIOS for my motherboard; I typically don't upgrade them without a reason, but it is worth noting I'm still on BIOS version 1101. I'm going to update this after work today, but by itself I'm not going to hold up an RMA over this given the two week monitoring period involved. (Might as well just get a new card during that time in case this doesn't resolve the problem)

2: So far I'm 3/3 on crashes watching Youtube. Kinda makes me suspect some power/performance related issue since the card runs fine on everything else I throw at it.

Anyone got any last-ditch ideas?

 

[Satan-IR]

BIOS update not a bad idea I'd try it anyway.

Have you tried changing Nvidia power settings? As in if it's set to 'Optimal power' or 'Adaptive' change it to 'Prefer maximum power'?

I personally would also do DDU in safe mode and install an earlier nvidia driver.

 

[gamerk316]

BIOS update not a bad idea I'd try it anyway.

Have you tried changing Nvidia power settings? As in if it's set to 'Optimal power' or 'Adaptive' change it to 'Prefer maximum power'?

I personally would also do DDU in safe mode and install an earlier nvidia driver.

The Power Settings could well be the issue...but the card should work in any case.

As for drivers, I tried three successive driver packages (January, February, March); I doubt NVIDIA has had a bug in it's driver for nearly three months.

 

[Satan-IR]

The Power Settings could well be the issue...but the card should work in any case.

Yes card should work and this might somehow shed some light on whether it's a power issue or just card acting weird.

Also it is recommended (and does make sense with cards with higher current spikes) to connect each 8-pin on card directly to a PCIE output from PSU with a separate cable. So for two 8-pins use two cables. If you're using one cable with two PCIE connectors at the end for card it might be worth a try using two cables.

As for drivers, I tried three successive driver packages (January, February, March); I doubt NVIDIA has had a bug in it's driver for nearly three months.

I agree, it's not likely for a bug not to be fixed with 2-3 consequitive updates.

 

[gamerk316]

BIOS update done; also put on maximize performance in the short term. Back to wait and see.