Does this look like a hardware problem? Or OC too high?

[Flogger23m]

Using a Zotac RTX 4070 Super. Been having a few issues with crashing initially as I was checking to find a stable OC. Gradually kept dropping the OC until crashing went away. Temperatures and fan speed is very low.

Today I was playing around in VR which worked fine. Though after removing my headset I was greeted with this screen:





Double checked my display port cable to monitor and GPU end as well as power connectors, but all of that appears to be snug and fully inserted.

Does this look like a GPU hardware issue? Or could it be the result of an overclock?

At stock it arrived with +100 on the core and +800 on the memory if I recall. This is what I am running at:



The FPS Review seemed to get a stable OC at +250 core & + 1500 memory as a reference in their review.

TLDR: Should I bother lowering the OC and testing more, or should I just replace it while I am within my return/replacement window?

 

[arbiter]

if your first thought wasn't to hit default to make sure not sure what else to tell you.

 

[oobymach]

These new chips are more sensitive than the last gen, and the gains from overclocking minimal, go back to stock speed on core and rams, set power target to 100%, pray you didn't fuck up your gpu. Just because some idiot oc's for a review doesn't mean you should run that speed.

 

[Flogger23m]

if your first thought wasn't to hit default to make sure not sure what else to tell you.

Interestingly, the default setting is lower than how it arrived from the factory. If you read the post you'd probably have realized that.

These new chips are more sensitive than the last gen, and the gains from overclocking minimal, go back to stock speed on core and rams, set power target to 100%, pray you didn't fuck up your gpu. Just because some idiot oc's for a review doesn't mean you should run that speed.

It did arrive with the power target at 110 upon first loading up any monitoring software. However, the "default" option does set it to 100.

I have been running an ASUS 4070 at 110 as well for around a year and that has no issues. Typically I increase the power limit and put a modest OC on all GPUs I received. Hasn't been an issue with past MSI/EVGA/ASUS cards for me.

Emailed Zotac, will see what they say. This one has been a bit troublesome since day one.

 

[tabascosauz]

Interestingly, the default setting is lower than how it arrived from the factory. If you read the post you'd probably have realized that.



It did arrive with the power target at 110 upon first loading up any monitoring software. However, the "default" option does set it to 100.

I have been running an ASUS 4070 at 110 as well for around a year and that has no issues. Typically I increase the power limit and put a modest OC on all GPUs I received. Hasn't been an issue with past MSI/EVGA/ASUS cards for me.

Emailed Zotac, will see what they say. This one has been a bit troublesome since day one.

From the visual artifacting I'd lean more towards VRAM. If core was the problem I'd expect it to just be unstable, clock stretching, or crash.

I'm gonna be completely honest, I don't see much point in pushing either VRAM or core on my 4070Ti. Undervolting was a better use of time. That includes DCS, especially in VR where CPU choice (X3D) matters a great deal.

 

[Flogger23m]

From the visual artifacting I'd lean more towards VRAM. If core was the problem I'd expect it to just be unstable, clock stretching, or crash.

I'm gonna be completely honest, I don't see much point in pushing either VRAM or core on my 4070Ti. Undervolting was a better use of time. That includes DCS, especially in VR where CPU choice (X3D) matters a great deal.

Swapped the display port cable out, have had zero issues since. I looked at the settings it seemingly came with from the factory, which is +100 core, + 1000MHZ memory and 110 power limit. Now maybe Zotac's Firestorm utility auto OCs something at first launch, but I kind of doubt it would.

Also been using DCS with VR more, and seemingly I am running out of VRAM which results in stutters which would be expected. Steam VR itself uses around 2GB of VRAM.

CPU is a 7800X3D, which should be good for DCS.

Backed down memory clock to +1000, which again is what it came with out of the box. Anything else I can do to check if VRAM is going bad?

 

[tabascosauz]

Also been using DCS with VR more, and seemingly I am running out of VRAM which results in stutters which would be expected. Steam VR itself uses around 2GB of VRAM.

I didn't have any issues with running out of VRAM while on Quest 3, but that was direct to OpenXR. If SteamVR uses 2GB VRAM that sounds pretty egregious.

I know you have 7800X3D; the point was more that pushing GPU more doesn't really change anything for the AD104 cards. At 4K pancake now and the card doesn't really run out of VRAM as much as it's just always a noticeably weaker choice at resolutions beyond 1440p

 

[Flogger23m]

I didn't have any issues with running out of VRAM while on Quest 3, but that was direct to OpenXR. If SteamVR uses 2GB VRAM that sounds pretty egregious.

I know you have 7800X3D; the point was more that pushing GPU more doesn't really change anything for the AD104 cards. At 4K pancake now and the card doesn't really run out of VRAM as much as it's just always a noticeably weaker choice at resolutions beyond 1440p

I will see if I can get my Vive to run without SteamVR. Picked up one second hand for a low price to test VR just a bit. Will look into upgrading to something better in the future.

For performance the higher OC seemed to get a few frame rates in the more demanding games like Cyberpunk, Alan Wake 2 & TLOU. Typically around 3-6 which is an okay increase when you're getting around 70-80 frame rates. I gradually lowered them until the crashes went away. Will be looking out for other issues and will load up those games again to see if anything props up.

Never ran into a VRAM issue at 2560x1440. Will look into OpenXR because SteamVR while it works well, that 2GB VRAM hit is quite large.