Help troubleshooting 3D operation on my main rig

[Voodoo Rufus]

Donnager is having some problems. Desktop use is fine. A couple months ago I fired up ARK and Expanse Telltale Games and played a while. Worked fine. A month later, fire up ARK again and the machine black screens and will not reboot without hitting the button. Stress test on other 3D things yields random black screen failure as well. I loaded ARK on my backup rig and it works fine, so it's not a game issue. Temperatures are not a problem, nor are drivers.

All I can think of is to work my way from easiest to hardest issues to eliminate: PCIE riser cable, PCIE power cables, PSU, and then the GPU itself. Open to other ideas.

 

[EarthDog]

I doubt it's the psu cables... pcie riser card is first on my list.

What psu?

 

[Voodoo Rufus]

It's a 5 year old Seasonic Titanium 1000W. I've lowered the PCIE speed to 3.0 from 4.0 to see if riser comm issues would be a thing and it made no difference. It would be odd for it to fail after testing good prior. But I don't know for sure.

 

[Voodoo Rufus]

Well, I removed the riser, and no joy. These are the artifacts I'm getting, and still getting black screen reboots at random. Underclocking did not help either.

RMA time?

 

[freeagent]

RMA time?

I would.

 

[Voodoo Rufus]

Took out the card and swapped in an air-cooled 1080Ti. The artifacting was still there, but I found out that it was a Nvidia driver setting issue. It still doesn't address the crashing I had with the 3080Ti, but the 1080Ti has been running a couple hours with no issues. Still tempted to go through with the RMA because of the crashing.


Is EVGA still swapping out these cards with new old stock, or what are they doing now?

 

[EarthDog]

Is EVGA still swapping out these cards with new old stock, or what are they doing now?

Last I saw, things are still being replaced as they obligated originally.

 

[Voodoo Rufus]

As long as it's the same PCB, then I can continue to use my Heatkiller block. If not, well.....I'd rather go to a newer gen card at this point.

 

[Voodoo Rufus]

I keep searching and black screens seem to be very common the past few years and it's not specific to any one gpu, but the 30 series may have a higher rate from power spikes.

The one thing I haven't tried is changing the gpu power cables, but I have reseated them. I don't have extra Seasonic cables, so I'll use one of my other rigs with an Evga or the new super flower one and see if that does anything. I really don't want to RMA this thing if I don't have to

 

[Voodoo Rufus]

Made more progress yesterday. Got out my Super Flower 1000W / Z390 rig to test the GPU in. Z390 wouldn't turn on, only flashing LED's every five seconds. Took it out of the case, no joy. Laaaaame. Got out my Z97 / 5775C Win7 system, and it fires right up. That Z390 might be bricked. Tried both bios switch settings and a different battery, too.

Anyway, the GPU refused to crash after hours of Heaven demo, even overclocked, so it would appear the Seasonic PSU in my main rig can't cut it. My Seasonic (Prime Ultra Titanium 1000W, 2018 model) was supposedly beefed for power spikes, so unless it's a PCIE power cable I don't know why it would be barfing. So, now I'm going to commit to a full PSU swap on the main rig from the Seasonic to the Super Flower. Going to test it on the counter first with the cobbled cooling system on the GPU before doing the full PSU change out and GPU integration back into the case's cooling system.

Orrrrrr, I could clean and sell the 3080Ti while I have it out of everything and upgrade while doing the PSU swap, and upgrade the GPU to a 40 series. Hehehe.

 

[EarthDog]

Orrrrrr, I could clean and sell the 3080Ti while I have it out of everything and upgrade while doing the PSU swap, and upgrade the GPU to a 40 series. Hehehe.

Yaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaas!

 

[Voodoo Rufus]

Don't tempt me!

The prospect of ridding myself of the 30-series power spikes and power consumption (who am I kidding, I DGAF about power draw) may be a positive.