PDA

View Full Version : GPU unstable, Bad Motherboard?



Hopeywolf
12-05-2017, 01:33 AM
So I recently had to RMA my GPU and today i got it back and its worse in terms of stability. I tried everything:

-Fresh install
-Clean Drivers (DDU)
-Swapped PSU
-Old system ( i7 4970k)
-Different PCI slot within new build.
-Memtest86+ for 10 hours
-AIda64 CPU stress with memory for 7 hours
-Realbench for 16 hours Stress test
-Prime95 for 10 hours

The only item that showed no issue was using my old system for my GPU gave no issues. So is the motherboard defective?

Raja@ASUS
12-05-2017, 01:55 AM
You'll want to detail the following:

1) Full list of parts.
2) Configuration - overclocked or not (yes, that includes XMP).
3) The type of instability - BSOD? System locking up? Something else?

Hopeywolf
12-05-2017, 02:22 AM
You'll want to detail the following:

1) Full list of parts.
2) Configuration - overclocked or not (yes, that includes XMP).
3) The type of instability - BSOD? System locking up? Something else?

1.
Processor: Intel® Core™ i7-8700K CPU @ 4.30GHz
Memory: Corsair LPX Series 32GB @ 3000Mhz
Hard Drive: 2x Samsung Evo 850 1TB, Samsung 960 Pro 2TB NVME
Video Card: NVIDIA GeForce GTX 1080 Ti FTW3
Monitor: ROG PG278QR + XB271HU
Operating System: Windows 10 64-bit
Motherboard: Asus ROG MAXIMUS X HERO (WI-FI AC)
Computer Case: Corsair 570x
PCI Addons: Soundblaster Z + Elgato Capture Card HD60 Pro

2. Tried XMP and non XMP, same issues. No overclocks past that.

3.
69393
69394
69395

Usually after this happens system recovers from drivers but more often, it just freezes and restarts.

Other than that, my system is stable while stressing the CPU and memory alone.
My system stutters for 2 seconds at the start but that could be normal.

Raja@ASUS
12-05-2017, 02:36 AM
That looks like a faulty GPU, or one that's overclocked too far. Of course, I'm assuming the card is properly seated in the slot, the display cable fully inserted on both sides, and that there are no misaligned pins in the CPU socket.

Hopeywolf
12-05-2017, 02:39 AM
That looks like a faulty GPU, or one that's overclocked too far. I'm assuming the card is properly seated in the slot, the display cable fully inserted on both sides, and that there are no misaligned pins in the CPU socket.


I double checked before installing my cpu and there are no misalign pins. This doesn thappen instantly its over 5-10 minutes. My last card before RMA was able to do heaven benchmark without any of the artifacting. I would be heartbroken if it was the new asus board I just got, im close to ordering a new board just to see if its the issue, but i dont see why it would be. I used a 1050 ti on the slot and its fine

Raja@ASUS
12-05-2017, 02:40 AM
I'd contact the graphics card vendor.

Hopeywolf
12-05-2017, 02:46 AM
I'd contact the graphics card vendor.

What should I tell them, They said they honestly think its the board since the GPU worked on an older x99 sabertooth on the same stress test, but the errors look like they come from the GPU, its confusing

Raja@ASUS
12-05-2017, 05:45 AM
This type of stuff may need a certain frame rate/load/temp to manifest, so I'm not sure the comparison to other platforms is valid.

Hopeywolf
12-05-2017, 06:20 AM
This type of stuff may need a certain frame rate/load/temp to manifest, so I'm not sure the comparison to other platforms is valid.


On top of that my old card that was RMA'd had gsync flickering constantly but with it off, it ran games and benched heaven just fine. This card I got as a replacement is a lot more unstable in comparison.

Raja@ASUS
12-05-2017, 07:15 AM
As I said, I'd take this up with the graphics card vendor.

Hopeywolf
12-05-2017, 07:16 AM
As I said, I'd take this up with the graphics card vendor.

Yes i will thanks.
so I forgot this info a while ago but Nvidia has a debug mode which disables any factory overclock from the vendor and gives it a stock setting. So i turned that on and it seems to not artifact anymore.

Raja@ASUS
12-05-2017, 07:21 AM
Yeah, unstable card at the applied oc.

Hopeywolf
12-09-2017, 05:12 PM
So I got the new card today and out of sheer luck, its a defective card. I was told that it was a new card that came in a brown evga box.
It doesnt run as it should, its 50% slower than the bad card and the voltages are not properly raising to increase the performance. Any changes to the fan profiles on EVGA precision will cause the computer to stutter heavily. Keep in mind the old card in debug mode works fine, it had a bad OC otherwise. I swapped the NEW card for the old one and it runs as it should in games and benchmarks.

I ran 3dmark and here are the results. This is the bad card on DEBUG MODE
https://www.3dmark.com/3dm/23850205?

This is the new card i got
https://www.3dmark.com/spy/2897251

It doesnt make sense to me. I doubt its my motherboard and CPU because if it was, then the results would have been the same.

I reinstalled windows and did DDU for both cards

People are telling me that the maximux hero X motherboards are prone to these issue with the GPU? which makes no sense to me at all since the bad OC card works fine with Debug mode on that removes any factory OC

Menthol
12-09-2017, 05:51 PM
I have no issues on my X Hero Wifi, currently using 1080 FTW, also have had an ASUS 1080 Strix, no issues

Hopeywolf
12-10-2017, 01:37 AM
just constant bad cards with evga then

Menthol
12-10-2017, 02:17 AM
That would be my guess, just keep on them tell your satisfied
One question about your power supply, is the 12 v a single rail, if multi rail make sure your GPU is on it's own rail with a separate cable to each plug on the card

Hopeywolf
12-10-2017, 02:20 AM
i have a 1000 T2 from EVGA as well as a 1000G from the same company. I dont think its the PSU though