Build freezing AFTER GPU load - Help needed

Thomas Lee · 9 Dec 2020 at 20:17

Benass said:
Do u have a good friend who has a pc?
u
ccacan
mmamakmake
yyoyouyour
llilifliflife
eeaeaeaseasieaeasier omg this forum bugged on my phone.Daamn

.
can u swap for another g card or other ram. ?

I do but don’t think I’d be allowed to test with a 3080FE given the issues I’m having

I have different RAM already that was changed out a week ago to try and tackle this (so don’t believe it’s a RAM issue but still doing all the suggested testing)

Vince · 9 Dec 2020 at 20:20

Thomas Lee said:
Here you go. Thanks in advance for any thoughts - this is well over my head now!

You are running memory at xmp 3600mhz, stock would be without xmp at likely 2133/2666. Put it stock and test from there

Thomas Lee · 9 Dec 2020 at 20:43

Vince said:
You are running memory at xmp 3600mhz, stock would be without xmp at likely 2133/2666. Put it stock and test from there

Great suggestion, but already tried - see post #9 where I had diminishing performance and stability with stock MHz options.

Thomas Lee · 9 Dec 2020 at 21:03

mickyflinn said:
Sorry skipped the memtest bit.

Memtest 4hours in shows no errors (https://imgur.com/HiaXShx). Will leave running overnight. Thanks

Webzta · 10 Dec 2020 at 08:40

Thomas Lee said:
Great suggestion, but already tried - see post #9 where I had diminishing performance and stability with stock MHz options.

If you're seeing errors at stock memory setup, on two different sets of ram, it's possible there's an issue with the CPU/Board.

Thomas Lee · 10 Dec 2020 at 10:32

Thomas Lee said:
Memtest 4hours in shows no errors (https://imgur.com/HiaXShx). Will leave running overnight. Thanks

No errors identified.

Thomas Lee · 10 Dec 2020 at 10:37

Webzta said:
If you're seeing errors at stock memory setup, on two different sets of ram, it's possible there's an issue with the CPU/Board.

I think part of the issue is I'm not seeing errors at all for memory. When the GPU isn't being utlilised with normal use, it's absolutely fine at all of the speeds tested. The crashes are only consistently in game. I'm going to order a new board today however reluctant to swap out the old until I know this is the issue. Unless consensus is to do this anyway?
To mention; I have fully removed Afterburner and associated files/profiles though this has not had any affect either.

adrichardson81 · 10 Dec 2020 at 13:23

Try running hwinfo in sensor mode and a 20 pass test with tm5. Hwinfo should show any WHEA errors.

Thomas Lee · 10 Dec 2020 at 13:52

adrichardson81 said:
Try running hwinfo in sensor mode and a 20 pass test with tm5. Hwinfo should show any WHEA errors.

Couldn't figure out how to adjust for 20 passes, though nothing picked up in the test run https://imgur.com/i3dim5j . Thanks

adrichardson81 · 11 Dec 2020 at 14:18

Have you tried underclocking your gpu?

Thomas Lee · 11 Dec 2020 at 17:45

Underclocked by 250Mhz - No change to stability unfortunately.

I’ve used it for the whole day today for work (Citrix) and a few apps open in the background and been absolutely fine. 2minutes into game play and it crashes - could it be an issue with the game download (should I offload and download a fresh version)?

Thomas Lee · 12 Dec 2020 at 16:00

Decided to physically pull it apart this morning and make sure there were no obvious issues with installing of the hardware. Ran Heaven for around an hour on Ultra/AA x4 with no issues. Ran PerformanceTest(3D mark) and it got through all except the particle screen (https://imgur.com/Ar0At3D) before crashing which I think is after DirectX12 test(?) . Any further thoughts?

Have ordered a new board and waiting for stock of a new GPU (whcih I get the feeling is the issue?). Thanks

Other screenshots from GPUz, HWMonitor, TM below:
Running Heaven - https://imgur.com/YMDCXlR
Crash - https://imgur.com/B1HNSO7 | https://imgur.com/geAgM0K | https://imgur.com/NgUmCvy

Thomas Lee · 20 Dec 2020 at 22:16

Wanted to provide an update on my issue. I noticed GPU sag when putting the rig back together and, having propped it up temporarily to ensure it was level, the issue seemed to completely disappear (extensive stress testing and game play - not a single reoccurence since 12 DEC) all settings I had in place are now all higher as a result seeing the 2070s can perform well.
I had been using the arm mount which came with the GPU, however wasn't secure enough it seems. Have since bought a GPU Stand and everything is ship shape and shiny. I would assume this means there is a lose connection on the board or connection on the GPU PCIE (?).
Plan to switch out the board this week once my new GPU arrives and can put this all to bed. My 2070s will be going to a family member and be giving them the stand to make sure they dont experience the same issue as I did.
Thanks all for your input, suggestions and helping educating me along the way!