Games freeze (or crash) once a while once the GPU is stressed almost at max after upgrading PC

Dec 3, 2018
5
0
10
0
Hello community!

Help a fellow enthusiast! :)

I'm currently experiencing frezes in pretty much all GPU-intensive games once the utilization is hitting max (99-100%). Although it seems like a freezing issue, games actually crash because sometimes I can see there's a second window message that pops-up after the freeze but I can't actually see it since everything on my first monitor is not accessible. I'm using dual monitor setup and I'm force-closing the game once this happens from the second monitor.

This is my actual setup:

- Intel i5-8500 (stock settings, the joke)
- Gigabyte Aorus 1080ti (stock settings)
- 2xHyperX Fury DDR4 16 GB (model here: https://www.kingston.com/datasheets/HX421C14FB_16.pdf), using XMP profile
- Gigabyte Aorus B360 Gaming 3 Wi-Fi (with latest BIOS update)
- Seasonic Prime Ultra 1000W Titanium
- running windows 10 64bit (legally licensed if that matters), nvidia drivers are the latest (417.22), also windows is stored on NVMe (Intel 760p)

I usually run games at max settings 1440p at 144hz (I am running dual Predator XB271HU, one TN and one IPS, I game on the TN panel though because reasons). I should say that the issue started occurring once I've upgraded the GPU and PSU altogether. Previously I had a GTX 1060 3GB card and used an eVGA 550W G2 PSU, on which I didn't had any issues whatsoever.

Once I swapped the GPU I did a clean-install of the nVidia drivers, as for swapping the PSU I did changed all the previously used cables with new ones as I've heard there may be compatibility issues. Although I've been building PCs before and installed windowses all my life, I'm no expert at debugging and any opinion or direction is greatly appreciated.

To summarize, or TLDR, the issue is occurring randomly, every 5-10-15 minutes of gaming at max settings once I've upgraded my GPU and PSU altogether. There is some coil whine, or at least I think it is, once the GPU is getting stressed, so that is a clue for sure. Only the application itself freezes/crashes, in this case the game and not the operating system, and I've never had a BSOD (blue screen) so far. Another clue might be that I regularly play a browser game in-between loading maps (World of Tanks) and once the issue occurs I do notice some artifacts in the browser game too.

I cannot say for certain that the GPU is to blame since the PSU might be affecting it.

What I did so far:

- did a clean installation of nVidia drivers once I swapped the GPU
- removed the old PSU cables with new ones before installing the new PSU
- stress-tested the GPU using SuperPosition benchmark, success only

What I didn't do so far:

- pinpointed the coil whine, could be the GPU or PSU
- reinstalled Windows
- tested the GPU in other PC
- tested the PSU in other PC

Do you guys have any recommendation as to how to start debugging this issue?

Thank you in advance!
 
Dec 3, 2018
5
0
10
0


Hey man, thanks for the reply. I ran the OCCT PSU test for more than 10 minutes, everything seemed fine. The PC was venting some heat I could hear the fans spinning a bit louder but other than that, everything was fine.

CPU cores sat around 55 degrees and GPU temperature hit a max 72 and was stable around 70-71. I think the power consumption hit a max of 230W.

Regarding the GPUz render test, I couldn't find the option to start it. Is it still present in the GPUz app?
 
Dec 3, 2018
5
0
10
0


I finished doing this test too, GPU sat around 60 degrees, I left it for about 10-15 minutes.
 
Dec 3, 2018
5
0
10
0


Hey man, sorry for the late reply, I was caught up with work lately and haven't even turned on the PC. I now have more time on my hands. To answer your question, I haven't done any overclock in my system, besides the XMP profile I've activated on my RAM.

I do hower have something, I've been gaming StarCraft 2 with GPU-Z turned on and on the sensor tab. This game doesn't crash but freezes from time to time, for like 5 seconds. Each freeze in the GPU-Z sensor tab is accompanied by a complete pause in rendering or load, every time this happens the GPU is at 75 celsius degrees or above, max I've seen was 81-82.

I think this is what they call a thermal throttling. I've read about it and it is related to the workload the GPU is exposed, the type of cooling the card has and the overall case ventilation. The workload is obviously to the max, I run every settings on ultra, on a 1440p resolution at 144hz, plus G-Sync on. The card is kind of famous for it's triple fan cooling system, so it's shouldn't have a problem here.

The case ventilation however is debatable, I have two 120mm fans for cooling the CPU radiator, which is AIO; configuration is push/pull and it is mounted up rear-back. I have another two 120mm fans for intake, mounted in the upper part of the case. And I have two 120mm fans for exhaust mounted in the front. It's not exactly by the book but somehow the airflow it's balanced, 3 intake vs 2 exhaust.

Question is why would this 1080ti card thermal throttle at 75-80 degrees celsius and also how can I improve things?

Much appreciated!
 
Dec 3, 2018
5
0
10
0


I did what you said and honestly I didn't think it would mean that much, but it does! I set my front fans to intake and the top mounted fans, along with the CPU AIO to exhaust and now my GPU stays at 64 degrees celsius. Granted I haven't stress tested it but it already looks like an improvement, wow.

I guess I should have thought about hot air trying to rise and not try to force its exit in the front of the case.

Many thanks man! Let me know if you have a patreon account or if you accept donations as I would like give something back to you :)
 

Similar threads


ASK THE COMMUNITY