PC restarts under heavy GPU load

[Yoshiwizard3]

So i recently upgraded my PC early January, and was one of the lucky few to get a brand new NVIDIA 3080 founders edition for a reasonable price, and everything has run perfectly fine up until the last week.

 

Suddenly around a week ago my PC just crashed and restarted in the middle of playing a game, the RGB flickered on and off the card and i heard the fans spin up and it would die again. It would fail to boot around 5 times before coming back up again and would send me straight back to windows with no error messages or anything.

 

From what i read around this sounded like it would be either high thermals or a bad power supply, but i assumed my power supply wasn't bad as i got a brand new one in January along with my new card to support it, but my GPU temps on afterburner were also consistently below 70 C when under load and around 30-50 idle.

 

I did read around that the founders edition cards were having issues with overheating VRAM chips, which weren't detected by the main sensor you would find on afterburner, so i checked this using HWiNFO64 and ran a heaven benchmark to just see what the temps would be. Sure enough i found that the memory was instantly spiking to 110+ degrees and throttling the card so i though that must be the issue with it. I went and bought some thermal pads and a couple aluminium heat sinks and fans to see if i could reduce these temps with some minor mods to the back of the card. After doing this i managed to get my memory temps all the way down to under 80 C and the card ran fine for around half an hour in heaven benchmark, however it crashed again, although this time it managed to boot straight away with no trouble.

 

Obviously this isn't ideal as it still won't allow me to play any sort of demanding game for a decent length of time, the only solution I've managed to find is under clocking the card in afterburner and turning the power limit down to 31%, this as far as I've found fixes the problem but isn't ideal having a high end graphics card to push it at a third of its capability. I'm just assuming there's something I've overlooked somewhere causing the issue however its worked fine for months with no complaints at all so its not like a sudden change in hardware has altered anything.

 

If anyone has any clue of a fix i'd love the help :)

 

PC specs are:

ASUS PRIME B250-PRO

Intel Core I5-7400 (was planning to upgrade soon)

NVIDIA GeForce RTX 3080 Founders Edition

GIGABYTE GTX 1660 Super

20 Gb ADATA memory

KIngston M.2 NVMe SSD 500Gb

1Tb HDD

Windows 10 64 bit

EVGA bronze 750BQ power supply

 

 

[jslane21]

could be the gpu, id honestly try a beefier psu first worst case send it back if its not the problem.

[Yoshiwizard3]

On 4/23/2021 at 12:47 AM, jslane21 said:

could be the gpu, id honestly try a beefier psu first worst case send it back if its not the problem.

yeah that might have to be my next shot, i'm debating whether its a CPU issue because it always caps out at 100% while gaming and can sometimes cause instability if i have and apps open on my second screen, not sure if that would affect anything but it doesn't seem to match up, just really don't want it to be the GPU.

[Yoshiwizard3]

Replaced the bronze 750 watt power supply with a ROG Strix gold 1000 watt supply, seemed to stop restart loops straight away after switching them, ran heaven benchmark for over an hour including a further half an hour with overclocked settings and ran perfectly compared to before crashing after about 15 mins on standard clock speeds.

 

Clearly the 3080 just needs that extra head room over the recommended power consumption.