Question RTX 3060 GPU clock and memory speed dropping ?

[marko4452]

Hi guys.

I've been having issues with my second hand RTX 3060 that I have bought recently (it was obviously used for mining as 90% of second hand GPUs nowadays). It works fine 99% of the time, but in 1% of cases, randomly, the GPU clock and memory speed drops for around 40 seconds, then comes back up. This happens once every two hours on average, but sometimes it happens more or less often than that. It obviously makes my FPS go down from 90 FPS quite a lot in VR and games become completely unplayable. The temperature seems fine, 67 degrees Celsius stable, so it's not that. I have attached a HWinfo .csv, it's very clearly visible there at around 46 minutes into the log.

https://drive.google.com/file/d/1khH-ZPrZ98z2F-Vpu3NeturyZ0RnfbN5/view?usp=sharing

PC Specs:

CPU: Ryzen 5 2600 stock
GPU: Gainward RTX 3060 Ghost 12GB stock
RAM: 16GB DDR4 2993Mhz
MB: Gigabyte B450M DS3H V1
PSU: Coolermaster MWE Bronze 550W

I have tried setting the power limit to 90% in MSI Afterburner and dropping the maximum clock to 1777mhz (memory clock stayed the same). I am out of ideas, maybe the PSU?

Thanks!

 

[Aeacus]

I've been having issues with my second hand RTX 3060 that I have bought recently (it was obviously used for mining as 90% of second hand GPUs nowadays). It works fine 99% of the time, but in 1% of cases, randomly, the GPU clock and memory speed drops for around 40 seconds, then comes back up. This happens once every two hours on average, but sometimes it happens more or less often than that.

Buying used hardware means that you get it "as is". And since GPU most likely was used as mining card, there's no telling what kind of abuse it saw. And with this (unknown history), it is nigh-impossible to find out and fix what issue it may have. If it even can be fixed.

My guess, it probably has custom mining BIOS (firmware) on it, which, periodically drops the GPU output. Probably for collecting telemetry of the mining operation. But this is just a guess.

You may turn towards Gainward and ask them. But since you don't have any warranty with the GPU, Gainward isn't obligated to help you.

All-in-all, if one doesn't want random issues with their hardware, they wouldn't buy used stuff, but instead brand new. But with used stuff, it having issues, is normal. This is compensated by the cheap price you payed for it.

My suggestion;

• Either live with the quirk your GPU has, while comforting yourself with the idea that you bought it cheap,
• Or buy brand new GPU, which doesn't have this issue, while paying full price of the GPU, as it is worth it. + You'd get warranty too with brand new GPU, whereby when issue(s) do arise, you have safety net in form of valid warranty.

Coolermaster MWE Bronze 550W

Poor PSU and a bit weak for your system as well.

With RTX 3060, i'd be using good quality 650W unit. E.g Seasonic Focus/PRIME or Corsair RM/RMx/RMi/HX/HXi/AX/AXi series.

If your PSU would be an issue, you'd see random system reboots/shut downs instead. Still, i'd replace the PSU nevertheless.

 

[marko4452]

The GPU did have a mining BIOS on it, I reflashed it the moment I got it. Strangely, the GPU passed two 1hr 40min Fire Strike stress test without any issues. It seems like the issue only happens in iRacing which I play in VR. Not sure what to do to be honest, the last thing I can try is limiting its power use in MSI Afterburner, but that's about it.

 

[Aeacus]

It seems like the issue only happens in iRacing which I play in VR.

If issue is only with 1 game, then it's the game's fault. Some optimization issues perhaps. Best to look into the game official forums and see if there is some fix for it. Or contact developer via e-mail and ask what can be done.