Can't find nonce with device restart in windows

[hahalo]

3060ti
0220505 12:40:26 TREX: Can't find nonce with device [ID=2, GPU #2], GPU doesn't appear to be LHR. If it is LHR, try updating video driver.: error code 999
20220505 12:40:26 WARN: Miner is going to shutdown...
20220505 12:40:26 Main loop finished. Cleaning up resources...
20220505 12:40:26 ApiServer: stopped listening on 127.0.0.1:4067

[tjayz]

"TREX: Can't find nonce with device [ID=2, GPU 2], GPU doesn't appear to be LHR. If it is LHR, try updating video driver.: error code 999"

I can confirm this has happened to me twice in ~2 hours since upgrading to both studio driver 512.59 and trex 25.15, on two seperate EVGA 3060 Ti ftw3 ultras in a rig of 10. Results in BSOD for me in Windows 10 x64.

[bugsybando]

3060ti hynix latest hiveOS and Latest drivers dealing with this too.

[IMMGO]

same problm with me ,3060ti hynix using in windows, studio driver 512.59 and try to change other 51x.xx drives.

[tjayz]

This happened two times more on my rig. Affects 3 out of 10 3060ti, both samsung and hynix.

Decreasing mclk 25 (windows) and -lhr-tunes .1 on affected gpus to see if resolves. Suspect the driver upgrade to 512.59 combined with lhr on all gpus now at 80.0 to 80.2, causing instability. Those new lhr points were found utilizing new settings --lhr-autotune-interval 5:20 --lhr-autotune-mode full. After lhr high points found, reverted full to down. Will report back results in day or two.