Recent content by PotatoVonEpicus

I just gonna assume it's not really worth worrying about hotspot temperature with Vega. It's supposed to shutdown PC or throttle at 120c or so anyway. Thanks for information.

I see. But if sensor is reporting inaccurate and high hotspot temperatures core should throttle down and drop clock. It doesn't do that and that is the weird part for me. It's reporting high hotspot temp without taking any action to drop it down. Also I assume it's average of multiple sensors...

Yesterday I played some games on 1440p with my undervolted Vega 64 LC ( 1650 MHz 1.1v core, 285w ). Highest core temperature reported by HWinfo was 71c and average was mostly around 65c. Hotspot temperature however was reported as 110c, which I doubt was accurate because I didn't see that...

Yes, that fixed my issue. Shouldn't Hwinfo only read settings and not overwrite if nothing is changed though? Probably shouldn't overwrite if nothing is changed, unless it doesn't check if anything is changed and overwrites anyway. Regardless, thank you for your time.

RtssInclude"=dword:00000001 is indeed present in backup.reg file. Also I did not have any issue before with this same Windows build that I am running now. Could it be caused by running Hwinfo before running registry file to add setting values? Gonna attach settings file.

I bounced back and forth using different Windows builds and had no issues with Hwinfo. Been using Win11 for few months, also without having any issues with Hwinfo. However when I decided to go back to Win10 due to unrelated issues it doesn't seem that backup.reg file works anymore. I have...

Looks good, did it resolve the issues you were having with your computer?

Are these 2 x 16 GB sticks are dual rank? Dual rank puts more stress on memory controller. L3 cache errors also indicate memory controller. Things you can try. - Slightly increasing memory voltage - Slightly increasing SOC voltage - Loosening RAM timings I suggest you try these voltages SOC...

Offset +, you add as much as you need to already set voltage. I strongly suggest you google this and better yet google specific motherboard that you use. It might explain everything better. I read all messages in this thread and it seems to me that this is either degradation or it's bad...

Well, one more area where these Vegas are buggy to oblivion. At least it reports values fine after disabling ULPS, which re-enables itself. However after disabling it, values remain correct. Also GPU itself should be fine, I did some rendering on it the other day. I am also wondering if this...

You can try increasing SOC voltage to 1.1v. If it's CPU issue, otherwise I would replace those RAM sticks.

I have one more debugging - file - with the issue occuring. Assuming vbios flash reset ULPS. Seemingly things are more stable with Vega 56 vbios. UE4 kept on crashing my PC, but that can also be caused by HBCC segment. Regardless, seems fine now.

I haven't, however whea errors usually are caused by memory controller. On both AMD and Intel. Also it may not be running in 1:1:1 ratio due to low RAM clock, memory controller can get unstable with weird voltage or even specific clock. Although I never tested this it could probably be getting...

In BIOS. I suggest you google the model of your motherboard and look up BIOS manual.

I think having ULPS enabled shuts one GPU completely off. I have seen second card turning red leds off and just doing nothing completely. Might as well report all values as 0 if it's disabled. I haven't seen it do that with this windows install or with this specific GPU driver. Thank you for...