Quote from: mrbillishere on December 18, 2021, 09:10:52 PM

Quote from: miner29 on December 18, 2021, 06:31:03 PM

Current is 497.09 for windows. This is most likely the issue.

@miner29 - if your reply was directed towards my post, just above yours, then first off thank you for weighing in... I should add that per the nVidia control panel we're still using driver package 470.05, and per device-manager's info the timestamp on the driver file is still 2021-02-26, so I believe I've ruled out Microsoft being helpful with a GPU driver update. (Also, I looked at update history and aside from the monthly cumulative update, nothing else was applied yesterday. bummer.)

That said, I'm not averse to cleaning up the existing nVidia drivers package using DDU followed by a reinstallation of the same to see if anything changes.

In the meantime I did get t-rex going to see what it thinks. Both cards are hashing and t-rex was able to set the memory and clock speed deltas, and the clock power delta. Of course nothing's perfect - it is claiming it can't see the temps of the graphics memory so I can't specify a fanspeed range based on mem-temps. IDK if that ever worked because I've never used t-rex for ethash before and heck, I'm not sure I want to switch because t-rex doesn't support AMD GPUs like PM does, which means I'd have to run PM for AMD and t-rex for nVidia on this mixed GPU rig. Yaay.

Here's a quick update re: a workaround I discovered that remediates the bizarre errors I was getting from PM...

So to recap, for the past five days the rig was running PM v5.9d for the AMD RX580 GPUs and t-rex for the nVidia GPUs because PM glitched on me after this month's Win10 patches. That was alright, but let's just say I am not a fan of the t-rex console, nor of its higher dev-fees, and I missed some of PM's features like being able to hit C while mining when you want the app to re-read its config settings and update them on the fly, so I didn't leave well enough alone.

This evening, I went into the UEFI of the motherboard (it's an ASUS brand with an AMD x570 chipset) and fiddled with the settings for the PCIe version assigned to each PCIe slot. Previously, all four were set to Auto and the newer nVidia RTX 3000-series GPUs were being set to PCIe 4 automatically while the RTX 580s were being set to PCIe 3. Pretty normal stuff. Don't ask me why but I decided to set all four expansion slots to PCIe 3 to see what would happen. Lo and behold, PhoenixMiner is back to normal! You could even go so far as to say that it's risen from the ashes. (cue rimshot)

So, take this fwiw. Maybe it'll help someone else too. Cheers.

mrbillishere

Posts

Merits