Very low VID problems (Nightmare week)

[13oots2]

My system had been stable with a nice PBO offset, had delided my 9950X and all was working well, am one for hitting Windows Update, so am not sure if it was that. Last week I started experiencing freezes and Windows crashes, and had recently installed a second hand 4tb Lexar NVME, everything I tried just lead back to the freezes and reboots. I noticed today my VIDs were much lower than previously and the CPU sipping power at idle, the VID was dropping to around 0.4V and the cores dropping to 600mHz, the voltages had been higher before normally dropping to around 1.2V, I then had a play around with Power Plan settings and disabled USB and PCIe suspend and set the processor usage to 45%, voltages are still a little lower, but I haven't had a freeze or crash since.

To round it off my cat broke my door and hinge on my Fractal Design Define 7XL and and was shipped an X299 board with nothing covering the CPU socket, I think the consequences of that will be apparent, I did get a refund.

 

[LtMatt]

My system had been stable with a nice PBO offset, had delided my 9950X and all was working well, am one for hitting Windows Update, so am not sure if it was that. Last week I started experiencing freezes and Windows crashes, and had recently installed a second hand 4tb Lexar NVME, everything I tried just lead back to the freezes and reboots. I noticed today my VIDs were much lower than previously and the CPU sipping power at idle, the VID was dropping to around 0.4V and the cores dropping to 600mHz, the voltages had been higher before normally dropping to around 1.2V, I then had a play around with Power Plan settings and disabled USB and PCIe suspend and set the processor usage to 45%, voltages are still a little lower, but I haven't had a freeze or crash since.

To round it off my cat broke my door and hinge on my Fractal Design Define 7XL and and was shipped an X299 board with nothing covering the CPU socket, I think the consequences of that will be apparent, I did get a refund.

Random question, but in the PBO menu in the BIOS, do you have the power limit set to Motherboard or manual?

 

[13oots2]

Discovered the root of the problem.... incompetent operator, PBO and AMD are fairly new to me, I had also done a BIOS reset and some settings from my 9900X must've lingered over despite the BIOS resetting when the new processor went in. I always thought that cores with lower VIDs would yield lower voltages, if I had stopped to think the cores would already be calling for a low voltage. Have been trying to find a baseline where I don't get a freeze when the CPU drops to a lower power state, despite my curves being much more realistic my power draw is lower at idle and maxed out at 255W, my Cinebench scores are pretty much the same. Things haven't been helped by an SAS drive playing up, when I checked my manager there were a lot of "Unexpected Senses", I have re-seated the cables and am pulling off over 5tb of data, to another drive, am hoping that will help too. I am hoping the drive errors were just a loose cable where I had been shifting the PC around a fair bit recently.

 

[13oots2]

Looks like 2 weeks of problems are finally over, the 9950X seemed to be running fine, after taking out my 9900X and put in the new processor the BIOS reset to defaults, I started get the odd random freeze. Had to reset recently due to some PBO and memory tweaks not working out, ever since the system has been behaving very differently, VIDs for CCD0 were very low and the PC was stable as a 2 legged stool. I did some memory testing and would get loads of errors within 5 minutes, my 64gb DDR5 6000 was stable at 5200 but I wasn't happy with that. A bit of browsing suggested a memory setting that I had never heard of, it recommended changing ProcODT to 36.9 Ohms which would stabilise 4 sticks, I now had RAM that would pass an hour of OCCT memory at 6000 and then passed test with PBO on auto. For now I have a -20 Curve on both cores and again OCCT has passed an hour of CPU+RAM. I have dialled +200 boost overdrive and set my PTT, TDC andd EDC limits, everything seems stable and max temps are cooler than before, Cinebench R23 is within error margins of my unstable settings.

I can now get on and finish installing Windows and all my apps, as another bonus my 10tb SAS in RAID 0 is now content with no unexpected sense errors.