Question Can undervolting your CPU cause Windows 10 instability?

[yaggaz]

So I decided to undervolt my i9-14900k by -0.050v in the bios settings. Also my XMP profile was disabled so I could increase my RAM from 4800mt to 6000mt

Now I have gamed for days. I have ran multiple CPU and system stress tests across various programs. I let unigine heaven run overnight on extreme..

Not a single problem. Not a single stutter, crash, BSOD and my CPU hovers in the 33c at idle to 65c at gaming, GPU at 40c idle to 55c gaming. My CPU actually dropped 5 degress after the undervolt.

But here's the thing. While sitting idle in the Windows 10 environment doing barely nothing, the system locks up forcing me to reset.

Why no problems under extreme stress with gaming, but problems when doing nothing at the Desktop?

My setup at time of this post:

|| CPU: Intel 14900k || GPU: evga 3080 XC3 Ultra Hybrid || MB: Gigabyte z490 UD AC || RAM: 2 x 32GB 6000mhz DDR5 SDRAM || Samsung EVO 970 Plus 2TB || Dell S2417DG Monitor || Soundblaster AE-7 || Phanteks Eclipse p600s Case || be Quiet! Dark Rock Slim CPU Cooler || Corsair AX1600i PSU || 7 Fans total in system ||

 

[bigpinkdragon286]

"Can undervolting your CPU cause Windows 10 instability?" Yes.

There is no guarantee a chip will run stably at settings the manufacturer hasn't sold them to work at regardless what one may read or watch online. Integrated circuits, while repeatably reliable, still have ranges in quality and tolerance.

Intel has even run afoul of the reality that chips run without end-user scrutiny of default voltage settings set by motherboard manufacturers, may degrade in short order.

It's simple to test, really. If your system was stable before you tinkered with the settings, undo one of them. Start by undoing your undervolt. If your system stops behaving in an unstable manner when idling at the Windows Desktop, that's the most likely candidate for the unwanted behavior.

When undervolting, pick a target. If instability is witnessed afterward, dial back your target undervolt. The goal is to walk the setting from unstable to stable until you find what your particular chip will tolerate.

Why would your chip have issues at idle instead of under load? Chips often use the equivalent of voltage tables. Different amounts of work being done by the chip will require different voltages to remain stable. Does your undervolt apply to just a single entry in the table or is it an offset for the entire range the chip will operate at? It's possible your chip does not need the maximum voltage given at the top end of the table, but when workload reduces and allows for idle voltages, your offset now falls below the minimum threshold needed for your particular chip.

 

[yaggaz]

Thanks for the

"Can undervolting your CPU cause Windows 10 instability?" Yes.

There is no guarantee a chip will run stably at settings the manufacturer hasn't sold them to work at regardless what one may read or watch online. Integrated circuits, while repeatably reliable, still have ranges in quality and tolerance.

Intel has even run afoul of the reality that chips run without end-user scrutiny of default voltage settings set by motherboard manufacturers, may degrade in short order.

It's simple to test, really. If your system was stable before you tinkered with the settings, undo one of them. Start by undoing your undervolt. If your system stops behaving in an unstable manner when idling at the Windows Desktop, that's the most likely candidate for the unwanted behavior.

When undervolting, pick a target. If instability is witnessed afterward, dial back your target undervolt. The goal is to walk the setting from unstable to stable until you find what your particular chip will tolerate.

Why would your chip have issues at idle instead of under load? Chips often use the equivalent of voltage tables. Different amounts of work being done by the chip will require different voltages to remain stable. Does your undervolt apply to just a single entry in the table or is it an offset for the entire range the chip will operate at? It's possible your chip does not need the maximum voltage given at the top end of the table, but when workload reduces and allows for idle voltages, your offset now falls below the minimum threshold needed for your particular chip.

Thanks for the explanation!!!

So I had figured what might have been happening is a fault with the USB ports on wake up. I moved my mouse and keyboard to the 2.1 ports and I havn't had a single lockup since.

I also, for the first time ever, got lazy and just swapped out the HDD into the new motherboard without freshly reinstalling windows. It works great, but I'm wondering if there is a leftover driver conflict with the USB hub or maybe the MB is damaged on those ports?

 

[USAFRet]

I also, for the first time ever, got lazy and just swapped out the HDD into the new motherboard without freshly reinstalling windows. It works great, but I'm wondering if there is a leftover driver conflict with the USB hub or maybe the MB is damaged on those ports?

Gremlins may come up to bite you at any time.

 

[Karadjgne]

Thanks for the


Thanks for the explanation!!!

So I had figured what might have been happening is a fault with the USB ports on wake up. I moved my mouse and keyboard to the 2.1 ports and I havn't had a single lockup since.

I also, for the first time ever, got lazy and just swapped out the HDD into the new motherboard without freshly reinstalling windows. It works great, but I'm wondering if there is a leftover driver conflict with the USB hub or maybe the MB is damaged on those ports?

Drivers by themselves are actually very rarely ever the issue nowadays, with more modern versions of Windows. Because Windows itself generally contains multiple drivers from multiple different brands, as well as it's own generic drivers, for just about every standard component there is.

That said, what's generally the issue is the registry and directions from such, it's (for instance) telling Windows to load asus mobo drivers, you have an msi mobo, so Windows cannot apply the drivers necessary, uses generic Windows drivers instead just to get a functional platform, and then you don't get the performance or sometimes compatibility necessary if msi is using something more proprietary. Gremlins.

Hypothetical, genetic BT or usb drivers are ancient, you are talking about version 1.0 etc, which work just fine with BT or usb 3.0, but you'll definitely not see power delivery or speeds close to what new versions can deliver, and if the added equipment is looking for minimum usb 2.0, you'll get glitches, inability to fast charge etc.

Many ppl don't actually see the gremlins, they don't show as critical system errors and bluescreens, they'll pop up as yellow errors in the log, get bypassed or worked around by Windows. Little stuff like having a 144Hz monitor that shows up as a generic monitor and only puts out 120Hz, most only see that if delving into system specs and making absolutely sure they are getting the full 144Hz.

 

[bigpinkdragon286]

Drivers may or may not be the issue and while I would say swapping a Windows installation from one set of hardware to another generally works, there are plenty of times it doesn't. If you're feeling concerned, I would recommend a stroll through the Windows Device Manager and see if anything is reporting an issue.

USB drivers are pretty generic and supplied by Windows itself unless you're running something a tad more gourmet, but your description sounds pretty straight forward.

I'm not certain which ports you are actually talking about here when you mention: "I moved my mouse and keyboard to the 2.1 ports and I havn't had a single lockup since." You didn't specify which ports your keyboard and mouse were plugged into previously and I'm finding your description ambiguous to which ports they were moved, as according to the manual for the Gigabyte z790M Aorus Elite AX ICE, there is 1 x USB 3.2 Gen 2x2 Type-C, 1 x USB 3.2 Gen 2 Type-A, 4 x USB 3.2 Gen 1 Type-A, and 4 x USB 2.0/1.1 ports.

As the USB 3.2 Gen 2x2 port is managed directly from the CPU, I would suspect if there were issues related to undervolting the CPU, they may affect this port. The rest being chipset-managed, I wouldn't lean in this direction.

I would however be concerned if I had USB ports that caused system instability when in use. What if you need them at some point? I'm not certain the undervolt has actually been determined stable. What do you have your Windows Power Profile set to? Have you verified your system has run the CPU at idle states since changing where your USB devices are plugged in? Have you tried moving one or both USB devices back to the original USB ports to see if the unstable behavior returns in a reliably repeatable manner?

While I wouldn't rule out the possibility, I'm not convinced the undervolt is stable or that the USB is related.

You mention dialing your memory up to 6000 from 4800. Have you run any memory tests to ensure the memory is stable across it's entire range? There are a few different memory testers you can find online, but even Windows has a built in memory test you can run overnight or when you won't be using the system for a while. I recommend running at the very least, the built in Windows Memory Diagnostic test. If the system can't fully pass even the built-in test, your memory is throwing errors. Doesn't mean the memory is defective, but your settings certainly could be configuring it to an unstable state.