Laptop keeps eating NVMe drives via 1000 power cycles per hour

Yep, classic modern standby response.

So that seems to cause a high number of power cycles and [perhaps] those by themselves aren’t as bad as I thought.

Need more data, I guess…

Thanks for your help!

I am not sure if I helped, but if I have something that can increase the surface area of information on an issue, I will try to provide what info I can. I haven’t heard of or seen anyone else with this kind of odd behavior until I saw this post lol. Let me know if you want me to check anything else on my laptop.

When I was looking for an easy way to disable Modern Standby and see how that would influence matters, I found an incredible number of registry data on power matters, including all kinds of assumptions about the milliwatt consumption of components and actions, which I suppose are there to help the OS decide on what to suspend, when and how.

The amount of control available seems huge and It’s certainly a complex set of decisions which are made there.

Lenovo has its own set of intelligent power management parameters which are active and I have no idea how they might conflict or harmonize with AMD chipset software.

I’d just like my old suspend to RAM back on a lid close, because that makes things very easy to control. Unfortunately that doesn’t seem to be available on Windows (“no firmware support”) while it works just as expected on Fedora 40.

Manual isn’t always worse than algorithmic and let’s not add AI anywhere in this picture!

Thanks, and if you have any other laptops (or people come to you with theirs), that would be nice to add to the data pool.

I know of a few machines that are of the same product line as my laptop. Ill steal their smart info soon. A few others I know have decently recent laptops, so i will check with them as well if they are willing. If I find anything interesting, I’ll let you know.

2 Likes

Just checked a couple Lenovo ThinkPad Ts:

  • older Intel without modern standby: SN730 with 6156 power cycles and 2043 power on hours, which is a pretty reasonable ratio given the sleep timer. 115 unsafe shutdowns.
  • newer AMD with modern standby: PC711 with 70205 power cycles in 4109 power on hours, 4 unsafe shutdowns.

So that’s 17 power cycles an hour versus 3.0.

1 Like

this is really lame though tbh. Designing a storage medium that wears out every time you lose grid power or the laptop battery runs out (although it should be timed to clean shutdown before there isn’t enough power to run) it just seems like planned obsolescence.

1 Like

I’m pretty sure that’s not how it was intentionally designed to operate.

To be frank I did a clean install of Windows 10 (originally) Enterprise on that machine, because the OEM variant from Lenovo had all kinds of crap on it and I don’t trust Lenovo to a) not try to collect tons of extra data b) do that in a manner that doesn’t huge security gaps: their public record on these matters is really bad, not that others are much better, mind you.

I then did install all the Lenovo specific drivers, but none of the other “tools”.

When I upgraded to Windows 11 (long past replacing the original NVMe), I installed the Lenovo provided drivers again, except that I later updated the graphics drivers to the AMD release, because Lenovo hadn’t updated their drivers basically since the launch of the machine and they were several major releases back.

I seem to have lost variable (below 60Hz) refresh that way, but that was an acceptable loss.

After observing the extraordinary amount of power saving data in the registry, I wonder if I should try a clean install on another medium to see if that makes a difference, especially with the “unsafe shutdowns”.

Or perhaps that just lets me get rid of Modern Standby (without losing standby alltogether), which I see more of a liability than an asset anyway.

Lenovo is known to put whatever SSD or WLAN module seems close at hand or cheap today into their lesser models and it’s easy to see how that could wreak havoc on their ability to actually validate a representative number of permutations. Add the complexity of the co-evolution of ACPI and Microsoft’s desperate attempt to squeeze fruity cult battery life out of Windows slabs, and it’s a miracle anything works at all…

1 Like

I always check for newer driver releases directly from the manufacturer over the motherboard drivers from the vendor. Only specific things like vendor specific tools I would use from the motherboard download page, after getting as many of the latest drivers as possible.

Even windows can lag behind on their auto installed drivers, it’s always good to bypass the middle man and double check

For desktops, that’s what I’d do, too.

With this notebook, I’m starting to think that Lenovo might have put tweaks and stuff compensating ACPI implementation bugs in there. I guess there is a reason why AMD warns you not to replace OEM drivers…

Yet of those OEMs don’t maintain their drivers, the onus is on the user…

However, the worst behavior was on the old NVMe drive which only ever saw Lenovo OEM drivers, if none of their tools. So I’m feeling somewhat safe that the 280,000 power cycles for 280 operating hours weren’t directly the result of me messing around.

Gotta recheck the ratio of the unsafe shutdowns on the original drive… tomorrow :sleeping:

disable S3 state
most UEFI have this option for just this reason (Windows is not very good at low sleep states)

nope, just normal behavior for Windows based non domain controlled laptops

1 Like

It’s sad how true this is.

I had a bunch of 12th gen machines that couldn’t even install windows from PXE server without a BIOS update due to power management or some other issue.

Would BSOD halfway through.

As stated, assume that the shipping BIOS (especially early on) is complete trash and has showstopper bugs, because its true pretty often :slight_smile:

2 Likes

Every laptop i have EVER owned has had inexcusable bios bugs. Some are worse than others, but all are inexcusibly incompetent.

1 Like

My MacBooks have mostly been fine, lol (yeah they do get firmware updates, but I’ve never had any showstoppers).

More seriously though, laptops from the early 2000s were much better int his regard - back when people cared about release software quality rather than “ship it! we’ll patch it in the field!”.

sigh

2 Likes

Sorry for the delay, was busy. Got some info from those who were willing to fork it over. Hopefully I will get more to help with this, but getting this info from people has been like trying to milk a cinderblock.

Laptop 1:
ASUS TUF f17 (intel cpu)
2747 power on hours
720 power on count

powercfg

The following sleep states are available on this system:
Standby (S0 Low Power Idle) Network Disconnected
Hibernate
Fast Startup

The following sleep states are not available on this system:
Standby (S1)
The system firmware does not support this standby state.
This standby state is disabled when S0 low power idle is supported.

Standby (S2)
    The system firmware does not support this standby state.
    This standby state is disabled when S0 low power idle is supported.

Standby (S3)
    This standby state is disabled when S0 low power idle is supported.

Hybrid Sleep
    Standby (S3) is not available.
    The hypervisor does not support this standby state.

Standby (S0 Low Power Idle) Network Connected
	Connectivity in standby is not supported. 

Laptop 2:
asus g14 (amd cpu)
3858 power on hours
190969 power on count

powercfg

The following sleep states are available on this system:
Standby (S0 Low Power Idle) Network Connected
Hibernate
Fast Startup

The following sleep states are not available on this system:
Standby (S1)
The system firmware does not support this standby state.
This standby state is disabled when S0 low power idle is supported.

Standby (S2)
    The system firmware does not support this standby state.
    This standby state is disabled when S0 low power idle is supported.

Standby (S3)
    The system firmware does not support this standby state.
    This standby state is disabled when S0 low power idle is supported.

Hybrid Sleep
    Standby (S3) is not available.
    The hypervisor does not support this standby state.

Laptop 3:
dell inspiron 16 plus 2024 (intel)
power on count 797
power on hours 1568

Summary

The following sleep states are available on this system:
Standby (S0 Low Power Idle) Network Connected
Hibernate
Fast Startup

The following sleep states are not available on this system:
Standby (S1)
The system firmware does not support this standby state.
This standby state is disabled when S0 low power idle is supported.

Standby (S2)
    The system firmware does not support this standby state.
    This standby state is disabled when S0 low power idle is supported.

Standby (S3)
    The system firmware does not support this standby state.
    This standby state is disabled when S0 low power idle is supported.

Hybrid Sleep
    Standby (S3) is not available.
    The hypervisor does not support this standby state.

Noticing a trend here? lol

1 Like

So I swapped out the drives that were bad/marginal, did clean Windows 11 23H2 installs on all affected notebooks and for now I see relatively normal power cycles, but then these machines aren’t kept suspended on power, but have been properly shut down most of the time.

Only time will tell if it was an issue with “suspended while connected to power” or modern standby doing its thing, or flaky drives.

I’ve stuck with Samsung Evo+ for most of the PCIe v3 drives I bought, WD850X for v4, but you get all kinds of odd DRAM-less garbage with notebooks you buy and I guess that’s how you pay later.

Special thanks to all of you, who provided me with valuable reference data!