Solved Investigating random BSOD since 2017 on a Thinkpad...

grospoulet · Apr 25, 2022

Hello everyone,

I'm still investigating from time to time some random BSOD I could never fix on a Lenovo T440P... I posted first in 2017 on tenforums: Random BSOD even with fresh reset, can't find the cause - Windows 10 Help Forums

Things went well for a long time, BSOD were so rare I didn't bother investigating.
But, recently, I got some file corruption issues (ie: multiple database corruption in Syncthing).

About my laptop: bios is modded just to remove the Wi-Fi card whitelist from Lenovo. I got BSOD before and after this modification, I'm pretty sure it's not related.
Otherwise, BIOS is reset to default settings except secure boot is enabled. I disabled Throttlestop so there is no undervolting or TDP modification.
Obviously, this machine was built in 2014 and is not under warranty. Also, this is not officially supported by Windows 11 although the behavior was the same in Windows 10 (a few random BSOD with apparently the same causes).

I did some pre-check steps:
- Windows 11 was installed from ISO not long ago (reset while keeping programs and user data)
- Lenovo BIOS diagnostic : OK
- Intel processor diagnostic tool: OK
- Windows memory diagnostic : OK (default settings)
- Passmark memtest86: default settings (parallel mode), 4 passes OK

BSOD happens specifically :
- after several minutes to one hour of prime95 blend torture-test
- after about 15 minutes of AIDA64 stress test
- sometimes when shutting down Windows (not "reboot")
- whith only one of any of the two RAM sticks, it may crash on shutdown but it seems much more stable with stress tests (didn't test more than one hour)

I tried to be concise so you can read it quickly and help me to diagnose this with the right steps...

Thank you very much in advance and thank you again for these forums (ten & eleven) which help a lot of users like me.

zbook · Apr 26, 2022

The logs displayed piracy / crack.

Similar piracy in the prior thread: (see post #2)

Random BSOD even with fresh reset, can't find the cause - Windows 10 Help Forums

Hi, I'm having BSODs on my laptop since months and I struggle finding the reason. I asked for help on the Microsoft community, but I thought I should give a try here as well. I work on a Lenovo T440P

www.tenforums.com

grospoulet · Apr 27, 2022

Hello @zbook,
Thank you for your quick reply.
I remember I used an unlicensed version of Adobe software (Photoshop & Illustrator), is it what you are talking about?
I use them so rarely I uninstalled them before doing stress tests and system dump with your script, but there might be some modifications still in place (hosts file...).

Could you please let me know which software / downloaded data could be causing the issue? I would be glad to remove them.

grospoulet · Apr 27, 2022

This is the first time I try to run Prime95 in safe mode boot and for now it is stable... almost 6 hours.
But the CPU clock stays at 2900Mhz (nominal frequency, without turbo) and the fan runs slower than in normal mode boot.

So I guess my RAM and my motherboard are okay and this is a software-related issue.

I have a friend having the same laptop (Thinkpad T440P) who kindly sent me a report (unfortunately in French). It might help us find the difference between the two configurations?

glasskuter · Apr 27, 2022

Are you monitoring your heat during these tests?

grospoulet · Apr 27, 2022

Thank you for your reply!

I didn't log the temperature but the chassis cannot cool it down correctly.
When running Prime95 (small FFTs for maximum heat), CPU core & package are between 90° and 93°C with some CPU throttling. When watching with ThrottleStop (I don't modify any settings, as stock), PROCHOT is often triggered.

Prime95 ran fine for more than 9 hours in safe mode without overheating: I don't have access to CPU temperature in safe mode but the fan runs slower and the laptop seems way cooler.

New thermal paste has been applied but the T440P is known to run best with 37W TDP CPU, and maybe it's because I upgraded it with a 4910MQ which is a 47W TDP CPU. With that said, the other laptop has a 4700MQ CPU which is also a 47W TDP CPU, and it runs Prime95 just fine.

I used Intel XTU a long time ago but I uninstalled it. Throttlestop is not launched at boot.

You may be on the right path, I guess some CPU limits are removed when Windows runs in normal mode and the CPU is way hotter than allowed. But I don't know where it comes from.

grospoulet · Apr 28, 2022

Hello again,

This might help other people in the same case.
I launched ThrottleStop and made sure in TPL settings that "Clamp" is enabled in power limit controls.
So, when the CPU is really stressed (such as with a Prime95 test), the package power never exceeds the long power (PL1) value.
The Lenovo T440P was sold with a 37W TDP CPU at most if I remember, but the bios allows up to 47WTDP CPUs.

When setting PL1 to 47W, I get some CPU throttle in Prime95.
When setting PL1 to 43W, I don't have any CPU throttle in Prime95.

In both cases, I didn't have any BSOD with Prime95, OCCT, Aida64 system stress test... It's currently stable for 4 hours, which never happened before.

Indeed, when "clamp" isn't checked in Throttlestop, the minimal CPU frequency under load is it's nominal frequency (2.9Ghz for my CPU). Even if the load exceeds the TDP PL1, the CPU won't be clocked under 2.9Ghz and it might exceed it's nominal TDP (47W) in some specific computation (Prime95, OCCT...).

When disabling clamping, my TDP somehow went up to 60W (!) in OCCT (AVX2) and slowly went back to 47W. after a while, it starts to throttle because of PROCHOT throttle (thermal throttle), not TDP. Then it goes back to higher than nominal TDP and I observe several TDP spikes like this before it crashes with a BSOD:

Now, I don't have any BSOD anymore, but I need a software like Throttlestop to handle this. I should continue to investigate why this behaves like that by default, but about this BSOD... I think the problem is solved. :)

TopgunB · Jul 17, 2022

grospoulet said:
Hello everyone,

I'm still investigating from time to time some random BSOD I could never fix on a Lenovo T440P... I posted first in 2017 on tenforums: Random BSOD even with fresh reset, can't find the cause - Windows 10 Help Forums

Things went well for a long time, BSOD were so rare I didn't bother investigating.
But, recently, I got some file corruption issues (ie: multiple database corruption in Syncthing).

About my laptop: bios is modded just to remove the Wi-Fi card whitelist from Lenovo. I got BSOD before and after this modification, I'm pretty sure it's not related.
Otherwise, BIOS is reset to default settings except secure boot is enabled. I disabled Throttlestop so there is no undervolting or TDP modification.
Obviously, this machine was built in 2014 and is not under warranty. Also, this is not officially supported by Windows 11 although the behavior was the same in Windows 10 (a few random BSOD with apparently the same causes).

I did some pre-check steps:
- Windows 11 was installed from ISO not long ago (reset while keeping programs and user data)
- Lenovo BIOS diagnostic : OK
- Intel processor diagnostic tool: OK
- Windows memory diagnostic : OK (default settings)
- Passmark memtest86: default settings (parallel mode), 4 passes OK

BSOD happens specifically :
- after several minutes to one hour of prime95 blend torture-test
- after about 15 minutes of AIDA64 stress test
- sometimes when shutting down Windows (not "reboot")
- whith only one of any of the two RAM sticks, it may crash on shutdown but it seems much more stable with stress tests (didn't test more than one hour)

I tried to be concise so you can read it quickly and help me to diagnose this with the right steps...

Thank you very much in advance and thank you again for these forums (ten & eleven) which help a lot of users like me.

I have the same problem have been searching for months finally found the solution. Re-installed win 11 twice! Here it is: Microsoft ACPI-Compliant Control Method Battery is a battery driver for Windows that deals with battery usage in your laptop. The ACPI stands for Advanced Configuration and Power Interface. The OS communicates with the BIOS, with the help of ACPI, to power up different internal hardware. There is a persistent Win 11 bug with this driver and computer usually crashes when battery reaches 100%. One solution is to pull out the power before battery reaches 100%--does not work for everyone, The other workaround is Device Manager-- expand Batteries-- and find "Microsoft ACPI-Compliant Control Method" Rt click and disable this then reboot-works for me! Unfortunately the battery icon and control disappears. I have informed both Lenovo and Microsoft of this hope they fix it soon!

Solved Investigating random BSOD since 2017 on a Thinkpad...

grospoulet

New member

Attachments

My Computer

System One

zbook

Well-known member

Random BSOD even with fresh reset, can't find the cause - Windows 10 Help Forums

My Computer

System One

grospoulet

New member

My Computer

System One

grospoulet

New member

Attachments

My Computer

System One

glasskuter

aka Mama Glass

My Computers

System One System Two

grospoulet

New member

My Computer

System One

grospoulet

New member

My Computer

System One

TopgunB

Member

My Computer

System One

Latest Support Threads

Latest Tutorials