Use Tigerlake CPU to execute Passmark BurnIn show no response after resume from S4

[phssub]

Hi all,

Recently, I use Tigerlake CPU to execute Passmark BurnIn show no response after resume from S4. And found out that also have similar issue in this forums. If burnIn support do hibernate and won't get no response after resume from s4? Or if burnIn this tool doesn't support specific CPU? Thank you.
BurnIn version: v9.2

[David (PassMark)]

Sounds like this issue
https://forums.passmark.com/burninte...ume-from-s3-s4

[phssub]

Yes, it seems the same issue.
But I wonder about this issue will fix by new version of BurnIn or it might relate to platform and can ask for intel's help?

[David (PassMark)]

At the moment we have no reason to think it is a problem with the BurnInTest software.

Sleep & hibernation have always been problematic on Windows. It is very hard to get the device drivers and hardware to transition from a high load state to a sleep state smoothly. And you only need one slightly bad driver to break the whole system.

[Barry]

Hi David,

I also encounter this problem on Intle Tiger_Lake CPU.
Do you have any suggestion for how to capture log? Since we used debug mode(https://www.passmark.com/support/bitdebug.php) cannot reproduce issue.
1. I used pure OS (OS without 3rd party driver) still can reproduce this issue. So, It seems not driver issue.
2. And I rollback OS to 20H1, it still can reproduce.

[David (PassMark)]

The base install of Windows, includes a lot of device drivers. So you can't rule out a device driver bug.

We have been unable to reproduce across several test machines (but we haven't tested with Tiger Lake as we don't have a machine with that CPU model) so likely there is some hardware or driver issue specific to this system that is triggered the lockup in BurnInTest.

If you have temperature monitoring enabled then try disabling it as this is very driver and timing dependant.

[phssub]

I had disable temperature monitor and found out it won't show no response after resume from S4.As your reply you mention that can't reproduce at other platform. I wonder if this only happen on Tiger lake? And why it won't happen after I disable temperature monitor?

[David (PassMark)]

We don't have any hardware that has this behaviour.
But probably what needs to happen is after the lockup you attach a debugger to the machine and break into the process and the hopefully identify the device driver that is locked up.

We collect temperatures from many sources (RAM, CPU, GPU, Disks) so there are several device drivers involved. Does this CPU you are using have a integrated video card?
Can you send us an example machine to test?

[phssub]

As another user mention that cannot reproduce issue when use debug mode. I tried debug mode and found out this really cannot reproduce issue. I wonder if it relate to temperatures why it will have between normal mode and debug mode? And only this tool will no response system still can work?

[David (PassMark)]

Our guess is that it is a timing issue on this machine that exposes a bug in it's device drivers.
It is only a guess, as we don't have an example of this machine to test on and the problem isn't happening on other hardware.

[phssub]

I wonder if there will have any impact to disable temperature monitoring? My test condition is to set CPU, 2D graphics, RAM and Disk each to 100% and run test for 15 minutes.