XPS 8940, randomly freezing

We are fairly close in Configuration:

• BIOS Version 2.9.0
• Nvidia Card RTX2060 6GB
• Nvidia Card RAM 6GB
• Monitor Resolution 3140x2160
• Monitor Port used HDMI 1
• Intel CPU i7-11700 @ 2.50GHz, 8 Core
• XPS Standard or S.E model S.E.
• XPS RAM total 32GB
• XPS DIMM size in GB's 16GB
• XPS DIMM slots used 2, 2 and 4
• DIMM Rank Dual
• Windows Version W11, 21H2 Home, Build 22000.856
• Any BIOS changes other than turning off Capsule Update Just to use AHCII from RAID for the SSD
• Is SSD running RAID or AHCI AHCI
• Any ADDED internal h/w (disks, PCIe cards, etc)
• Any ADDED external devices (USB Hard drives, non-Dell wireless Keyboard/mouse, etc.) Logitech Wireless mouse, M310 and Keyboard K540 and a WD USB 3.1 External 4TB hard drive

The M.E. Interface is MS's and I've got the same.

Why did I go back to V2.4.0 vs. staying on V2.8.0? Easy, after 5 days it didn't happen. So since it took only a day or two on V2.4.0, I figured it would be a better test of Dual Rank... If it wasn't fixed, I'd hit the problem sooner. If it was, I'd stay on it for maybe 2 weeks and declare it 'probably' fixed. Didn't matter, within a week I hit the lock up. Dual Rank was NOT the answer.

So I went to V2.9.0 as each new BIOS seems to take longer before it fails. CVE's are something you'd want fixed anyway. Running on a very old BIOS loses the protections the new BIOS provides.

You wrote, "Well, that might mean it was fixed in the V2.8.0 BIOS even, so I went and installed V2.4.0 again."  Why go back to v2.4.0?    Why not wait to see if it failed on 2.8? 

Here's my configuration again, along with some additional info you requested, but don't know about AHCI vs RAID.  I have not changed anything related to the drives except to disable SATA-1 and SATA-2.

Desktop PC:
XPS 8940, Special Edition
Purchase date: 07/22/2021
CPU: i7-11700 @ 2.50GHz, 8 Cores
GPU: Nvidia RTX 2060 Super, 8GB
Memory: 16GB single rank stick (SK Hynix) in slot# 2, added a second 16GB dual rank stick (Crucial) in slot# 3.
Crucial part# CT16G4DFD832A
CPU-Z shows: memory is dual channel: 2 x 64-bit, Hynix rank single, Crucial rank dual
Drives: C: 512GB SSD, task mgr shows NVME BC711 NVME SK hynix 512GB
1TB HDD, task mgr shows WDC WD10EZEX-75WN4A1
E: CD-ROM, ATA PLDS DVD+/-RW DU-8A5LH
Motherboard: 0K2CM7
O/S: Win10 Home--> 21H2 (OS build 19044,1889)
BIOS: v2.8.0
Intel Mgmt Engine Interface: v2145.1.42.0, 10/31/2021
Cooling: Replaced rear stock fan with Noctua NF-A9 92mm fan, replaced stock CPU fan with a Vetroo 5,
added Dell VRM heatsink, using HWinfo64 to monitor CPU/GPU

Monitor: Dell_U2715H, 2560 x 1440 QHD at 60Hz, using Nvidia port
Devices: Logitech USB wireless mouse M510 and keyboard K330

I ran the Memory Mark test selecting run all the tests for my memory in slots# 2 & 3.   Here are the results:

Interesting, your overall speed is 3636, mine was 3473? I expected mine to be faster by 3 to 5% based on not running 2 Channels and one being Single Rank in your case.

So I tested again, and got this:

Now it shows 3660, slightly faster than your test? Since we both have 32GB's of RAM, each time I have less free? That should not play into this though.

As for you slot usage, I found this under the 8930 FAQ about slots, which I'm sure applies to the 8940 and 8950 as well:

I wonder how accurate that Memory test actually is? I ran it again, got numbers closer to you?

It could more have to do with some background tasks also accessing RAM than anything else?

---

@ispalten  schrieb:

I wonder how accurate that Memory test actually is? I ran it again, got numbers closer to you?

 

It could more have to do with some background tasks also accessing RAM than anything else?

---

Forget about the accurancy of these software tests.
The results are never be exact the same, also an equal machines.
There are slight differences, that doesn't matter. Run at least 10 tests and middle the results.
But that is also only chit-chat and has nothing to do with any lead to a  final solution.

Well, today the XPS8940 of my girlfriend locked-up, after a couple of days without any problem.
And again, it locked-up when there is no workload on the maschine. We paused our gaming session and left for a break. When she came back - screensaver was frozen -only way out - power off/power on..

After that I updated both machines to v2.90 - my machine still runs flawless without any problem, 
but hey I update both machines identical. The differences of both machines are written in the other topic. So i don't repeat it here. Now I have to wait how long it takes, until I hear again *oh no! it is locked-up again* - hopefully v2.90 fixed the problem - we will see.

____

DK

Yeah, I feel all those 'Performance' tests give you a number to rank against others. Too many 'things' can decrease the values or even increase them. Depending on configuration and running processes at the time can give different results.

Thanks for the update on the fail...

It seems to me that with each BIOS Release since the problem started on V2.4.0, the time between failure is longer.

S/W will usually always do the same thing unless it is a timing issue or data input difference normally.

There are changes in the time between fails though from release to release, we've all seen it.

Is it Dell making some coding changes?

We seem to know it isn't RAM DIMM Rank, at least 3 of us here have all DUAL RANK RAM and fail, others either Single Rank, mixed Rank, or have not specified what they have.

All have Nvidia cards, but not everyone fails with it.

Both Windows and Linux users have the fail, so it can't be Windows, Linux or any program.

If one focuses on the BIOS, then it is either a coding change or possibly something that changes with most releases, the Intel Management Engine. Both the BIOS and Intel M.E. have changes made for CVE's. It is also possible that Intel provides only a Binary to Dell for inclusion in the BIOS?

Next, staying with H/W, are all Nvidia cards created equal? Power consumption,  RAM speed, other parts?

BIOS also on power up has the job of setting up the RAM. It loads the SPD that the OS (CPU-Z shows you one for each DIMM for instance) for the Memory Controller Chip to access the RAM. It does this by reading the RAM itself. However, if it has a problem reading, it loads a DEFAULT SPD. There is a lot more to this, what if the DIMMs are different for instance I've not seen described. So could this be part of the problem, but it shouldn't be, I'd expect the proper SPD loaded for each DIMM, unless it couldn't be read... but then again, too many of us have this problem?

Same Nvidia cards are in the 8950, and possibly most of them in the 8930's, yet those users don't seem to be reporting the problem.

So what IS left? Minor differences in the h/w?

You probably more than any other user (and Dell should too) have the capability to possibly narrow it down. Swapping the RAM DIMMs between the XPS's and see if the fail follows the 8GB DIMMs. Still works (can take days though, running with BIOS V2.4.0 might speed it up though), put the RAM back and swap the Nvidia cards. If either of these are the problem, you'll know it. If the fail stays on the XPS that had the problem before, well, I'm not sure where it might be? At one point Dell wanted to replace my Motherboard, but eventually users with the newer one reported the problem, so that isn't it.

I almost feel that if Dell can't solve the problem, they should replace the Nvidia cards with AMD ATI cards of the same capability and be done with this. Losing data is NOT fun.

Our configurations are indeed quite close, and yet I do not have lockup or freezing issues.  Hard to believe it's the slight difference in our RAM configuration. 

It would be good if the Dell moderators here would consolidate these various lockup/freezing discussions, where no known solution has been found, into just one discussion.  My recommendation for Dell is that they offer to replace those machines that are experiencing these symptoms.   Perhaps a letter to your local attorney generals office would spark some consumer relief, since, although am not trying to be troublemaker for Dell and have used their systems for years, but I don't know how you and others can tolerate using a system that can lose data.  It's just not reasonable. 

---

@ishmael wrote:

Our configurations are indeed quite close, and yet I do not have lockup or freezing issues.  Hard to believe it's the slight difference in our RAM configuration. 

---

I wouldn't this that could/would make a difference. Too many other configurations failing with the LOCK-UP and of course many that have no problem.

One discussion, fine, but that is really 1/2 the story, the other 1/2 (or more) is those that have no problem.

That is why @DataKill  ll has the best shot to sort of pin it down or point in the right direction. Swapping parts and see when the fail follows the parts. Hopefully it would, but if it doesn't it can be murky... who wants to swap out motherboards, PSU, Cooler's, or even the CPU? These parts especially when under Warranty.

Dell has to do this, but they need a failing system. Having a failing system is only part of the equation, how to recreate the lock up is the hard part.

All they need replaced is the Video card, but NOT with an Nvidia, ATI would probably be good enough for most of us.

BIOS V2.9.0 is not the answer either... lock up yesterday trying to close Zoom, a 'new way to fail' for me.

I was searching the web yet again for Lock Ups, found this You Tube video, https://www.youtube.com/watch?v=27zapKfts-0&ab_channel=Pennicles 

Already had the Link State set right, but not the Settings for the card?

Mine was set to 'Optimal Power', so I changed it:

Will it work? Probably NOT, but if this IS a power issue, something to look at?

Time will tell I guess?

Just a wild idea...
Open Device Manager and expand list under System Devices. Find the entry for Intel Management Engine and on its Power Management tab, uncheck the box "Allow PC to turn off...", and reboot PC.

Intel Management Engine Interface #1 is the one I meant. Are you saying that "Allow PC to turn off..."  isn't checked by default on the Power Management tab?

I don't have freezes on my XPS 8930 but that box is checked for IMEI#1 by default here, which is why I mentioned it...

Ahh, good IDEA!!

Uh, can't really find that specifically in Device Manager, but I can get 'close', the OS interfaces it seems. The actual 'engine' is installed with and part of the BIOS I think?

Anyway I did find these 'Intel' devices:

Last one is the only one even with that Power option to turn it off enabled.

Now, what is it actually is the question? Even Dell has a few pages on it, but most are for tablets or NUC's (Next Unit of Computing, the small PC's).

Two things I can think of:

• Possibly installed on ALL versions of Windows, even if not needed?
• I'd not think this is the cause, since the Lock-ups occur when the PC is in use, although there have been some when the PC was idle.

It also an MS driver, not Intel's.

Google isn't much help, and in some cases I've seen where it just isn't known what is does exactly?

You have any better luck on your XPS?

If that one that can be left on was doing that, you'd think other things would have problems? Even if it did, I'd think 'whatever' was trying to call it or talk to it would get an error and post it?

Yes, it IS not checked by default:

So it should be ALWAYS on.

The Host Controller DOES have it checked by default so it can be turned off.

Again, if anything, these being allowed to turn off to save power should only be an idle state action, not when the PC is in sleep, no?

I don't think any of these are related to the lock-ups. Odd that you have the ME Interface #1 check on the XPS8930? Again, can't see having it unchecked will do anything but use more power in Sleep?

I do see many web pages about "Intel Management Engine Interface #1 property allow the computer to turn off this device" when I Google it.

For instance:

https://techcult.com/fix-windows-10-will-not-shut-down-completely/ See #4 and 5 only.

https://techunwrapped.com/get-rid-of-black-screen-error-before-turning-off-pc/ 

This explains somewhat that the M.E. is, where it is, and what it does, https://www.intel.com/content/www/us/en/support/articles/000008927/software/chipset-software.html  and from it, I gather:

• During Install of the BIOS, ME is installed, either in the BIOS RAM itself or a Northbridge Chip.
• It is always running.
• It is needed to wake up the rest of the system.
• It resides in Protected Memory (which is set up by the BIOS.

Given that, turning it (ME Interface) OFF seems counter intuitive?

It also seems all that I've looked at mention W10, not W11? However, I've not seen a single report of owners with Nvidia cards having a lock up on an 8940 or 8950? Why is that?

So, is it possible that setting on all W10 PC's have it off by default? Mine was, and it is on however, but I went from W10 to W11... and thanks to Dell Support, I had to reload the OS when I was on W11 and had to use the the Windows Media Creator to do that. Didn't want to do twice the work recovering W10 and then going to W11. Some XPS8940's probably came with W11 on it as well.

Bottom line, I can't see if/how allowing the Interface to turn off helps here vs. it always being on?

By the way, the problems the links describe is EXACTLY the problem my wife's XPS8500 on W10 had for quite some time. Rarely did Shutdown complete. Had to finish it be manually using the Power Button. Nothing in the Event Viewer, not even an Error 41 for powering it off. Just had to wait for the power light to stop blinking and there was no disk activity and power it off. Then, months later, it 'cured itself'. Bet a Windows Update for the Interface was changed (or a new one installed)? Odd.

As I read the above link, I'm not sure I want to turn it off? Especially since I have Sleep NEVER enabled so it shouldn't matter and I am on A/C power. Might be a different story for a Laptop though or someone that does allow Sleep?