XPS Desktops

[kalarous]

So here's a fun one. I too (as many on this forum have), have been having random system freezing issues with my Dell XPS 8940 with Nvidia RTX 3060 Ti video card. After a lengthy phone call with Dell support, the support engineer in their infinite wisdom has decided that the system doesn't have a hardware problem and if I'd like any further help resolving this situation, I can fork out another $100 for "software" support. System is less than 1 year old and still under warranty. I spent close to $2500 for what was supposed to be Dells "top of the line" Desktop. So, I'm left with 2 choices - either live with a problematic dell desktop or shell out another $100 for their software support and HOPE that their software support personal can help me fix this issue. Needless to say, I will be seriously considering whether or not to purchase ANYTHING else from Dell, ever.

[ispalten]

Well, I see Support has changed its tune. Used to be re-install the drivers, doesn't work, Windows... and then might be motherboard of video card need to be replaced.

It is not Windows or the Video card.

I wonder if they want to find someone that they can 'test' on to determine the cause?

Anyway, some more info from you, what BIOS are you running?

Did you try back-leveling the BIOS to V2.3.0? That fixed it for most of us.

I assume you've used the Intel HD750 and had no problems as well? Ask Dell how can they explain that if it is a s/w problem?

Another choice, get a different Video card, but NOT Nvidia.

[kalarous]

Hey Irv,

Received system with Bios 2.4.0 - Currently running 2.8.0 (recently updated BIOS)..  It is my understanding that there are some significant security vulnerabilites that have been patched in BIOS since 2.3.0, I'm somewhat hesitant to downgrade that far back.  I have not pulled the RTX card out of the system as I'd also lose 1 of my 2 monitors to do so.     I may have to bite the bullet and go find me a decent AMD Radeon someplace and throw away the $$$ that I spent on the Nvidia...   It's more an annoyance than anything else as a simple FORCE power off, reboot and I'm up and running for anywhere from 3 hours to 5 days before the system locks up again.   I used to be a Dell fan, but after this experience and being invited to spend yet more money if I want support for a system less than a year old that I've already spent a ton of money on, I'm much less inclined to either A) buy future Dell products and B) recommend Dell to others...    I've been in IT for more than 30 years (primarily in Systems Administration and Engineering and I do have many people ask me what I think of various hardware/software companies.   I can tell you that after this support call and the current "recommendation for further help", my willingness to recommend Dell has decreased significantly.

If Dell is looking for someone to "test" on, I'm waiting with an open mind and ears. Unfortunately I have yet to find a way to reproduce this error on command.. It seems to happen most often when I have like either Netflix or YouTube open (not necessarily playing anything - but open).  However even that is not 100% reliable to cause the system freeze.   If Video is playing, it'll start to stutter and then the keyboard/mouse become unresponsive, but not always.     

If I were the only person to have experienced this problem, I'd be a lot less inclined to complain this loudly about being asked to spend even more money for any help.  The hardware support guy did NOTHING more than I have already done at least a half dozen times since I've had the machine trying to figure out what's going on.  System logs are useless at finding root cause.  

Thanks for listening.

Steve

[ispalten]

Steve. I hear you.

I've been in this boat since V2.4.0 came out.

Yes, going back that far is not the wisest thing to do. But I suspect CVE risks to a home user is minimal.

I am on V2.8.0 myself. I've not had, what I call, a 'Lock-up', since 7/15.

Take a look at my Reliability Monitor:

Concentrate on the red box.

I put on V2.8.0 on the 7th of July I think? On 7/19 I added some RAM. Since then, working fine.

I do have another question? How much RAM? Total size and number of DIMMs? Do you KNOW the RANK they are running at, Single or Double.

I ask as that could be a hint or possible verification of why?

See, there is another user here, with 2 'identical' configurations, when shipped basically. One doesn't lock-up, the other does. The User did make a change though. Took the 2 8GB DIMMS from one XPS and added them to the other one. Then Bought new RAM so he'd have 64GB''s and the other XPS (that does Lock-up) has 32GB's.

What is the difference you ask? The RANK of the DIMM's. Locking up one is SINGLE Rank, the one that isn't, DUAL Rank..

I suspect you have 2 8GB DIMMS for 16GB (standard base RAM on 8940's). You can use CPU-Z to get that info.

Yes, we all have the same issue, happens randomly, no one can find any info why, and no common action or cause it seems.

I am sure you know the hardest to debug are random event, you never know what kicks it off.

Me, I'm sure Dell is working on it, as each release of the BIOS from V2.5.1 to V2.8.0 seem to have the problem occur less often. They just have not fully located it in the code (I don't think they have a system that can reproduce the problem?).

Yes, it has taken way too long... and I have lost data in some programs (settings mostly as the program had them open, reading, or writing them when it happened and the file was corrupted... not fun at all.

I would if nothing more put down V2.3.0 to VERIFY you can keep running a few days with no lock-up. Then use V2.8.0 as it seems the most stable to me... and maybe SOON Dell will have a new BIOS with the fix.

[kalarous]

<Profanity removed.TOS73> And I thought I jumped through hoops to find the root cause....      Very thorough there Irv. 

At the moment I have 32 GB (2x16).   Am considering an upgrade to 64 (4x16) as even at idle I sit around 19-20GB used RAM (Actual IN USE - not cache in use)....   Between Docker, WSL, VScode, and a few other utilities that are generally running on my machine - I keep it busy.  I've only been on 4.8.0 now for about a day now - I'll have to see what's what over the next few days.   Fingers crossed "we're getting warmer"....   If it continues to be a problem, I'll have to dump the Nvidia (now there's something to make me sad) and go with an AMD...    Maybe I can find a friend that has an extra that I can borrow for a few weeks and see if I get the same results with an AMD card (not expected based on what I've read in the forums).    It would be nice if Dell would step up to the plate, admit they have a compatibility problem of some type with their 9840 system boards, recent BIOS builds, and Nvidia GPU's.  Would even be nicer if they would offer to trade for comparable AMD card for folk that are affected by this and are still under warranty, but I'm not going to hold my breath on that one.   Agreed CVE risk is minimal, however - I use my machine for both personal and professional purposes and as such am loathe to not be on the most current BIOS/Firmware/OS/app patches so 2.3.0 is a no-go for me as a solution.  I think at the moment I'm going to let 2.8.0 cook for a few days and see if it's any better.   With your mention of dual rank not locking up may accelerate my plans to move to 64GB RAM on the system (4x16) vs single rank RAM config.   I've kinda been kicking around a RAM upgrade anyway.  Anyway, your recommendations and insights are greatly appreciated.   Hoping someday soon I can update this post that I've not seen the lockup in months and all is well, ideally without having to dump the Nvidia card...   

Have a great day!

Steve  

[ispalten]

Steve, it seems that Dell ONLY ships SINGLE RANK DIMMS. It you still have the original, either 2 8GB's or 1 16GB's (which I have, I intended to go up to 32GB's later and I did it this month finally) you more than likely have SINGLE RANK. Do not confuse that with Single or Dual Channel. It is the Access Method basically of using RAM.

The thought is Single RANK is the cause. Please use CPU-Z to check. It is NOT a function of how much RAM you have.

They are not the same, other than they use the same 'name'.

Look at these links:

https://www.oempcworld.com/support/singlevsdualram.html 

https://www.hardwaretimes.com/single-channel-vs-dual-channel-ram-which-one-is-better-for-gaming/ 

RAM DIMM and Channel are not the same other than by name. I suspect you already know this though. Still good reading.

I had worked with Dell Support I guess in Jan. when this first hit. I 'suffered' the re-installs of Windows, Safe Mode, all that, and all the drivers. Discovered using the HD750 on-board video I had no problems. So Dell Support replaced my Motherboard AND Video card. Still happened. Tried another motherboard, still a problem. I asked for an AMD card... and they said sure, until they discovered they could only replace what I had. At one point Dell wanted to exchange my XPS, but I didn't like the terms. May be refurbished, will be what you have or better, might be new, might be the 8950. I could NOT delete anything. No-go, some programs I have are single use license, and I'd have to remove them before I put them and register them on a new PC. Worse, it took me a few weeks to get everything the way I want it. New system I might not be able to restore it with my back-up either... refused.

If you are not into heavy gaming, try the HD750 on-board video, it will not fail. Once Dell fixes it you can go back to the Nvidia card. Other choice you have.

[kalarous]

Irv,

Thanks much for the background info.   I was familiar with # of channels but not particularly familiar with single vs dual rank.   Those articles were awesome...   I have double checked my PC configuration using CPU-Z... Sure enough, although both 16GB DIMMS were dual channel, they ARE in fact Single rank.  That should be resolved on Thursday, I have 2 new 32GB dual RANK DIMMS coming my way.  I'll pull the single rank 16s and put them in a corner somewhere to be replaced by the new 32s.  Maybe between that and 2.8.0, this will stop being an issue for me.  My biggest issue with the HD750 is that there is only a single port.   I would not be able to survive (ok, maybe that's a little over dramatic) with only a single monitor, but my life would <Profanity removed.TOS73> big time and I'll be a very sad engineer trying to fit my entire view onto a single monitor.     Cross your fingers and say a little prayer for me that the dual rank and current BIOS put this silliness to bed.   Again, you are a Godsend with your knowledge on this subject...  MUCH more helpful than the Dell support engineer.   I greatly appreciate your time and efforts on my behalf.   Dell, if you're reading this - hire Irv.... He knows his stuff!

-Steve

[ispalten]

Steve,

When you get the NEW DIMM's in that are Dual Rank and they are the ONLY ones in the XPS, please PM me or reply here with the details. I can probably pass it onto someone in DELL, it might be the CLUE we are missing.

Only problem, we might think it is fixed only to have it pop up again days or weeks later.

If you DO work, go to DUAL RANK, I might just buy another 16GB Crucial DIMM and junk the Sumsung 16GB Single Rank Dell provides.

Good Luck.

Oh, Dell didn't hire me but they did notice I help people and made me a 'Technical Expert''. I can put that on my signature and use a unique Avatar, I've chosen not to. Thanks for the compliment though as well. Glad to help.

[ishmael]

My XPS 8940 has a 16GB single rank stick (SK Hynix) in slot# 1, and I added a second 16GB dual rank stick (Crucial) in slot# 4.  CPU-Z indicates memory is dual channel.   Motherboard is 0K2CM7 and GPU is a Nvidia RTX 2060 Super.   Have not experienced freezing problems while in Windows (using Win 10 Home), but did experience PC locking up during startup shortly after the Dell logo.  This seems to be BIOS related, (currently at 2.8.0), but also disabled unused Sata drives per one recommendation here. 

[ispalten]

Interesting....

I am set up slightly different after buying and adding a Crucial 16GB DIMM (can operate as both Single and Dual Rank) to my existing Single Rank Samsung.16GB DIMM.

Original Samsung was in slot 2, 2nd below the CPU. I put the Crucial into Slot 4. I am on V2.8.0 and have had no lock-ups since 7/15, having added the DIMM on 7/19.

I've begun not to trust either CPU-Z (and other programs) or the XPS how it is reporting DIMM slots.

For instance, this is what I see in CPU-Z for both DIMM's:

This shows I'm running Dual Channel and speed is 2933Mhz as it is DDR so it is 2X

Here is where it starts to go wrong I think?

First, it is called Slot 1? But the DIMM is in Slot 2? Rank is Single which I know the DIMM is.

The Crucial (actually it is Micron) it says it is in Slot 2, but I know it is in Slot 4. Also Single Rank?

Another program however does call them by the correct slot, AIDA64:

DIMM2 and 4 which corresponds to the slots.

You are actually running 2 channels though. Slot 1 and 3 are the 1st bank (channel) and 2 and 4 are the 2nd. Normally one would put the White Clips first (at least that is how Dell does it I think?).

Check this link, https://www.intel.com/content/www/us/en/support/articles/000005657/boards-and-kits.html  as it explains more what DIMMs go where. From it:

This is how Dell ships a single DIMM system:

This is what I've done:

What I believe you have is 2 channels each with a single DIMM in it. However, the Crucial should have been in DIMM1 or Slot 2. Maybe that is why you having a problem after the Dell Logo.

Yes, you are Dual Channel as you are using both. It is possible the Hynix will show Single Rank, and the Dell Dual Rank as you've got them in different banks (channel).

I'd really be interested in seeing your CPU-Z output for both DIMMs?