I have not used MSFS in a while (2-3 months).  The release of the PMDG NG3 and the Maddog had me dust this back off.  I updated MSFS and all the add-ons.  The only items in my community folder are the Maddog, FBW A320, and WT CJ4.

Since returning, MSFS is causing my computer to restart suddenly, as if the power went out (but it hasn't).  There's no error message.  No warning.  Performance seems very smooth all the way up to the abrupt end.  It has happened twice with the Maddog, once on the ramp and once just after takeoff.  And I think it happened with the PMDG NG, but I had stepped away while it was compiling WASM for the first flight and came back to my login screen. 

Thinking of what I assume to be the common first responses:

Power Supply: I have a brand new (3-4 months) 1000W power supply.  I have had no issues anywhere else including DCS World in VR, and a handful of racing sims that stress the system pretty hard.  I understand this doesn't absolutely mean a new problem hasn't just developed, but it doesn't seem terribly suspicious at the moment.

Overheating: I don't have a temp monitor up in the sim, but the fans are relatively quiet when this happens. Once Windows is back up (20-30 seconds), the graphics card temp is about 40C, and the coolant temp on my H150i cooler is 33-34C.  They've had a few seconds to cool, but I'm not getting any hints this is an issue.

Overclocking: I don't have any overclocks

Drivers: This is one I haven't been able to fully test yet.  I was running on older drivers (maybe even 4xx.xx).  I've updated with a clean install to the latest Nvidia game ready drivers, but haven't had a chance to test yet.

Does anyone have any suggestions for further troubleshooting or consideration?  I appreciate any thoughts or advice.

 

My system: i9-12900k, Asus Z690-E motherboard, G.skill 32 gb DDR5 RAM, PNY RTX 3090, eVGA 1000W PS, NVME drives x 3, 1 2.5" SSD, Corsair H150i liquid cooler, Windows 10 Professional

 

I have experienced this myself when i was optimizing my system and i  went  too tight on my memory sub timings. PC will post normal and everything else runs nicely but FS reboots without warning in a cpu and memory intensive section (loading screen) and have confirmed the issue being memory by the first or second run of Memtest 64

So i would Recommend downloading Memtest 64 and ensure that your ram is able to pass a minimum of 10 cycles with no errors. (this is probably not quite a long enough time frame but if it passes there is a good probability that you can remove memory as being a possible cause. and move to the next possible issue.

I am also going with the above recommendation due to the very new nature of your system and the fact that new platforms usually have a few memory controller niggles that a round of bios updates usually solve within a few months of that new platform launching.

 

And as Bob pointed out PSU would be my first guess under normal circumstances but your PSU is a thousand watts and is new so unless something is not plugged in properly i moved straight to ram but making sure connections are firmly plugged in is worth the effort.

Hope you are able to quickly isolate the issue.

Regards

Edited May 10, 20224 yr by Maxis

When a computer restarts under heavy load, especially in the absence of overheating, the most likely culprits are related to insufficient power delivery--weak/bad PSU, bad or loose cabling from the PSU, and power regulation on the motherboard itself.  I'd start by reseating all of the power cables (including the power main from the wall to the PSU).  A PSU tester might show a problem, but if it's a weak component in the PSU failing to keep up under load, it might not.  Here, I keep spare PSUs on-hand, so swapping out the PSU would be my next step to isolate the problem.  If it's not the PSU, odds are the VR on the motherboard is the problem, which would entail a motherboard replacement to fix.

I've personally seen both issues over the years--the easiest to fix was a loose power cable from the wall (embarassing), the worst was a mobo VR that required a system rebuild with a different mobo.

Yep, definitely a hardware issue of some kind. RAM would be my first guess as Maxis pointed out. PSU as Bob pointed out would be my second guess.

Edited May 10, 20224 yr by Fiorentoni

what speed is your RAM running in Bios Setting ? I had to have mine run rather slow to keep the system stable with MSFS.

The RAM is seeming a likely culprit.  I started Memtest64 and walked away to let it do its thing.  I came back to my login screen suggesting the computer restarted.  I wasn't around to see any errors, but it would seem pretty convincing.

So, I am not much of a RAM expert.  Is Memtest64 testing the physical RAM?  Or is it possible some software or BIOS setting?...on that note, my BIOS are up-to-date.  Any suggestions how to troubleshoot the RAM settings? 

7 minutes ago, ESzczesniak said:

The RAM is seeming a likely culprit.  I started Memtest64 and walked away to let it do its thing.  I came back to my login screen suggesting the computer restarted.  I wasn't around to see any errors, but it would seem pretty convincing.

So, I am not much of a RAM expert.  Is Memtest64 testing the physical RAM?  Or is it possible some software or BIOS setting?...on that note, my BIOS are up-to-date.  Any suggestions how to troubleshoot the RAM settings? 

On my 690 motherboard it has Memtest built into the Bios.  I had a very similar problem.  The bios Memtest was showing red lines after just 90 seconds.

I bought some 4400 DDR4 RAM (4x8 GB in two separate kits.) I didn't realise (until I checked on the internet) that 4400 was only guaranteed for each single set of two sticks, not for all four sticks together.

Eventually I took it down to 4166 (but at slightly tighter timings).  It is now dead stable.  Passed the Bios Memtest - all green after four hours.  I didn't seem to lose any performance strangely enough - or not a measurable amount anyway when doing before and after tests.

Also, with the bad settings, it always crashed in MSFS, but only once in Windows (straight to blue screen), but that one gave it away really.  I believe I had a couple of black screens straight to reset from MSFS.  All stable now though.

45 minutes ago, ESzczesniak said:

The RAM is seeming a likely culprit.  I started Memtest64 and walked away to let it do its thing.  I came back to my login screen suggesting the computer restarted.  I wasn't around to see any errors, but it would seem pretty convincing.

So, I am not much of a RAM expert.  Is Memtest64 testing the physical RAM?  Or is it possible some software or BIOS setting?...on that note, my BIOS are up-to-date.  Any suggestions how to troubleshoot the RAM settings? 

Well .. at least you have narrowed the problem down. Yes Memtest is testing all the memory addresses within the memory module to make sure data can be read from and written to it.

Anyways .. now to the solutions part.. and this can be applied from multiple angles.

I would

1 Check to see if the ram is on your motherboard's QVL (qualified vendor list). If it is then you should be able to get some information as to what speed and voltage the ram is supposed to be running at and crosscheck in the bios to make sure its at those verified settings.

2. Check to see if there are any Bios updates for your motherboard ..especially ones that are geared for fixing memory stability issues.

3. If you have XMP enabled and the system still isn't stable and you really want to keep it.. Tack on the smallest memory voltage increase available and retest.

4 last resort ... run at the default timings (You will be giving up performance with this option)

5 The I don't know what i'm doing option .. Bring the PC to an expert and properly explain the issue so they can repeat the test and focus on fixing the issue

 

Hope you are quickly back up and running. Dealing with these issues can be a bit rough when you thought you had a rock solid system and find out that some component is not stable or flat out doesn't play well together.

All the best

Edited May 10, 20224 yr by Maxis

I've had RAM go bad or unstable at XMP settings on me before, but never seen a PC reboot itself as a result.  A BSOD fault is the expected mode of failure with a bad DIMM.  If there's another hardware issue in play (and I suspect there is), rebooting during a run of Memtest64 does not necessarily point to a RAM issue any more than rebooting while moving the joystick points to a joystick issue.

I think what I'd do to try and isolate this to memory, if you think that's the issue, is run the test with only one DIMM installed, and at stock SPD speed/voltage/timing rather than the factory-overclocked XMP settings.  If it reboots during the test with one DIMM, try again with the other.  If it fails during both, it's highly unlikely RAM is the issue.  If both DIMMs individually test OK during a prolonged run of Memtest86+ at stock SPD settings and there are no reboots, then test with both DIMMS installed at SPD settings, then repeat while bumping up settings incrementally towards the XMP settings.

 

 

As Bob Scott the XMP settings might be the cause. They are usually an overclock of your memory and it doesn't always work. Often these XMP settings don't take voltage into account.
I usually disable XMP altogether and test the system with default Intel/AMD settings and than adjust them until I am happy with the performance.

Although your PSU should be powerfull enough, did you connect your RTX 3090 with TWO seperate powercables, or did you use two connectors on one cable?
The latter often causes problems because a single cable isn't really designed to carry that much power.

Disable the option to automatically restart upon system failure.

Go to Control Panel and select System and Security > System > Advanced system settings > Startup and Recovery.
Choose Settings.
Uncheck the box next to Automatically restart.

This should, but not always give you an error message instead of restarting. Might help narrow down your search.

I am bored testing and trying too with similar and similar issues.

 

Yesterday I did 4 flights all failed, the second with in the two 2x16GB DDR5, with always XMP I or II, with both voltages (standard is 1.25v) at 1.26v, third at 1.27, today I am with one flight now more than an hour with one at 1.29v, fingers crossed, and still no CDT or freeze with reboot, better than yesterday....

 

Today there was an W11 update, maybe... who knows....

 

If I do not edit later, all was well in the longest flight in months without failures, sorry I could not wait to the end of the flight to write the post.

 

After some more troubleshooting:

1. XMP was already disabled.

2. My BIOS does have a memory tester and has not found any issues with memory.

3. It was only a 10 minute run (needed to use my computer and had to interrupt), but no errors returned in 3 rounds of Memtest64 when repeated

4. My gut says the problem is getting worse.  I'm only going on 5-6 times right now, so this isn't a very well established trend yet.  But for instance I tried another flight and it crashed within about 5 minutes at the gate.

5. I dropped my memory speed from 4800 MHz to 4400 MHz.  I've only been up testing once since then, but it crashed rather quick once in the sim (5ish minutes at the gate).

I have a new power supply on the way, as I don't have any spares to test.  While it seems like I have a good, adequate, and new power supply that shouldn't be failing yet, I think this is the most suspicious.  I have not seen any voltage abnormalities in any monitoring tool, but I know this isn't very reliable.

I did have my graphics card vertically mounted for a period, and it was sensitive.  It would fail to boot about 1 in 5-6 attempts.  I moved it back to a straight in PCI mount and haven't had issues since, but makes me wonder if the graphics card is a little borderline.  Under stress such as MSI Kombuster, it will run about 77-80C on the GPU, and 90C or just a little over on the memory.  Seems a bit hot to me, but that's been the status quo for as long as I can remember watching it (4-5 months since I rebuilt this PC). 

I did disable the automatic restart as suggested above (seems like a great idea), unfortunately no error messages show.  This strikes me again more as PSU, as in the past when I've had graphics card or memory issues, I'd get a BSOD rather than immediate restart. 

I have not found in BIOS yet where to simple set the RAM voltage.  There's a lot of settings, ratios, auto voltage, etc.  So I have some learning to do about those settings. 

Edited May 10, 20224 yr by ESzczesniak

2 hours ago, orchestra_nl said:

...Although your PSU should be powerfull enough, did you connect your RTX 3090 with TWO seperate powercables, or did you use two connectors on one cable?
The latter often causes problems because a single cable isn't really designed to carry that much power.

I was in fact using a single cable with 2 connectors.  I am not on 2 cables, but haven't had a chance to really test it yet.  I don't know that this means anything, but unplugging the PSU, it would typically take 10ish seconds on the BIOS screen.  This time it was to the Windows log in in 5-10 seconds total.  Most likely meaningless, but when troubleshooting, one tends to question any sign.

8 minutes ago, ESzczesniak said:

I was in fact using a single cable with 2 connectors.  I am not on 2 cables, but haven't had a chance to really test it yet.  I don't know that this means anything, but unplugging the PSU, it would typically take 10ish seconds on the BIOS screen.  This time it was to the Windows log in in 5-10 seconds total.  Most likely meaningless, but when troubleshooting, one tends to question any sign.

You need to use 2 separate cables.  Especially on a 3090 you are exceeding the power limit for a single cable and probably overloading that single output on the PSU.  Those dual pigtail cables are intended for much lower power cards.