CTD is back again

[Willie Nelson]

This CTD is back again (or didn't go away for long enough for a family man who doesn't get anywhere near enough time on DCS these days) @Flappie can you help?

These files represent the two most recent crashes only. 

I have already deleted all mods as far as I can tell and completed a repair and clean-up. 

I appreciate whatever you can do.

dcs.log dcs.log.old debrief.log

[Flappie]

Hi. Does DCS really close abruptly and you're brought back to Windows desktop?

I don't see any crash in the logs. I see no sign of memory issue, no trace of mod, no sign of overwhelmed CPU...

The last message is always about reprocessing some texture.

2024-01-21 05:21:37.377 DEBUG   LuaGUI (Main): ----onSimulationPause---
2024-01-21 05:21:37.480 WARNING BACKENDCOMMON (4896): Need to reprocess [DDS] image '/textures/F18C_1_DIFF_STAY_nm.dds'
2024-01-21 05:21:41.462 DEBUG   LuaGUI (Main): ----onSimulationResume---
2024-01-21 05:21:41.479 INFO    TACVIEW.DLL (Main): Recording flight data in [C:\Users\Aidan\Documents\Tacview\Tacview-20240121-162142-DCS-formation_trainer_by_d0ppler.zip.acmi]
2024-01-21 05:21:41.488 INFO    TACVIEW.DLL (Main): Export rights granted by the host: PlayerPlane=[1] PlayerSensors=[1] OtherObjects=[1]
2024-01-21 05:21:41.589 WARNING BACKENDCOMMON (628): Need to reprocess [DDS] image '/textures/F18C_1_DIFF_STAY_nm.dds'

2024-01-21 05:37:55.664 DEBUG   LuaGUI (Main): ---DCS onReloadEvent--- 	3
2024-01-21 05:38:07.661 WARNING LOG (564): 1 duplicate message(s) skipped.
2024-01-21 05:38:07.661 DEBUG   LuaGUI (Main): ---DCS onReloadEvent--- 	2
2024-01-21 05:38:07.763 DEBUG   Scripting (Main): make: country: 	2
2024-01-21 05:38:07.763 DEBUG   Scripting (Main): Phrases make : key = 4449
2024-01-21 05:39:20.852 WARNING BACKENDCOMMON (1928): Need to reprocess [DDS] image '/textures/kraz6322_normal.dds'

Try this:

1. Browse to "Saved Games/DCS..."
2. Delete both "fxo" and "metashaders2" folders. They will be remade during the next DCS run.
3. Browse to "C:\Users\your_login\AppData\Local\NVIDIA\DXCache" and delete its content.

 

[Willie Nelson]

Thanks for getting back to me so quickly

I have tried delting the fxo and metashaders2 previoulsy as I am aware that sometimes that helps, I did it again this morning as you reqested. I dont have a DXCache in that location just a GL Cache which I tried deleting but one of the folders was in use at the time (DCS was closed but the Pimax client was open, so I am guessing it was that...?) 

I have attached a screenshot of what else other apps that are in use for some context and some of my NVIDIA settings, if it makes any difference. I am also using Joytokey for Apache Xbox controller mapping. 

Still crashing after clearing all of your suggested folders were cleared (that could be cleared) here's this moring's crash logs.

I appreciate you looking in to it.

dcs.log dcs.log.old

[Flappie]

Some things you can try:

1. Browse to "Saved Games/DCS...", then rename the "Scripts" folder to "Scripts.bak" (it will disable TacView,and any other export script).
2. Not using the OpenXR Toolkit with DCS.

[Willie Nelson]

7 hours ago, Flappie said:

Some things you can try:

1. Browse to "Saved Games/DCS...", then rename the "Scripts" folder to "Scripts.bak" (it will disable TacView,and any other export script).
2. Not using the OpenXR Toolkit with DCS.

Thanks for that, just disabling OpenXR Toolkit for the time being to see one change at a time, So far, three successful flights without crash, I'll see if that causal and report back, thanks again.

[Willie Nelson]

So I have now done both of those suggestions and while I have had some seesions, perhaps four maybe, without a crash, I have also had two to three that have continued to crash.

Here's the latest log.

dcs.log

[Flappie]

It's still the same kind of crash: no trace of it in DCS log. Does your PC reboot or show a blue screen of death when that happens?

2024-01-28 02:55:14.361 DEBUG   Scripting (Main): make: country: 	80
2024-01-28 02:55:14.361 DEBUG   Scripting (Main): Phrases make : key = 5
2024-01-28 02:55:14.362 DEBUG   Scripting (Main): Phrases make : key = 4279
2024-01-28 02:55:17.115 DEBUG   LuaGUI (Main): ---onBdaShowF10()--- bValue , isOn(): 	false	true
2024-01-28 02:55:32.593 INFO    ASYNCNET (Main): Current ping: 293.4ms
2024-01-28 02:55:32.844 WARNING WORLD (Main): ModelTimeQuantizer: SAME MODEL TIME
2024-01-28 02:55:34.041 DEBUG   LuaGUI (Main): --- onChatMessage---	BLUE player _\399/_Allfornone died	nil	nil
2024-01-28 02:55:34.041 DEBUG   LuaGUI (Main): --- onChatMessage---	RED unknown in MiG-29S killed BLUE player _\399/_Allfornone in F-16CM bl.50 with R-77 (AA-12 Adder)	nil	nil
2024-01-28 02:55:40.429 DEBUG   LuaGUI (Main): ---onBdaShowF10()--- bValue , isOn(): 	true	true
2024-01-28 02:56:02.596 INFO    ASYNCNET (Main): Current ping: 315.0ms
2024-01-28 02:56:04.039 DEBUG   LuaGUI (Main): --- onChatMessage---	RED unknown in SAM Chaparral M48 killed BLUE player Noob 6-9 in F-16CM bl.50 with MIM-72G	nil	nil

 

 

[Willie Nelson]

Both have in fact happened but it almost always just a garden variety CTD. 

[Flappie]

Then maybe you should check:

• Your PC temperatures
• Your RAM, with the help of MemTest86:

A third possibility is a defunct PSU but that is a difficult thing to check.

[Willie Nelson]

Thanks again for the response. I have a 4090 and a 14900KF with a high quality air cooler although I am told that these will often go up to 100 degrees which makes it hard to determine whether that's the issue becuase as you can see, I get there on several cores at least for a time (I am not if there's a more detailed log as there was nothing in the folder depsite me switching on the Core Temp log). This sceenshot shows a MP seesion on an empty 4YA server that I played for perhaps half an hour without fault.

I have followed the instructions for the Freeware Memtest86 and cannot for the life of me get it to boot from the Freeware file saved to a USB stick, I have checked that I am using UEFI BIOS mode and that secure boot if off. When I attempt to run it from my F12 boot menu list it keeps coming back to the same F12 boot menu list within a few seconds. Any ideas? I did run a Windows 11 inbuilt memory test but apparently that's checking several different things than Memtest86. 

The PSU is a definitive possibility, it is also new and I sometimes have to switch it on by recycling the mains power point switch! I know that's not normal in itself but once it's on it does stay on and my time limit constraints mean that if I take it back to the store I am likely going to run in to the problem where they cannot get it to crash while bench testing it because theyre not using DCS in VR with a Pimax Crystal. The stress tests that I have previouly run don't seem to cause it any promblems either. Not sure how useful a PSU tester would be in this case but I'm open to any of your suggestions. 

[Flappie]

I see your CPU is overclocked @5.7 GHz which is a lot, and that some of its cores reach 100°C.

You should first ensure that your motherboard does not cut the power when your CPU goes over 100°C, for it's often the case by default. Find the setting in your BIOS and see if you can set a little bit higher (guys at TechPowerUp had to set it to 115°C)

If that doesn't help, or if you've done that already, then consider lowering your CPU frequency, or even using stock frequency, since it seems OC does not change performance much according to this chart.

Last but not least, what is the power delivered by your PSU?

[Willie Nelson]

I was fairly confident that in fact I had not overclocked this CPU! I see what you’re looking at, is that not the native setting potentially being displayed? 
 

I am informed by somebody that know more that me that there’s not a lot to be gained from OC ing a 14series Intel.

With regards to the Memtest 86 do you have any guesses what I might be missing there with my attempts to run the USB image? I’m a bit concerned about setting anything to 115 degrees Celsius. :- 0

[Flappie]

Maybe you have not overclocked your CPU. It seems the turbo mode can make it go up to 6GHz.

I'm reading that even without OC, it consumes a lot of watts., which explains the high temperature. And you have a 4090, which needs a lot of power too. We need to ensure your PSU is enough for them both. What is your PSU model?

[Willie Nelson]

It's a Corsair RM1000e ATX variant. It is purpose built to accommodate the 4090 three conncetion requirement. in theory, it should be fine but there may well be some fault in it causing stability issues. I would like to rule out other lower hanging fruit first possibly , particularly now that I am at home for a few days with manflu. 

Again, do you have any theories as to why I might not be able to get the Memtext 86 to load from the USB that I have installed it on. It just keeps coming back to the boot menu without running despite me clearly following the installation and loading instructions.

Edited February 4, 2024 by Willie Nelson

[Flappie]

OK, let's rule out the PSU for now. Corsair is a good brand, and 1000W is enough for what you have.

Make sure your BIOS is allowed to boot on USB devices.

[Willie Nelson]

17 hours ago, Flappie said:

OK, let's rule out the PSU for now. Corsair is a good brand, and 1000W is enough for what you have.

Make sure your BIOS is allowed to boot on USB devices.

In confirming the BIOS allows me to boot the USB "image" I have UEFI BIOS in my MS Config settings and the following images from my BIOS might show you a little more. When I select USB from the boot menu, the screen momentarily goes black, then instantly comes back to the same boot menu that you can see in the first image below.

Withing the FAT USB there is a PDF document with instrucitons and an 'SRC" folder and a "SysLinux folder'

[Willie Nelson]

This is perhaps an aside but what's the rationale of dcs.log versus dcs.log.old?

[Flappie]

dcs.log is from you latest (or current) DCS session.

dcs.log.old is from the session before.

If you haven't tried it yet, maybe you can try and update your motherboard BIOS.

You should open a support ticket. Log into DCS website, click "Support", then click "Create a new ticket". Add a link to the present thread.

Edited February 5, 2024 by Flappie

[Willie Nelson]

I have done as you suggested and created a ticket with a link to here. The BIOS is updated to October 23 as F28a for the AORUS Z690Pro

[rfxcasey]

I'm not familiar with Gigabyte's particular flavor of UEFI/BIOS but I'll tell you from someone with years of experience, these days many motherboard manufacturers have their boards set way to aggressively set as default and they often need quite a bit of somewhat involved tweaking to get temps and power consumption within acceptable range. First thing I'd do is check to see if there is somewhere in the UEFI/BIOS to set your CPU cooler type to "Boxed" or "Air" which will limit the power applied to your CPU. Usually these boards are set for liquid cooling power profiles by default, but even after making the change to "Air" or whatever it is, there will still be some advanced core power settings you'll need to research and tweak. My MSI had all cores' turbo maxed by default and a far to aggressive dynamic voltage profile. Hope that helps, I really don't have time to do an in depth analysis if that's what you're looking for, but it's probably what you need. It might also help to verify the CPU cooler is installed correctly with the proper placement of thermal compound just to be sure. Did you build this machine yourself? 14900KF is most likely overkill for the average user and will only result in higher temps and power consumption vs something like the 14700K which will most undoubtedly give you the same FPS in DCS and every other game for that matter give or take 2 or 3 FPS. That is historically how it goes with these things. People often go for the flagship CPUs believing it will be better for gaming or daily use when in reality its completely unnecessary for anything other than professional full time video encoding, compiling or perhaps CAD applications. 99 percent of software won't even utilize the difference.

But I digress, make sure your turbo mode is set to "per core" or "turbo core ratio" and not "all core". Your mileage may vary.

Edited February 6, 2024 by rfxcasey

[Willie Nelson]

Thanks for getting back to me so quickly with some suggestions. I cannot see anything to that effect (boxed or air) in the BIOS menu and I know that the Noctua NH-D15 has been tested and shown to be just as cool as many AIO cooler add ons when fitted correctly of course, which it certainly seems to be. 

Having said that, I am happy to try anything. In that vein, here is the BIOS menu options shown under the BIOS "Tweaker" menu. 

I have also attached a picture of the paste that I had on the CPU, I have wiped off the old paste (that was in this picture which shows it was even at least) and reaplied a slightly smaller amount. I then reconnected everything and I turned it all on, ran another session and it crashed again with Core Temp showing that the highest temp had peaked at 96 degrees C. Those temps are representative of what appeared in the crash log attached. 

I show all this becasue I am not sure which settings Turbo settings you're asking me to change. Perhaps this will help you to determine which way to go.

Is there any other app that I could perhaps run on the side to see whether it is indeed the temperatures that are causing DCS to wig out. "CoreTemp" shows me high temps (which at least theoretically should be fine) but anything that might show the cause of the CTD if it is not coming up in the DCS log perhaps.

dcs.log dcs.log.old

Edited February 6, 2024 by Willie Nelson
Further info

[rfxcasey]

Active Turbo Ratios set to auto is one potential issue. As said, I'm not familiar with Gigabytes Bios, the best suggestion I can give you is to head over to the Overclockers.com forum and see if anyone will point you in the direction of a tweaking guide for your particular or similar setup. Those guys over there are intimately acquainted with current advanced EUFI/BIOS settings. There are other overclocking forums as well you can check out, not that you're wanting to overclock, but overclockers are the ones who really dig into the details of system tweaking. Google is also your friend. Try Googling a setup guide or temp issues for your particular CPU and MOBO.

Edited February 7, 2024 by rfxcasey

[rfxcasey]

As said, I really don't have the time (or desire) to do an in depth analysis with you. If I had a dollar for ever minute I've spent helping folks with PC issues I'd be retired already. I do this kind of stuff for a living so in my free time I only deal with my own issues when they come up anymore. Here, read this thread it might get you going in the right direction. https://www.techpowerup.com/forums/threads/14900k-tuned-for-efficiency-gaming-power-draw.315783/

 

Edited February 7, 2024 by rfxcasey

[Willie Nelson]

So after much grinding and gnashing of teeth, a friend of mine was able to look into this for me and provided me what I needed to ultimately fix the problem.

Enforce a 253 Watt power limit and disable multi-core enhancement. He did some searching around and said that some of the 14900 series is susceptible to a low I2 voltage whci can be tweaked at some other point. 

[Hawkeye60]

After reading this,

I wonder how many other MT issues are being caused by low voltage?

It is something most will never check,

And it is a common issue with Power supplies.

Heat and age will kill them.

With the new demanding specs DCS requires, a good PS is needed more than ever.

If you have strange issues with your PC,

Random BSD, crash to desk top, reboots,

And the PS is old, or low wattage,

Up grade. 750 to 1000 Watts ,

It is a minimum size IMHO

I had one go bad once.

Destroyed My MB!

I bought a top of the line one after that event.