23 minutes ago, pvupilot said:

I'm having this same problem. I created the other thread topic referenced in here, so I'll just continue in here with you guys. I ran the DDU in safe mode, updated to the latest driver, and still the same crash. GPU (GTX 1080) usage was only 72%, GPU Memory 81%, and voltage 72%. Nothing anymore than it's been for all of my other time in the sim. I already have TdrDelay=8 too. I'm really stumped on this as I never had this before 4.4.

Have you tried the only fix that worked for me and using MSI Afterburner of EVGA Precision to reduce Power Limit setting to 80%

 

6 minutes ago, w6kd said:

There's another TDR setting you can try in addition to TdrDelay, which is TdrDdiDelay (a DWORD value created the same way and in the same registry location as TdrDelay).  The default is 5, I have been using 20 for years, along with 10 for TdrDelay.  The TdrDdiDelay sets the amount of time a thread can leave the driver.  Note that if you set TdrLevel to 0, none of these delay settings are used--they're meaningless--as the detection of driver errors by the TDR code is disabled.

Also, for those using factory-overclocked GPUs, a factory overclock IS still an overclock.  I don't know what sort of binning/testing process is used by the various OEMs, but several years ago I had a couple very high-end 980Ti water-cooled boards in an SLI configuration that would not run stable at their aggressive advertised factory overclock, but they worked like champs with a relatively modest downclock from the factory settings (still overclocked significantly above the factory settings).  So it would not be at all surprising to me that some folks might run into stability issues at the factory settings on some of the factory-overclocked boards.  And overclocked w/r/t GPUs usually means both the core clocks and the memory clock, which are set independently.  So if you're having trouble, I'd start with both core and memory clocks at nVidia stock settings and work up from there.

Regards

No TDR edits of any sort worked for me. I tried both the level and delay variants in both 32 and 64 bit entries. Tried 0, also 8 and even seen a post claim to try 60. Most did nothing or completely locked up the pc forcing me having to reboot.

My 1080ti is factory overclocked and one of the earliest things I did was to down clock both the core and memory to a big standard 1080ti speeds..didn't help at all

Chris

Edited December 4, 20187 yr by cj-ibbotson

Hi all:

The dreaded DXGI error has struck me. Just updated to a 1080ti. Tried the registry key with no luck. However, FWIW, on my last flight I was running GPU-Z in logging mode. TBH, I don't know if these numbers will help diagnose the problem, but thought I would contribute something. Have posted it in my dropbox account below.

Click me to see the GPU-Z error log at time of error

4 hours ago, Rob Ainscough said:

Sorry, I thought you were running 2 GPUs not 3?

No, always 3 on this rig since July 2017. It was clear in my earlier exchanges with you, I think, but easily forgotten I expect.

4 hours ago, Rob Ainscough said:

3 GPUs on the X series MBs will work fine so long as you dedicate them to each display

That was the intention, but we found that the performance was noticeably poorer than with everything on one GPU. So then we tried SLI and that was just the same, but smoother.

4 hours ago, Rob Ainscough said:

But in your case I think you have distortion issues that need to be corrected by external software/hardware for your projectors?

It's a version of Immersive Pro. It deals with the otherwise curved image (shame we can't switch the "flat screen" distortion added by Projectors off! ;-)), and merges delibrately ovelapped images to ensure smooth transitions.

But performance is identical without that process running.

4 hours ago, Rob Ainscough said:

Per my testing, SLI adds CPU overheard, as much as 17%.  SLI is great IF (important IF) you aren't capping out your CPU.  If you're capping out your CPU then SLI is probably going to make matters worse, not better.

Well, as I said, the performance as measured by FPS was identical with or without SLI. It was smoothness which was better with SLI. and, yes, with this multi-scenery screen setup core 0 has always hit 100% with my test scenario.

4 hours ago, Rob Ainscough said:

those are going to stress the CPU and SLI will just stress the CPU that much more and FPS will actually be worse.

Not here, sorry.

Since disabling SLI my FPS have been the same, (not checked smoothness because not had time to fly), but the single operating GPU is now at 90-95% usage in my test scenario. I've not checked the temperature. That's instead of up to 45% on each of 2 out of 3 with the 3-way SLI.

Will try to configure 2-way SLI tomorrow. 90-95% is worrying. 

Pete

 

16 minutes ago, Pete Dowson said:

and, yes, with this multi-scenery screen setup core 0 has always hit 100% with my test scenario

I think you missed my point, if your CPU is at 100% 1,2,3 GPUs will not make any difference.

16 minutes ago, Pete Dowson said:

It was clear in my earlier exchanges with you, I think, but easily forgotten I expect.

Very likely, your information unfortunately stayed in volatile memory and got lost when I turned in for the night. 🙂

1 hour ago, pgde said:

The dreaded DXGI error has struck me. Just updated to a 1080ti. Tried the registry key with no luck. However, FWIW, on my last flight I was running GPU-Z in logging mode.

I took a look at your data and looks to me your card went south (memory corruption) a few milliseconds before the DXGI error point:

Sudden flush of memory usage and drastic frequency drop ... 

Cheers, Rob.

7 hours ago, Rob Ainscough said:

I think you missed my point, if your CPU is at 100% 1,2,3 GPUs will not make any difference.

1. It is not at 100% all the time. That is just in my test scenerios, heavy airport plus traffic.

2. SLI did make it run smoother.

3. Without SLI enabled I am now seeing 90-95% usage (according to Task Manager) on the one GPU in the test scenario, rather than 2 at 45% and one around 5% max. I find that a concern, and haven't even dared look at the temperature!

7 hours ago, Rob Ainscough said:

I took a look at your data and looks to me your card went south (memory corruption) a few milliseconds before the DXGI error point:

That data seems to date back to April. Is that when I got my first DXGI error? I don't recognise the data. Did i send it to you? What's it from?

More to the point, can you identify which card it is? 1st slot, 2nd or 3rd? The 2 which run at 45% might be 1st and 3rd if the order of GPUs in the Task Manager display is anything to go by. Didn't I read somewhere that those are the main slots in any case (and used for 2-way SLI) with the second slot only being for 3-way?

Or do I have to do trial and error?

Thanks,
Pete

 

Edited December 5, 20187 yr by Pete Dowson

3 hours ago, Pete Dowson said:

That data seems to date back to April. Is that when I got my first DXGI error? I don't recognise the data. Did i send it to you? What's it from?

Hi Pete,

I think we are seeing the American date format. I’m assuming these figures were recorded on 4th December, I.e. yesterday. Does that make more sense?

Regards,

Mike

3 hours ago, Pete Dowson said:

That data seems to date back to April. Is that when I got my first DXGI error? I don't recognise the data. Did i send it to you? What's it from?

Rob was responding to someone else

15 minutes ago, pracines said:

Rob was responding to someone else

Ah, so he was. Whoops! The first two parts were for me!

Pete

 

So Rob, does the memory corruption mean that the card is defective? Still under warranty so I can RMA it. I never had this problem with my 1070....

Thanks for your analysis.....

P. (the other Peter 😁)

33 minutes ago, pgde said:

So Rob, does the memory corruption mean that the card is defective? Still under warranty so I can RMA it. I never had this problem with my 1070....

Thanks for your analysis.....

P. (the other Peter 😁)

Sadly my own card is just outside the warranty (if it was just 12 months) I've had crashes since June or July but was convinced my card wasn't faulty as so many experience these crashes. I also did not want to be without a GPU for a month waiting for tests etc.

I might review running a log like you did and post results. Only thing is I can't sit in front of pc for hrs whilst it's testing so unsure at which point the log is applicable as I wouldn't know the exact time of the crash unless it happens whilst I'm present. Any tips?

Chris

Edited December 5, 20187 yr by cj-ibbotson

common issue this, mentioned on here countless times, mentioned and and deflected many times on the LM forumn as well.

Try 397.64 drivers that should eliminate the issue.  It has with me. Ive said a million times youll never get a huge multinational company like LM admitting theres an issue.

On a separate note I was watching matt davies on youtube the other night in work flying into singapore in some bad weather and the sim threw him into the deck.. 20 years on we still have wind shift issues and sillyness with p3d/fsx activesky which has nevber been fixed not in 20 years.

but as much as i found the video very funny ,  listening to the other dude (no idea who is he)  talking about airline operations and diversion airfields he clearly was absolutely clueless  about what happens with commercial and bad weather alternates  , I was glad its not just me that get throw to the ground in a little bit of turb.  

Same with this dxgi error in a way,  youll never ever ever ever ever ever ever get them to admit their software might be at fault, same as at my airline youll never get the bosses to admit the IFE is gash on a certain fleet,  too much liability and brand damage if they admit it.

 

 

Edited December 5, 20187 yr by tooting

29 minutes ago, tooting said:

...Ive said a million times youll never get a huge multinational company like LM admitting theres an issue...

Same with this dxgi error in a way,  youll never ever ever ever ever ever ever get them to admit their software might be at fault, same as at my airline youll never get the bosses to admit the IFE is gash on a certain fleet,  too much liability and brand damage if they admit it.

Respectfully, this is gibberish to me.  LM never developed Microsoft ESP or FSX for which Prepar3D is based on.  They brought on a team of experts who were developing Microsoft ESP, a program based off of the FSX engine.  We should be happy LM decided to bring the team aboard to further develop this wonderful product for the flight simulation community!!  I just cannot see how this error caused this particular crash. 

Well, since the DXGI errors occurred with my 3-way SLI (which is apparently not supported by P3D in any case), but not with SLI disabled (but with the horrible result that I was getting 95% GPU usage -- not much headroom!), I tried 2-way SLI.

Since the 3-way usge for the 3 cards was maxing at 45% + 5% + 45% I disabled the middle one in Device Manager (to save fiddling about with the connectors for now).

After mny hours of testing I can attest to the fact that this eliminated the DXGI crash, even at night which was where I could reproduce it within minutes 100% of the time.

The loading was maxing at a healthier 65% + 65%, but more usually around 50% each. Nice sharing of the workload.

My next step (tomorrow) wil be to re-enable that GPU but disable the third one. This might not work (not sure how fussy SLI connections are), but if it does and gives the same good result, then I'll know the video card is okay. If not then it is definitely suspect.

Either way, one of them is coming out.

Pete

 

i dont get this error with any other software I use.  only P3D.  and if I use ANY driver later than 397.64  I get the issue.  But ANY driver before 397.64 I dont get it.

Please explain how this cant be an issue with P3D then. 

4 hours ago, pgde said:

So Rob, does the memory corruption mean that the card is defective?

There is obviously an issue there, I can't see software doing that ... BUT, you might want to hit up nVidia engineer and present them with your data an see what they suggest.

34 minutes ago, tooting said:

Same with this dxgi error in a way,  youll never ever ever ever ever ever ever get them to admit their software might be at fault

Who's the mysterious "them"?  Jumping to conclusions and point fingers doesn't help anyone and that approach never solves problems?

A quick google search with "dxgi error device hung" and you'll see the problem is common across all kinds of games/sims/etc., the solution in most cases is to reduce GPU/Memory frequency or return the GPU under warranty (RMA) or replace the GPU with another one.  

Cheers, Rob.