Possible new VAS issue in v3.4

Any issues, problems or troubleshooting topics related to the Prepar3D client application.
User avatar
aeronauta
Posts: 1661
Joined: Sun Oct 23, 2011 1:56 am
Location: Melbourne, Australia

Re: Possible new VAS issue in v3.4

Post by aeronauta »

possibly Genoa aerosoft?
Jorge
Jorge Roldan
OZx Developer (aero) - Hovercontrol Helicopter Certified Pilot - Spad.Next Beta Tester - VATPAC Pilot PR
System i9-9990 4.7Ghz 32GB W11 Nvidia GeForce RTX 2080 8GB
Kosta
Posts: 1173
Joined: Sun Sep 18, 2011 10:20 pm

Re: Possible new VAS issue in v3.4

Post by Kosta »

Don't have any addons in vicinity.
Humpix
Posts: 22
Joined: Fri Feb 03, 2017 2:22 pm

Re: Possible new VAS issue in v3.4

Post by Humpix »

Nope, no other Addon installed.
First I thought that the mesh maby is the problem, but meanwhile I have an answer from FSP. FSP think it has to do something with how Prepar3d reads out, or handle the mesh. And the fact that freemesh have the same problem, it talks for FSP... Btw, I does not have this effect with the older versions of Prepar3D, so this also talks for no problems with the mesh itselfs.

Now, the real problem is not this single spot, but if you fly long ways u hit several of them. Causing OOM and make some routes impossible.

So first workaround for me is - dont use addon mesh!
Second hope is, that LM look at this and maybe find some fix... :)
Kosta
Posts: 1173
Joined: Sun Sep 18, 2011 10:20 pm

Re: Possible new VAS issue in v3.4

Post by Kosta »

OK, makes sense, I also think LM should look into it.
ScottecLEMD
Posts: 5
Joined: Mon Aug 11, 2014 8:03 am

Re: Possible new VAS issue in v3.4

Post by ScottecLEMD »

I´ve just updated to V 3.4.22.19868 (hotfix 3) and I´ve had my first OOM in months.

I departed a default airfield in Spain, LEZG, with Orbx Global Landclass and Vector running as well as AS16 (which I´ve had running without issue on previous version) I started at LEZG in the PMDG 737-800 and my remaining VAS was around 250mb - astounded. I departed but soon had to give up as the VAS decreased rapidly. I only upgraded to this version to solve a problem with the cycling of views causing a CTD. Frustrated!

Rolling back to 3.3 tomorrow.

Rgds

Ken
Kosta
Posts: 1173
Joined: Sun Sep 18, 2011 10:20 pm

Re: Possible new VAS issue in v3.4

Post by Kosta »

Weird. I've been quite happy with HF2, although I rarely do flights over 3-4hrs. Those are my longest, mostly ranging from 1-2 hours. Family-man nowdays...
dikkeduif
Posts: 9
Joined: Tue Jun 10, 2014 6:45 pm

Re: Possible new VAS issue in v3.4

Post by dikkeduif »

I have been trying to complete a flight from EBBR (mega airport v2) to EGLL in the pmdg 737 with no success on 3.4 hotfix 3. It always runs out of memory when I'm within 20-30nm from EGLL. The FPS isn't too bad really, but suddenly the VAS drops dramatically from 500000 to all of a sudden nothing (I'm using fsuipc vas logging). I have the following addons installed

* EBBR mega airport v2
* FS Pilot Global 2010
* FTX Global base pack
* FTX Global openLC Europe
* FTX Global Trees HD
* FTX Global Vector
* Active sky next
* Chaseplane

I tried lowering the sliders, that was actually the first thing I did, but that didn't help much. It crashed when passing EBOS (Ostend) somewhere in the middle of the north sea.

I read here on this forum that maybe FSP mesh might be the problem so I turned off the Europe mesh and local meshes in the library, to no avail, although I did get near EGLL this time, within 20 or 30 nm of EGLL I think.

The other thing I did was turn off shader cache in my nvidia control panel (I have a 970gtx), which also didn't solve much. And just now I installed 376.33 to see if that solves anything. It still crashes with a huge spike when I'm getting near 20-30NM.

Now I have to say I did do this route in the past, on older versions of p3d and I never had any issues. Heck I can't even remember when p3d crashed on me because of OOM issues.

Also this is a fresh new install on a ssd. The only thing I want to try now is to try again and lowering more sliders to see if that helps.
zap737
Posts: 216
Joined: Wed Apr 29, 2015 3:27 am

Re: Possible new VAS issue in v3.4

Post by zap737 »

dikkeduif and ScottecLEMD, are you on Nvidia drivers 378.49?
There is an issue with those drivers and VAS with the PMDG 747 V3.
See here: http://www.avsim.com/topic/503393-747-v ... s-not-747/
The solution is to try Nvidia drivers 376.33, it may help with the PMDG 737.....
Asus Z370-A | 8700K @ 5.0 Ghz | 32GB DDR 4 | GTX 1080TI | 4K Monitor @ 30Hz | Samsung 970 EVO | Track IR 5 | Windows 10 Pro x64 | Virpil Mongoose Stick | TM Warthog Throttle | MFG X-Wind Rudder Pedals | Orbx Global | Orbx Vector | Orbx OpenLC NA, EU
User avatar
Martyson
Posts: 15188
Joined: Sun Mar 25, 2012 11:08 am

Re: Possible new VAS issue in v3.4

Post by Martyson »

Just for info:

Nvidia drivers 378.49 installed 1-24-17 all ok here so far with Win7-64 , P3D HF3, Orbx Global base / N.America / PNW / a few Orbx airports , all my payware aircraft including PMDG737NG (FSX and P3D).

No payware Euro airports.

I will keep testing.
Best Regards,
Vaughan Martell PP-ASEL (KDTW)
Kosta
Posts: 1173
Joined: Sun Sep 18, 2011 10:20 pm

Re: Possible new VAS issue in v3.4

Post by Kosta »

dikkeduif wrote: Mon Feb 06, 2017 12:20 pm I have been trying to complete a flight from EBBR (mega airport v2) to EGLL in the pmdg 737 with no success on 3.4 hotfix 3. It always runs out of memory when I'm within 20-30nm from EGLL. The FPS isn't too bad really, but suddenly the VAS drops dramatically from 500000 to all of a sudden nothing (I'm using fsuipc vas logging). I have the following addons installed

* EBBR mega airport v2
* FS Pilot Global 2010
* FTX Global base pack
* FTX Global openLC Europe
* FTX Global Trees HD
* FTX Global Vector
* Active sky next
* Chaseplane

I tried lowering the sliders, that was actually the first thing I did, but that didn't help much. It crashed when passing EBOS (Ostend) somewhere in the middle of the north sea.

I read here on this forum that maybe FSP mesh might be the problem so I turned off the Europe mesh and local meshes in the library, to no avail, although I did get near EGLL this time, within 20 or 30 nm of EGLL I think.

The other thing I did was turn off shader cache in my nvidia control panel (I have a 970gtx), which also didn't solve much. And just now I installed 376.33 to see if that solves anything. It still crashes with a huge spike when I'm getting near 20-30NM.

Now I have to say I did do this route in the past, on older versions of p3d and I never had any issues. Heck I can't even remember when p3d crashed on me because of OOM issues.

Also this is a fresh new install on a ssd. The only thing I want to try now is to try again and lowering more sliders to see if that helps.
Do you have PFPX and can send me the flightplan? I'll be happy to fly it quickly, it ain't a long route.
dikkeduif
Posts: 9
Joined: Tue Jun 10, 2014 6:45 pm

Re: Possible new VAS issue in v3.4

Post by dikkeduif »

Hi Zap, I did install the 376.33 drivers but it also crashed for me.

Kosta I'm not entirely sure how I can save the flight plan in pfpx, but I exported it for the NGX in pfpx and for p3d. I also managed to save it as a txt file. Let me know if you want another format and I'd be happy to export it for you. Much appreciated!

The txt file from pfpx
https://dl.dropboxusercontent.com/u/497 ... 20%231.txt

The NGX file to load into the FMC
https://dl.dropboxusercontent.com/u/497 ... EGLL01.rte

The prepar3d flight plan
https://dl.dropboxusercontent.com/u/497 ... EGLL01.pln
Kosta
Posts: 1173
Joined: Sun Sep 18, 2011 10:20 pm

Re: Possible new VAS issue in v3.4

Post by Kosta »

Quite fine, thank you. I'll give it a go...
Humpix
Posts: 22
Joined: Fri Feb 03, 2017 2:22 pm

Re: Possible new VAS issue in v3.4

Post by Humpix »

For me, it did not help to just turn off the europe FSP 2010 mesh. I have to turn off ALL of them. Even if I just fly in europe, what is very very strange.
Maybe there is some wrong data in other regions, which leads to a double load at some spots, with can explain the sudden VAS increase. But that is highly speculative!!!. Because the meshes always woks perfectly in earlier versions of Prepar3D (up to 3.2, the mess starts with 3.3). Or they are to detailed and the VAS logic can not handle them? I don´t know...

In any case, after turning off all of the mesh entries in the scenery library, I have no problems with the VAS. No sudden increases.
I did an 5h flight, Europe to Dubai, starting with 2600MB and end up at the Gate with 3100MB. (v3.4 HF2)
Also my VAS-test-routes work without problems.

Kosta, after that, overbearing as I am, I tryed the same 5h flight with the freemeshx - and gues what - also no problems... Also not from LOWI to LEPA (our test route).
Im not sure what happens, because you get a 100MB+ increase at similar conditions...


But! What I also find out: If the LOD is on max, the VAS goes up and up and up... (slowly, but always up)
With or without mesh. With LOD just one step to the left, the VAS did not increase very much.

So, at the moment:
- LOD one step back
- Scenery complexity maxed out
- Autogen 2 steps from max
- water one step from max
- all other sliders (mesh, tesselation, 7cm per pixel, 4k textures, ...) maxed out
- freemeshx
- All ORBX Stuff, AustriaHD, tons of airports, AS2016, REX HD Airports, EZdok2, all active without problems
- NVidia drivers 378.49 (but I think this does not matter at all if you do not have the 747v3)
- 4k resolution and about 40-50fps and great visuals - amazing visuals!
- Fun as hell (ok, I know, P3D is not for fun :) - Im sorry for having fun!

Next days I get the FSP ultimate next gen. meshes. I will report how they affect the VAS.

Chears!
Kosta
Posts: 1173
Joined: Sun Sep 18, 2011 10:20 pm

Re: Possible new VAS issue in v3.4

Post by Kosta »

I did a just a directed flight in an NGX today going 180degrees down from EKBI.
I flew this route the first time, however...

I know this is something that never happened: I started with no sceneries, only FreeMeshX, ORBX Global, UTXv2 EU und AS16.
Free VAS at takeoff: 1400mb
Free VAS after about 2 hours: 800mb

Next time, I'll do VAS-logging, so that I can see if this is gradual or it happens at one place.

I know that something like that doesn't happen on my system usually, at least it never happened before this whole story began. So there must be something off. LM have done some VAS management stuff as of late, so they possibly broke something... I don't know. Speculating here.

But it's definitely time to start some serious testing and logging.
Kosta
Posts: 1173
Joined: Sun Sep 18, 2011 10:20 pm

Re: Possible new VAS issue in v3.4

Post by Kosta »

Because the meshes always woks perfectly in earlier versions of Prepar3D (up to 3.2, the mess starts with 3.3). Or they are to detailed and the VAS logic can not handle them? I don´t know...
This is one of my thoughts... this needs some extensive testing...
Kosta, after that, overbearing as I am, I tryed the same 5h flight with the freemeshx - and gues what - also no problems... Also not from LOWI to LEPA (our test route).
I have a feeling that tests need to be repeated to get a consistent result. It seems to me that not every test is repeatable.
But! What I also find out: If the LOD is on max, the VAS goes up and up and up... (slowly, but always up)
With or without mesh. With LOD just one step to the left, the VAS did not increase very much.
This is one of the things I don't want to do. Anything less than 6.5 looks very blurry to me, especially landing in Innsbruck or other mountainous areas.
Besides, it used to work fine!
Locked