I'm not sure why you're so astonished; how else are you supposed to know exactly how much VAS the add-on you're developing is using?

 

Why am I astonished? I get your point, what my point is why didn't they test over add-on scenery as I indicated in my post?

Doess PMDG actually think we're all going to be flying the 747 over vanilla sims?

And yes, I know there's almost an infinite amount of scenery add-on configurations out there but to test with just vanilla doesn't reflect most people's scenery setups.

Why am I astonished? I get your point, what my point is why didn't they test over add-on scenery as I indicated in my post?

Doess PMDG actually think we're all going to be flying the 747 over vanilla sims?

And yes, I know there's almost an infinite amount of scenery add-on configurations out there but to test with just vanilla doesn't reflect most people's scenery setups.

 

Jack, I think you misunderstand the original post. The VAS "programme" used this methodology but it's highly likely the aircraft was tested in a multitude of real world (virtual) environs. It's not as black and white as you presume.

 

All the best and enjoy!

Jay

The PMDG FMC memory leak is/has been an issue for quite a long time on the 777 and now looks like it is present in the 747. PMDG a long time ago made some fixes that cured this issue at KLAX but many more airports still have it. KSFO, CYVR, KSEA, VHHH, Lihue's single ILS approach and all the OCK tranistions into Heathrow suffer from this. It's the missed approach data in the navdata, irregardless if from Navdata or Aerosoft. You plug your arrival rwy in and it has that certain missed approach data in it and you will get a memory leak. Something to do with the "intercept" commands. Very long thread in the 777 forum about it. PMDG will need to issue the fix. A workaround is to go into those offending airports and delete the missed approach data. I have a separate folder with the modified airport files that I copy and paste back into each new AIRAC cycle.

Why am I astonished? I get your point, what my point is why didn't they test over add-on scenery as I indicated in my post?

Doess PMDG actually think we're all going to be flying the 747 over vanilla sims?

And yes, I know there's almost an infinite amount of scenery add-on configurations out there but to test with just vanilla doesn't reflect most people's scenery setups.

 

Why is it a developers job to balance everyone's machines? If a dev tests their aircraft on a vanilla build, and they see they still have 1.8-2GB VAS still available, then that means they leave the end-user a good chunk of VAS to add in their own sceneries. It's down to a user to balance their own machine. My question to the general flight sim populace who seem to have endless OOM issues is: why is it ok for a scenery to take up to 500MB VAS for a single airport yet when an aircraft takes 750MB the world falls in? Why is it that people value a $35 scenery (which you'll see for a few minutes each flight) over a $135 aircraft (which you're in for the entirety of your sim session)?

 

This is all academic anyway - as Jason said above, I'm fairly sure a wide variety of installs are used in testing. In fact, Kyle Rodgers mentioned he had a number of setups with varying add-ons.

Gentlemen,

 

Now that i have a concrete report to come back on with MY issue since this was MY thread that I started, everyone can step down off their high horse lol...

 

Coming out of KDEN (Flightbeam) going to KSDF (Blueprint) in N581UP i am at cruise and i havent gone above 69% 2800MB...currently holding at 68% 2770MB

 

What was my fix?... a legacy install of the previous Nvidia Driver 376.33!!!

 

Thank you so much to all who suggested amazing troubleshooting techniques...The Queen is restored to her full glory!!!

 

Now this doesn't necessarily mean this is the driver you need to use, this is the driver my system and aircraft like!!!

It may vary based on setup!!

 

Rich M. 

I found that turning off the NVidia shader cache, an option in their control panel, saved me significant VAS.  I've not done a serious study but it is around 200-300 MB.  I wonder if your driver issue is related in any way?

Gents,

 

We spent a LOT of time (well...full disclosure, the beta team spent a LOT of time) testing the 744 in multiple areas to make sure leaks were caught and squashed. Some of you referenced the 777, and you are correct: it had a VAS leak. All the same, with that fact in mind, I don't think it's entirely logical to say "well the 777 had it so the 744 probably does," unless you were to make the argument that we never learn (although, in my case...  ).

 

Regardless, instead of pointing fingers and getting spun up about the situation, as I've mentioned in other threads, the important issue is data. We need to know exactly how to replicate what you're seeing on our systems. If any of you see oddities, we want to know about them, but we also need actionable data: Where were you? What were you doing? What livery were you in? What was your exact route? What exact approach were you flying (or had you selected)? What scenery are you using on that route (or did you cross over)? We just need something that we can go on to replicate the situation to determine where the leak is.

 

It may be us - it has been before. It may also not be us - this has also been the case (see the post two above this one). Cool thing about data, though, is that you really can't disagree with it. If the available VAS is steadily decreasing, then there's something to look into - either in our product or elsewhere. Having the data and the scenario allows us to at least look into it.

Why am I astonished? I get your point, what my point is why didn't they test over add-on scenery as I indicated in my post?

Doess PMDG actually think we're all going to be flying the 747 over vanilla sims?

And yes, I know there's almost an infinite amount of scenery add-on configurations out there but to test with just vanilla doesn't reflect most people's scenery setups.

As Rob stated earlier, they used a plain vanilla install to check for vas leaks, just to eliminate other factors. Obviously at other times they would test it with many variations of add-ons installed over the course of beta and within their own staff

Disable Nvidia shadder cash help me. I am on 368x Nvidia driver. Performance better.

Hi Gents,

 

As promised I report back:

 

Rolled back to NVIDIA driver 376.33 (from 378.49) and my OOM issues with the Queen vanished.

Same flight LFSB-LIRF mentioned in my previous post (didn't change a thing), landed safely and smoothly with 500 MB VAS remaining and slightly better FPS average (35-40).

 

Reccomend stand-by updating NVIDIA driver to 378.49, wait for next release.

Hope this may help all with OOM nasty dings flying the QOTS II.

 

Happy flying,

Some of you referenced the 777, and you are correct: it had a VAS leak.

 

Kyle, it still does. There is hard data with graphs and everything in that memory leak post in the 777 forum. All related to the FMC missed approach nav data I referred to above. We are still experiencing the same memory increase at various airports.

Hello,

 

I just have to report, that VAS consumption is noticeable higher then in the 737 and especially the 777 (which is better regarding the VAS as the 737 on my system)

 

I can't say if there's a leak, I just have to say that it demands noticeable more than the 777.

 

 

I'm still on drivers 376.33, wasn't updating. I have of course installed demanding scenery like Germany North, Tegel and EDDB. But I have a "test flight" with this scenery and settings and have no problems with 737 or 777. But 747 produces OOM (or very close to it, when not switching into external views.)

So, under same conditions, the 747 needs much more VAS. (on my system!)

 

 

So, I would give you any datas, but I do not know, what devs need to know?

 

Addons used:

Chaseplane

GSX

EDDT

EDDB

Germany North

SimServer

 

Win10 64bit

GTX 970 driver 376.33

Hello,

 

I just have to report, that VAS consumption is noticeable higher then in the 737 and especially the 777 (which is better regarding the VAS as the 737 on my system)

 

I can't say if there's a leak, I just have to say that it demands noticeable more than the 777.

 

 

I'm still on drivers 376.33, wasn't updating. I have of course installed demanding scenery like Germany North, Tegel and EDDB. But I have a "test flight" with this scenery and settings and have no problems with 737 or 777. But 747 produces OOM (or very close to it, when not switching into external views.)

So, under same conditions, the 747 needs much more VAS. (on my system!)

 

 

So, I would give you any datas, but I do not know, what devs need to know?

 

Addons used:

Chaseplane

GSX

EDDT

EDDB

Germany North

SimServer

 

Win10 64bit

GTX 970 driver 376.33

 

 

I have same scenery and addons and 368.69 Nvidia Driver but iam running on 4K monitor.   When i disable Nvidia shadder cash in Nvidia control panel,  VAS probem was gone and i had about 600 free VAS memory. 

 

777 which is better regarding the VAS as the 737 on my system

PMDG 777 is higher VAS than the 737. Its far more demanding its seem very odd that you find the NG on your system is higher, very odd. The NG is light on VAS.

 

The 747v2 needs no more than the 777, PMDG tested to death the VAS on it.

 

Here Rob tests NG and 777. As normal the 777 used more

 

 

9 hour flight from KSFO to EHAM VAS test ~(not me)

 

Rolled back to NVIDIA driver 376.33 (from 378.49) and my OOM issues with the Queen vanished.

 

Yes, I also rolled back to 376.33 and the OOM issue is gone. Tested with the same flight KIAD to TNCM after 30min of flight only 400MB VAS left with 378. With 376.33 I had steady around 800MB for the rest of the flight to TNCM.