P3D 2.0 Lessons Learned

[jjjallen 37]

I still remember driving 100 miles to get my copy of FSX as all the local stores were sold out.  Now after 7 years of learning how to configure it and with the many addons and now on my 5th computer build, FSX is finally at the stage where it runs in a performant fashion with a high level of realism.

 

Now, after all these years I now have an upgrade to FSX to play with called P3D Version 2.0.   After some initial trials (haven't reached the tribulations yet) I want to offer a few observations to try to gain a perspective outlook on Flight simulation and the future of this exciting hobby.

 

1) FSX remains as a testament to an excellent flight simulation architecture that was achieved by a team of dedicated individuals who, working under the leadership of an individual with a vision (aka Phil Taylor), created an extensible flight simulation platform that to this day has not been excelled.

 

2) P3D 2.0 can IMO be considered SP3 to the FSX platform and offers many optimizations such as implementing the latest version of D3D 11, dynamic lighting, shadowing, HDR and AutoGen optimizations.  In hindsight some of these where expected in the initial release of FSX.  Remember the FSX prerelease marketing glory shots of water that were much talked about?

 

3) P3D 2.0 has been marketed (possibly inadvertently and subversively) to the general Flight Simulation community as the next gen simulation.

 

4) P3D 2.0 now runs very well with said optimizations on medium to high-end systems.  FSX will perform similarly on these systems but without the optimizations.

 

5) P3D 2.0 can reduce fidelity in some areas such as AI, AntiAliasing, Addon support ect.  These should work themselves out over time however.

 

6) P3D 2.0 does not offer an improved AI subsystem, cloud shadows over land or an FSX equivalent start screen.

 

7) The Flight Simulation community in general is still hoping for a new visionary flight simulation platform that offers both support for their existing assets but is forward looking and offers revolutionary advantages.

 

In summary P3D 2.0 offers an evolutionary upgrade to FSX but unfortunately falls short in some areas to appease most FSX users.  

 

Going forward we need an Individual cast in mold of a Phil Taylor and who has the backing from a large software development company such as Microsoft who can move the Flight Simulation franchise to the next level. Most importantly, this consortium will need to be wedded to no one entity but only to the betterment of Flight Simulation as a whole.

 

Regard

jja 

 

 

[MarkSC 122]

The main lesson learned is that P3D has advanced the hobby tremendously because the flight sim is no longer CPU bound. With a high end GPU you can now fly with most sliders full right. As GPUs advance, performance will only get better. That plus the active development by LM means that P3D has a bright future. No use wishing for an alternative.

[Matthijs 79]

The most important point: LM is still working on updates and new features of Prepar3d V2.

So I believe it has a bright and great future. Just to be realistic: FSX is a great platform but it's getting older and older, it is time for retirement.

[mgh 89]

The Flight Simulation community in general is still hoping for a new visionary flight simulation platform that offers both support for their existing assets but is forward looking and offers revolutionary advantages.

 

If there was a market for such a platform surely someone would have entered it? The reality surely is that the hopes will remain just hopes.

[jjjallen 37]

Not true Mark - check your task manger CPU usages and you will see that several cores are maxed at 100% which is possibly due to the fact that P3D 2.0 (like FSX) does not optimize HT core use.

 

This is another reason why I developed FSXAssist which helps spread the load across all real and virtual (HT) threads and why one of the LM tweaks is to add an Affinity Mask setting of 14 to the p3d.cfg file.

 

Regards

jja

[Noel 4,486]

 

 

or an FSX equivalent start screen.

 

Boy I'm with you there.  I posted this on the P3D forum but it's been buried and unresponded to.

[jjjallen 37]

If there was a market for such a platform surely someone would have entered it? The reality surely that the hopes will remain just hopes.

 

The point is that LM (if they will be the best Flight Simulation provider) needs to break out of the FSX mold and develop a totally new platform but at the same time support legacy FSX assets.

 

This is how it is traditionally done when moving a platform forward.  

 

Regards

jja

[WR269 367]

Not true Mark - check your task manger CPU usages and you will see that several cores are maxed at 100% which is possibly due to the fact that P3D 2.0 (like FSX) does not optimize HT core use.

 

This is another reason why I developed FSXAssist which helps spread the load across all real and virtual (HT) threads and why one of the LM tweaks is to add an Affinity Mask setting of 14 to the p3d.cfg file.

 

Regards

jja

Why would it be an issue if the sim works steadily? I have a bottom of the range cpu and card, can run with third party add ons at egll with 40fps. I dont need any more, but it is smoith like fsx never was and stutter free. Lets face it, some will complain because fsx has not advanced and some will complain because it hasnt advanced enough and stay with the buggy old one.

 

It is all about personal preferences, my preference is to take any advancement in the market and enjoy it, however small it may be perceived to be.

 

Sent from my Mobile thing

 

 

[Noel 4,486]

 

 

The main lesson learned is that P3D has advanced the hobby tremendously because the flight sim is no longer CPU bound. With a high end GPU you can now fly with most sliders full right.

I'm not so sure this interpretation really holds true as it first might seem.  I'm finding what is added w/ the outsourcing to the GPU has less to do w/ the basic tasks of executing the simulator engine and quite a bit to do w/ DX10/11 effects.  What I notice is that in FSX moving in-game sliders had a profound and linearly scaled impact on performance such that it was possible to run the sim in a big range of computer processing power environments.  W/ V2, what I notice is that this is less the case--I can add lots of effects w/ virtually no impact, but the big ticket items like autogen still take a toll on performance.  About the only thing that impacts DX effects is putting Shadows to Ultra mode for me.  So it seems like adding the high end GPU, which I am fortunate to have, adds primarily special effects.  They are compelling, don't get me wrong.   But I would have hoped more of the basic work of the simulator--not visual effects--were going to be ported to the GPU.   I could be wrong on this for sure, but this has caused me to to believe SLI'ing my Titan will likely yield very little more than I already see in V2.  Quite frankly, I can't see any difference between Very High and Ultra for shadows.  Moreover, I also have the vegetation box checked for shadows--along w/ all other ones--but I can't really see the difference w/ that box checked, or not.  64-bit & building a core that can access LOTS of physical memory, since it's now dirt cheap, AND really aiming for maximum multithreading, would be the key to an engine every 3PD & simmer would want to jump on board with.  We'll see how P3D progresses & XP 64 as well. 

[jjjallen 37]

The point is in regards to architecture.  A computer platform should optimize based on need across all available CPU assets.  Given that, to a modern computer calculations are trivial so why would P3D 2.0 require 100% of a core when 5 others are lagging at 20%. This goes back to the basic design introduced in FSX (and before).

 

As an aside my team (in my day job) recently reduced a process from 33 hours to just 18 minutes by optimizing and off loading how calculations were done.  Basically you want to utilize the subsystems based on potential.  FSX had very different subsystem potentials when it was written.  This is why I emphasize a revolutionary code change rather than an evolutionary one.

 

Regards

jja 

[mgh 89]

The point is that LM (if they will be the best Flight Simulation provider) needs to break out of the FSX mold and develop a totally new platform but at the same time support legacy FSX assets.

 

This is how it is traditionally done when moving a platform forward.  

 

Regards

jja

 

Maybe Lockheed Martin doesn't want to be "the best Flight Simulation provider"  (what ever that may mean). Its objective could be to  concentrate on its training business - which it has always said from the beginning was its primary interest. It obviously decided not "to break out of the FSX mold and develop a totally new platform" with V2.0 and there's no evidence that it intends to do so.

 

Prepar3D is the best hope for the future, even though it's little more than FSX with vastly improved graphics. Lockheed Martin's strategy seem to be to provide "hooks" so that developers can create their own improvements. I suggest that will be of more benefit to those who want to develop Prepar3d for serious training applications rather than those who want to support the enthusiasts market simply because there'll be more money in the former.

[jjjallen 37]

Spot on Gerry and your post emphasizes the dilemma that we Flight Simmers are stuck in.

 

Regards

jja

[DarkstarF16 72]

7) The Flight Simulation community in general is still hoping for a new visionary flight simulation platform that offers both support for their existing assets but is forward looking and offers revolutionary advantages.

I'm really really struggling with this point. Am I missing some deep meaningful point here? I'm confused as to what it is your actually asking for?

[jjjallen 37]

LOL Darkstar - I'm just a flight simmer who like many is looking for the ultimate flight sim experience so I cannot give you any god-like answers. Suffice it to say that many like me are still waiting for FS11 in our hearts but alas MS disavowed the Aces team so we are crying in the wind at this point.

 

Suffice it to say that It would be really nice if.. Well we have P3D at this point for good or naught.  The optimist in me says let's do this thing right and if I had booku $ then it would be done.

 

Anyway if a company like say ORBX decided to do a new Flight Sim version then I would be fully on board with $ to back the effort.  It will just take someone to push the effort forward is all I'm saying.

 

Regards

jja 

[DarkstarF16 72]

LOLOL - IC your looking for nirvana.