P3Dv2 vs FSX: Addons Comparison

My dear simmers!

This is the 3rd and most probably the last comparison I will do for P3Dv2. It has become apparent that my tests are considered unrealistic, and bogus if you will. The thing is, I read many forums, including the biggest ones too, and I see many negative reactions to my tests. I see also many happy simmers out there, and I don’t want to the one to sour the apple.

The thing is, P3Dv2 isn’t at all that bad, as some have come to conclude from my tests. People, I do the test, I look at the results and share them with you. Nothing else. And this test was actually a positive one, but more on that later.

This test was constructed to see the difference between the sims when using addons. I have chosen two really great products that I own:

Flightbeam KPHX
Aerosoft Airbus Extended

Let’s begin with showing off my basic situation, from which I conducted the test:

FSX:

Image

P3D:

Image

And the settings used:

ImageImageImageImageImageImageImageImageImageImage

As you can see, autogen is maxed, scenery is maxed, GPU stuff is turned off. In both sims, I have comparable settings (no, autogen is NOT comparable, but max nonetheless).

So what happens when we add Flightbeam KPHX, observe:

ImageImage

Alright, Aerosoft Airbus Extended:

ImageImage

Well, I can only say disappointing. Or, is it?

Now let’s do another another small step, which I admit didn’t want to do, but did it to really compare apples to apples. Let’s try what happens if I remove ALL the tweaks from FSX, and leave stock FSX + these addons. And for the sake of being correct, I made a screenshot with both mouse off and mouse on the sim results, as we all known that mouse on lowers the FPS in FSX.

FSX, mouse off:

Image

FSX, mouse on:

Image

And then, P3Dv2 (no mouse bug):

Image

And this is what I was talking about before.

P3Dv2 actually performs on par with UNTWEAKED FSX. Not better, not worse. So LM team did a good job of importing many good stuff into the sim while retaining the performance of the STOCK FSX.

With my guide, users have become used to having good performance in FSX, under certain conditions of the tweaking guide.

I also now do understand why there are users out there reporting great results with P3Dv2:
These people are MOST LIKELY those that never had anything better than stock FSX running and were used to poor performance. To them, P3Dv2 is god-given, if it has no bugs. It is smoother than FSX, retains the “low” FPS number and it retains it steadily. But since I have a highly tweaked FSX, my FSX is performing better than P3Dv2 does. And the tweaks aren’t really working in P3Dv2.

To even throw a twist into the whole performance thing, I hit the situation with and manual overcast via weather page, and got this:

Image

As the conclusion, I will simply say this:
If you had FSX running well and tweaked according to my guide, P3Dv2 is most likely going to be a slight disappointment. It will perform slower when it comes to FPS.
However, if you did not, P3Dv2 will be a joy, since the performance is same (or maybe even better?) and the stability is given or coming.
They still need to fix some quirks and give it Nvidia support, but to achieve the tweaked FSX level, they will need something else. Whatever that might be.

Advertisements
This entry was posted in Testing. Bookmark the permalink.

69 Responses to P3Dv2 vs FSX: Addons Comparison

  1. John Orfanos says:

    Your tweak guide is the best thing that happened to Fsx! Thank you. Your tests trully add to the fs community!

  2. Alain says:

    Hey Kosta,

    Are you “gonna” let the fan boys tell you how or what to do with your result or how to test for them to see what they would like to see?? Keep doing what you want the way you want it and D’ONT LET anybody tell you otherwise.

    You would have been a God in their eyes if your results had been favorable to them, it’s not like you had something against P3D, you are just impartial and some can’t cope with that, either you are on their side or you are a naysayer. Again, the vast majority of simmers that don’t like the truth are the one with a lot of money invested in 3PD add on, they are the one saying that you should not say anything negative about P3D no matter what, I can’t blame them as they want somebody to improve a platform for all their add on, like that they can still use them without having to invest in new one…= $$$$ saved.

    The fact of the matter is that there is a LOT more simmers having problems with P3D compare to others saying that P3D is the best thing since beer in a can, and that is the truth as you can read it all over the web, the bottom line is that P3D was NOT ready for release, far from it.

  3. John says:

    I agree with your conclusion.
    I had a well tweaked FSX, but could never get a smooth ride and I always felt one more fiddle would get it perfect…but it never did become “perfect”.
    After one reinstall of P3d v2 (after an abortive attempt of installing all my add-ons in one go..how daft is that?) and a fiddle with the mods suggested in the learning centre, I now have a beautiful looking simulation that flies smoothly, and one by one the add-ons are becoming perfectly compatible. I have not looked at the set-up sliders screen for many days and have no desire to…it is working beautifully and I will leave it well alone. If any patches cause more efficient use of the GPU etc then that will be a great bonus.
    There are OOM and runtime errors to be fixed but I have avoided these luckily.
    I think we have a winner here as long as we don’t fret too much about what is going underneath if it is giving us what we want!

    • Kosta says:

      John,

      “I think we have a winner here as long as we don’t fret too much about what is going underneath if it is giving us what we want!”

      I agree. However, P3Dv2 isn’t giving me what I want. It slows down below my tweaked (and smooth) FSX and that’s already not enough for me.
      Currently there are no heavies available for P3Dv2. Wait till some come out (PMDG).

  4. Andrew 737 says:

    Great tests

    Thanks Kosta

  5. abraham says:

    Thanks Kosta…FSX and P3D2…”the war” but…what could we do with the great “volumetric fog” and “cockpit shadows?”…who can now ignore them? …Maybe we do need a third one. A clever enterprise that mix the best of them. Thanks from Spain!

  6. Rob says:

    Many thanks for always taking the time to do all these comparisons. Your time devoted to the FS community is much appreciated

    Rob

  7. Torkermax says:

    Kostas, many, or most, of these “people” are over zealous blinded fools drenched in denial hoping in there minds the Flight Sim Savior has arisen. πŸ™‚ Hope that wasn’t too caustic.

    “I see many negative reactions to my tests. I see also many happy simmers out there, and I don’t want to the one to sour the apple.”

  8. Emin says:

    Hi Kosta! I agree with Alain. Forget about fanboys and please continue to do what you’re at best . I’ve also received many negative reactions when I posted my initial opinion about V2 at Avsim and Orbx forums. Fanboys are generally those newly converted FSXers. These are same people were saying negative things about P3d 1.4 while I was defending and using it from the beginning.

  9. dazz says:

    Great comparison Kosta my friend (this one and the GPU one). I was wondering this same exact thing, how would P3D v2 perform without the new features compared to FSX. Thanks for taking the time to test it man.
    I won’t be an early adopter, although I still think there’s promise in P3D so most definitely you’ll be one of the go to guys for accurate info and meaningful tests.

    Take care bud. All the best

    • Kosta says:

      “Great comparison Kosta my friend (this one and the GPU one). I was wondering this same exact thing, how would P3D v2 perform without the new features compared to FSX. Thanks for taking the time to test it man.”

      Indeed dazz,
      This is what it does and it’s actually no better than FSX. Still the very similar CPU bound app, which in best case scenario works a little bit better when fully loaded with autogen, cars and AI. Versus tweaked FSX it fails miserably, however.

  10. I’m totally agree with you…i used your guide and my fsx was really good. Right now I’m really disappointed with LM. I was testing AXE in Aerosoft Tegel with OPUS Weather..12 FPS, in FSX 25/30. The graphic is pretty good, but not the performance, so i hope they will fix and improove it.
    Thank you for your tests and i hope in a new guide to use P3Dv2 at the maximum.

    Best regards

  11. Kosta says:

    Guys, I can’t tell you how much I appreciate your support here. I know these are only fanboys, and I certainly won’t allow anyone tell me what I should do. However, the tests done show what needed to be shown and anyone can base their decisions on that.
    The fact is that P3Dv2 looks better, fog, autogen, shadows, water, clouds… we can’t dismiss that.
    It is, however, still way too CPU bound in the core sim! The autogen is still way too much CPU intensive, and that needs to be changed if we are to have better performance with FSX addons!!
    That other stuff is GPU intensive is clear and good so. We need autogen as GPU intensive part of the sim thus freeing us completely from the CPU bounded state, so that there is enough computational place for the addons. And then, the possibility to run the addons from the GPU, or at least optimized, that would indeed give us a boost, especially if dual or even quad SLI gets involved.

    • Paul says:

      I agree Kosta.

      It also would be nice if NVidia would start to give P3D,FSX, and any other serious flight simulator support like it does for all the violent shooting games. There has not been hardly any mention at all about flight simulator in any NVidia beta updates or WHQL drivers ever. But our simulators should be the standard of performance for every PC program, because they are the most demanding of performance (if only we could be at 64-bit).

      Too bad flight simulators are not as popular as games full of killing and other forms of brutal violence.

      This is certain:
      much greater and real (not fake) violence will be the result, its already happening and everybody is looking for answers as to why….I say ask NVidia, see what they say. With out them the violent game producers would have a difficult time *like we are* having their games run smoothly if at all. I’m not saying NVidia or video games are the only cause of all the violence; parents, movies, and even cartoons can share the blame as well. I know that NVidia supports other genres of PC games too but there is no doubt that high favor is given to the violent games ( you know the titles : call of duty, assassins creed, and the list goes on and on). Just wait until the next world war happens and see if any of the fans of these games enjoy war anymore.

      I say all this to measure how long it takes NVidia to get around to supporting P3D V2 …a non violent (in other words *not* popular) flight simulator.

      If P3D V2 was all about destroying cities full of women and children, I might even bet my life that we would have a specialized video card driver made just for it and it would be quick!

  12. Feno says:

    Hi Kosta,
    I totaly agree with your conclusion on P3Dv2. He still needs to grow a little !
    one day, it could be the FSX killer … but not now surely

    Thanks you for your test, continue.
    Best regards from France
    Pascal

  13. Joseph King says:

    Good job Kosta…you speak English !

  14. smcnitt says:

    One big issue I see with this comparison is that you used the highest (Large) LOD radius in FSX which you know is 4.5 in the config file. In P3D you used MAX which is a 6.5 LOD in the config. To get apples to apples you need to set LOD in P3D to High which will yield 4.5 LOD. Otherwise your result is skewed because P3D had to render approx 44% more autogen due to the higher LOD.

    • Kosta says:

      smcnitt,
      You are right, I overlooked that.
      However, the “big” issue, as you mildly put it, is in FSX 56 -> 53 in a tweaked scenario and 40 -> 38 in the untweaked scenario.
      It really doesn’t change the comparison much.

      • smcnitt says:

        Agreed. I’ll have both sims installed for a while until theres some quality tubeliners available anyways. Until then P3D will be good for GA VFR flights πŸ™‚

  15. Marco says:

    Maybe it’s time for X-plane?
    The only platforms with future are P3D V2 and X-plane at the moment.
    I’m just realistic.

  16. yayuyo says:

    Is this comparison rellay valid? As LM many times says, the setting are very different between FSX and P3Dv2 even thought slide bars look the same. On which basis you can say that the setting you have made is exaclty the same between FSX and P3Dv2? How about Tessilation factor? Number of autogen per a ceartin m2? Another point. Is your conclusion really general? Some simmer has a slower CPU with a powerful GPU and vise versa. Is FSX faster or slower much depending on hardware configuraiton? It would be nice if you can clarify those points.

    • Kosta says:

      yayuyo,
      Of course it is, why wouldn’t it be?
      I know what LM said, I was on the beta team, and discussed this quite many times.
      The thing is, P3Dv2 renders more autogen far away, but renders less autogen near the aircraft. I can’t say if the general number of autogen objects is the same, and they said it’s not. However, that being said, I have confirmed that a situation that performs the same in P3Dv2 concerning autogen as compared to FSX contains WAY LESS autogen than FSX in a whole area. Basically sparse in P3Dv2 performs the same as full in FSX, approximately.
      I have never said the settings are the same, and one would be stupid to say that. The thing is, CPU is extremely busy with the object placement, but the GPU is the one rendering them. And this is what I am urging them to optimize and/or change.
      Tessellation has been confirmed that it works better, and is not the point of the discussion really. The fact is, CPU gets bombed by the autogen and still performs poorly, just like it did in FSX.
      Your questions above are valid, but have been answered by me numerous times already. Just read my comparisons (all of them) and comments to them.
      I compared IB that is faster than my computer, and also a Titan GPU. It was no faster in a CPU limited scenario (which implies LOTS OF AUTOGEN), and neither it was when I added some typical addons.
      LM is going all out saying P3D has been put onto GPU, which is actually a bit of a two-sided sword – they put the terrain on the GPU, which is working great. They “fixed” the autogen to not pop up, which is fine, but nothing changed much as how the system is handling the autogen. Instancing was introduced which reduces the draw calls on the CPU (or, it should), but it still is very CPU intensive – way too CPU intensive. And they shifted the autogen outwards, putting less autogen below the aircraft.
      In the end, you can’t get higher frames than you ever did in FSX, if you use lots of autogen (from Dense up).

      • ChaoticBeauty says:

        Since the autogen needs to be completely redesigned to be moved to the GPU’s workload (essentially breaking compatibility with add-ons), wouldn’t it be a good idea to modify it so that it takes advantage of multiple cores? In the DirectX 11 developer blog, Lockheed Martin promised multi-threading support, but from what I see, it’s just Core1 getting hammered and the rest of the CPU pretty much sleeping. Not much of an improvement compared to FSX in my opinion. Maybe this is something that you should urge them to improve too.

        By the way, thanks a lot for your observations and testings. They’ve been a real eye-opener. Prepar3D 2.0 has great potential, but it seems that it’s not the revolution that everybody expected (so far).

      • Kosta says:

        Chaotic,
        If you follow the performance thread on LM forum, I have a nice conversation with Beau and CPU usage in the P3Dv2.
        He claims it’s impossible to move the tree and object placement onto GPU, at least that’s how I understand it.
        He said however, going for multiple cores would make more sense. I agree, if it actually boosts the performance.

      • ChaoticBeauty says:

        That’s the thread I’ve been following for the past few days, and it contains quite some interesting information. I’m kind of confused by what Beau says. He says that working on multi-threading would improve hardware utilisation (and I’m sure it will increase performance, right now Core1 is getting hammered with all that autogen, AI and physics while the other cores are sleeping), but did he confirm at all that they’re going to work on that?

      • Kosta says:

        Chaotic,
        Let’s see what they say onto my latest post. Let’s assume that object placing is a CPU problem and intensive task, non-changeable.

      • ChaoticBeauty says:

        So if I understand correctly, Beau says the autogen in Prepar3D 2.0 is already multi-threaded as much as it can be and there will be no more performance enhancements?

      • Kosta says:

        Chaotic,
        Sounds like it. That’s about it for me, I’ll be refunding P3Dv2 at the end of January (approaching the end of 2-month period) and will probably use FSX on a monthly basis, like 1-2 times to make a flight, until either something else comes up or v2 really becomes better performing and addons greatly support it (PMDG…). I don’t like being a “non-paid” beta tester, as we didn’t even receive the software for free after beta testing, I had to buy it. There were never any promises of that, but it’s simply a custom of giving beta testers the software they helped the developer to test and make better (essentially making them more money) for free.

        And what really makes me tick off, is that Beau said that my scenario covers only a little portion of the spectrum and a single machine test – he either missed, or simply chooses to ignore the fact that I parallel-tested an IB machine with the Titan, and the results have been proven the same. So no, it’s not only ONE machine, it’s TWO, and the strongest card currently on the market.

      • ChaoticBeauty says:

        I agree that this was a very disappointing answer. I honestly can’t see how Prepar3D 2.0 has improved multi-threading when CPU utilisation stats show Core1 getting hammered and all other cores almost sleeping. If it were a properly multi-threaded application, it would be able to use all cores to the max, and I doubt they threaded out the AI, physics and other CPU work too. They can’t call it a GPU-bound application just because it has shadows and water that stress the GPU a lot. When GPUs advance so much that Prepar3D 2.0 will look like an easy task, but CPUs get more cores and show no benefit, it will have the same fate as FSX. I wonder what performance will be like when heavy add-ons like PMDG aircraft are used.

        I don’t think they handled the whole beta-test thing correctly either. The release product is riddled with tons of bugs (at least many are getting fixed for 2.1), and lots of add-on developers were caught off-guard with many of the compatibility issues between Prepar3D 2.0 and their FSX add-ons at release, while many of them had been testing their products in the beta. I agree that there should have been at least a discount for beta-testers, but it feels like everyone is a beta-tester right now, with so many bugs and compatibility issues going on.

        So much for Prepar3D 2.0 being the true successor to FSX.

      • Kosta says:

        Chaotic,
        Couldn’t have put it better myself. My points exactly. But neither those blinded by LM nor LM themselves will recognize it. I have tried talking to Beau, and his answer, being as complicated as it sounds, tells the same story over and over again, that they improved the performance and how I need a stronger hardware. Yet, on the stronger hardware, the performance backgrounds are still the same. I have proven him wrong, yet he still is holding to his own story. Fine by me, I am exiting the grounds and will maybe reenter when I see others do test (however, not believe it’s going to happen). I won’t be paying for the stuff only to get told that my tests (although true) are faulty.

        I’ll be keeping it until the end of two months now, see what they can improve, but that’s about it for me.

      • ChaoticBeauty says:

        This is the problem, that too many people are blinded and are thinking that current hardware is just not enough, but in reality it is underutilised, just like FSX. I’m sure that this is going to show in a couple of years when we have GPUs like Volta from NVIDIA, and Prepar3D 2.0’s performance will have remained the same. People on AVSIM are already reporting marginal differences with new high-end GPUs, but I guess that if I try to explain that they’re actually CPU-limited, I will get flamed.

        This is why I thank you for all those tests you made, they are huge eye-openers and expose the reality. I’m hoping that one day Prepar3D 2.0 will become what we want, a great platform for all those amazing FSX add-ons, but with Beau’s latest answer, that looks like a pipe dream.

      • Kosta says:

        Chaotic,
        People will keep on building the “effects” and will have better performance, that’s the fact. However, the core performance will never get any better. Like this: people will install PMDG NGX in the v2, and will be getting 15fps in some scenario. LM claims for P3D to be GPU limited. So people will buy the next gen GPU, and what will happen? You will still have these 15fps in the NGX, however you will be able to turn up the shadows! Yay!

        You’re welcome, man πŸ˜‰

        Hope dies last.

  17. abraham says:

    You should work in LM. Like this, all this great analysis would be destined to make a good work!. We need people like you working there to feel the products are well done before being saled.
    thanks from spain!

  18. niggle says:

    Kostas keep up the good work, dont let the flamers get to you, there may well be other systems out there that contradict your findings, but you are saying it as you see your system does it, more power to your elbow

  19. Rune says:

    Sirdan Kosta, keep up the good work πŸ™‚ We need some “unbiased tech” reviews from you also in the future, keep posting very good guides and stuff (and ignore “the very negative” guys)
    That said, have a stil Merry Christmas and a soon Happy new year also. πŸ™‚

    I have also bought Prepar3D v2 now and a Titan 6GB card and agree with what you said earlier that the boost from 285 to newer card “is big” even in fsx and that is from releativly untuned and defragmented harddrive hehe πŸ˜‰
    (I have just applied som bojote cfg-tweak service, sorry to say just too lazy right now :-p)

    yours and Nick N guides, are just a bit too involving for me “right now”, BUT I really appreciate the information, and read with great interrest about the inner workings πŸ˜€
    (been tweaking quite a few years)

    and might try some stuff of it in the future.
    I also understand from Nick N. bible, that the i7 Haswel 4770K is a really big boost combined with gtx780, and with good mainboard and quality fast ram.

    Thanks
    Kind Christmas regards from Norway

    • Kosta says:

      Rune,
      Yes, it indeed has more information, but as you progress, I am sure you’ll find it useful!
      4770K + GTX780 is surely a sweet combo, however I question this “boost” over 2600K/2700K and GTX580, in any kind of scenario. Of course, if you get GPU limited by using high IQ and heavy clouds, and max out the card, the faster one will help, but that is not the point πŸ™‚

  20. jackson says:

    Very nice comparison, Kosta….
    I am still “on the edge” and thinking of making the jump…the point is this year I have spent my best moments ever since I have started flightsimming thanks to a perfect combination of hardware and tweaking (mainly thanks to your guide..:-) ) which have made my FSX finally a joy to use, and despite I see a big potential in P3DV2 I am honestly scared to start everything over again…
    So, in a word to put it simple, would you reccomend to move from FSX to P3DV2 or not…or suggest to wait for P3DV2 to reach its full maturity before switching to it…?
    Thanks

    Ciao

    • Kosta says:

      Jackson,
      Your decision, and I would recommend you try the developer license. Cheapest and you will quickly see what it really is. But even with that, wait at least for the first patch. If not 2nd or 3rd… I know I won’t be trying it out again until they announce some further core-changes, Nvidia and driver support, 64-bit, and until PMDG comes out for it. Then I will wait to see what people say about it. My belief into P3D has been lost really, and will be very hard to gain again.

  21. Greg says:

    Appreciate your work Kosta.

    Just wondering how you get away with no SMAA in P3D?
    On my system that produces a blurry mess.

    Thanks,
    gb.

  22. Samuel Dickes says:

    Kosta, I want to thank you for all your analysis, it’s very interesting and so are many of the thoughts of people posting on your page. I have to admit that I am a little bit surprised by the results and I do not contest anything, the opposite, IΒ΄m also convinced more than ever that most poeple did not take advntage of the tweaking possibilities FSX offers to them. The stock configuration files parameters were done by Microsoft 8 years ago with the hardware of 2006 in mind and they do no fit to the powerfull hardware of nowadays!
    I do own a MSI gaming laptop with a Haswell i7-4700MQ which overclocks to 3,3GHz and a Nvidia 780GTX mobile graphics card and yes, I tweaked my FSX. While I don’t benchmark the performance of my system in kind of a scientific way that you did I noticed two things:
    1. FPS didn’t go up significantly in P3D compared to FSX, but smoothness went up in a way that changed my sim experience. My first reaction was that FPS does not tell the whole story after all (but I admit it is more objective if you want to talk about numbers). I didn’t use fraps to take a look at FPS but just the shift+Z function.
    2. The look of the whole sim has changed. The atmosphere looks much more realistic in P3D. Just look at the DX11 effects on haze and fog or the light in the atmosphere at dawn and dusk and the impact this light has on ground textures. Not to mention the effects of the sun on clouds. It’s worlds appart from FSX. As I am a private pilote in the real world, I can tell you that it comes much closer to what I see when I’m in the real plane from our aviation club (even when colours still look to punchy, even in P3D).
    3. Support and development: Take a look at all the enhancements and updates LM did provide for P3D 1.x over time. That should give us a glimpse to what we can expect in the future. I have to say that I’m disappointed that LM went to the market with so much bugs, delivering us nothing more than a beta version in my opinion. But I’m convinced that they will fix things and deliver even much more features we even don’t know about today over the lifecycle of the product.
    FSX will remain on my system for a long time to come, but P3D will remain there too for sure!
    Cheers

  23. Samuel Dickes says:

    One more thing, LM state that they did manage to knock down on OOM problems while recoding parts of the core FSX engine. This does mean that FSX has a barrier of 4GB of RAM usage. This 4GB RAM includes the normal system RAM used by FSX as well as the RAM used by the GPU! P3D v2 now shifts this barrier on 64bit systems in the way that the simulator can now use up to 4GB of system RAM plus up to 4GB of GPU RAM!!
    Now if somebody has real experience on this topic, it would be nice to here from you πŸ™‚

    • Kosta says:

      Samuel,
      Thanks for your posts, I have a question about the 2nd one: where is the thread you are referring to, about LM knocking down the VAS memory usage? I read about the UI menu fix, but what about the scenery, or were you referring to the UI fix?

  24. ChaoticBeauty says:

    Hey Kosta, I took a look at the thread in which they’re posting progress on the upcoming 2.05 update, and under the list of issues they’re working on, they’ve added “further optimise autogen system”. Do you know if there’s already an issue that is to be fixed, or could it be a general improvement in autogen performance?

    • Kosta says:

      No idea man, sorry.

      • ChaoticBeauty says:

        Have you refunded Prepar3D 2.0? If not yet, when does your 2-month period end? If they can get out the patch before that period ends, it would be great if you made a new test.

      • Kosta says:

        End of January. I would also like that. Failure to do that would mean me getting a refund probably until P3D v3…

  25. Kosta says:

    And there we go. There might have been more reports, but I just decided to visit AVSIM’s P3D forum to see if there’s anything new and was greeted with one of the newest posts:
    A user (pmb) has upgraded from 560Ti to the 760, and didn’t have any FPS raise. After all BLAH-BLAH-BLAH from P3D team, about how they moved the sim onto GPU (not they did NOT, they merely put GPU intensive stuff INTO the sim), this post, as it stands, shows the same thing I have been testing all and all over again. I have been right all along, and I knew that.

    http://forum.avsim.net/topic/432730-news-graphics-adapter-same-fps/

    “After inserting the new card and installing most recent Nvidia drivers I was amazed to see exactly the same 18 fps on Sydney runway. Actually, I should not complain, after takeoff fps go into the 20’s, this is dense scenery, it actually looks quite stunning, as I said, settings are certainly on the high side. Overall I enjoy a quite fluent ride with very few stutters, if any at all.
    I just expected fps to raise.”

    ” I did not expect a fps boost in Prepar3d1.4 being CPU bound, but wasn’t I told 2.0 being much more GPU oriented which should pay off?”

    Hahahaha, it was surely not me who told him that. I remember LM saying that to me however, when I told them that I have a GTX580 – the expression was in lines of “get the Titan and you’ll have 30fps in every scenario”. But they didn’t mention addons…
    Well, I guess some people will possibly and eventually learn that I *do* have some experience on the matter. Thanks again to everyone who helped me test too.

    • ChaoticBeauty says:

      What happens when you try to reveal the truth on AVSIM? You get flamed, and unfortunately that’s true. People won’t realise that Prepar3D 2.0 is still CPU-bound (or maybe Core1-bound) until GPUs can handle the shadows and water easily, and that’s when you’ll be limited by the CPU all the time.

      And 18 FPS? Oh dear, I was hoping that maybe we could actually achieve 60 FPS with Prepar3D 2.0, yet you get the same (and worse) performance with FSX with just a few effects more. Matters will become even worse when everyone realises that G-Sync will never work because that requires exclusive fullscreen (the fake fullscreen in Prepar3D 2.0 won’t cut it). Same with ShadowPlay. I can’t believe Lockheed Martin aren’t aware of that yet, and are trying to “continue to work with NVIDIA”.

      Starting to lose hope with that one. Hopefully someone else will get an ESP licence and make the simulator we all want.

      • I think you’re too idealize gsync:) only light/shadows and fog system too good for be true, nvidia, intell and amd doesn’t sleep, just same as lokheed
        lokheed has to be give a some more scheat about addons compatiblity, but as they say this sgead not for entertainment lol, but who care:)
        now i use fsx with stev’s fix dx10, and willn’t install p3d2 about half years, till new bought new age vcard:)
        but people with good cpu ram and vcard could be happy and entertainmention with this not entertainmention product hehehe
        well, i think after half year for p3d2 will so mych great addons for absolutely kill fsx, not for everyone, some people fly fs9 right now, and good feel with this, so may be same will with fsx, any way for p3d2 realy big future even if lokheed will close project like microsoft did

    • must be user use some cpu depended scenes, because some users said what their vram load around 4gb, i dont belive what 760 no give some after 560ti in much scenes, with good grafic settings, i sure that user got good cpu amd ram too
      my 560ti after upgrade system use around 100% fpu in dx10, and about 50% vram, i even wont try p3d till got new vcard and some patch from l.martin, but preaty sure this will eat my 560ti and all vram like kids in school

      and i think every chip have to use so much vram how was meant to be play, 760 with loaded 4gb could just botlenecked by gpu, what couldnt work with this massive correctly
      more ram or more vram than enoug is realy big mistake, firstly more same ram will slower in any way, secondary loaded ram realy depended of content and chipset what will work with this content, so user with 760 4gb have to use some other settings or upgrade vcard

      but problem not in above causes, just something was changed in settings
      you self belive what same settings will give same performance at thos cards, at his 2600@3.4:)?

  26. Isaiah says:

    I find these tests very useful. Too bad we – the flightsim community – always need to come down to one vs the other; Airbus-Boeing FSX – FS9 and now a new battle has immersed FSX v P3D..
    I have a nice set up at the moment and FSX is running smooth. As good as it gets within the limitations. I do keep a close eye on the developments in P3D and might one day transfer to that platform. For now FSX will do

  27. Pat says:

    Indeed very good testing kosta!

    The best part is, that I also think we are going to see a huge amount of vas issues when the heavy hitters like pmdg will be released.

    The funny thing with that is, that like u said, LM are advising on using a high end gpu.

    Problem is, if you already have vas issues on your current setup and you expand your gpu memory, guess what’s going to happen? Even more OOM’s

    I’m sure that LM must realize by now that they have made some mistakes and wrong choices, but I also think they are kind of stuck now and have to make money 1st before they can even think about a 64bit app with actual Multi threading going on.

    I think we will be on fsx for at least a couple more years.

    • Kosta says:

      Pat,
      Exactly. This is when problems will arise.
      You are right. Higher hardware doesn’t help when you introduce addons and when you add higher VRAM, OOMs happen even sooner.
      But yeah, I wondered why the pre-released P3Dv2 before it was really ready, I guess it was for financial reasons.

  28. Marc says:

    I couldn’t agree more with your assessment of P3Dv2. I have found my perfectly tweaked and balance FSX (as per your guides) with a myriad of addons to be a better overall performer compared to vanilla P3Dv2. The jury is still out on whether P3D will outperform FSX with the same plethora of addons. Also we do not know how P3D will perform with SLI support down the road.

    • Kosta says:

      Marc,
      Indeed, SLI could be a big help, however very questionable, as P3Dv2 still relies MUCH on the CPU – see my current thread on texture and autogen loading, I am succeeding, and that very easy, to overload my CPU just by mere flying (F22 at 1100kts). GPU is at 60%, and that is my “old” GTX580.

  29. Marco says:

    Kosta i agree with you i’ve done a little comparison with a pc with 4770k 780 GTX same scenery aircraft and settings like yours. P3D V2 doesn’t perfom well as a tweaked FSX on the same machine, now i see people around screaming about the beauty of P3D and the stability. First propably they are not able to tune FSX and second they are only flying little vfr plane around scenery with only mesh and orbx global. I would imagine this people when they will try a futore P3D PMDG 777 into scenery like Heatrow with 5 fps and an istant OOM.

    Regards

  30. Massimo says:

    Hi guys! In these days I’m trying P3D V2. I see some differences with FSX. I’m using it without any tweak and I notice less or no stutters vs FSX with the same hardware. Probably both sims are quite the same in some areas, but what I’d like to have is a sim without annoying behaviours, like low fps without any reason (you restart the sim, reload the flight and fps goes high again!), CTD, low loading time when you start the sim and so on … Probably in this sense P3D is much better and, in any case, it could be improved. Anyway I’ll try it more and finally I’ll decide what to do. Stay with FSX and wait a new improved version of P3D or change to it immediately πŸ™‚
    Now is there any tweak for P3D V2 ? And for Nvidia drivers settings as well ?

    Thanks!
    Bye, Max

  31. MrTact1 says:

    Very interesting site! I was looking for more info about P3D. At the moment I have a (5 year old) mid-end PC and wants to buy an extra high-end PC, because I’m intending to build a fully functional Cessna-like home-cockpit. Should I leave FSX and move on to P3D or Xplane? But I had my doubts, I’ve read too many ‘not so good’ reviews about P3D. I’ve also seen loads of in-game vids/pics of P3D and unfortunately I’m not really impressed. My graphics still looks better in my tweaked and with (the right) addons FSX, including a stable 25-30 fps and an average of one OOM per year. Would be great if the ‘out of the box’ version of P3D would have the same graphical/performance level of my customized FSX.
    So for the time being, I’ll stick with my good old FSX.

    • Kosta says:

      MrTact,
      I must agree with you. Although P3D is quite advanced, it is still not usable as an allround simulation. It will take some more performance development and Nvidia/AMD support to make P3D what FSX is or was.

  32. Michael says:

    The contract of P3D and FSX is not obvious yet. Since FSX is static and P3D is an on-going project of LM, I will keep an eye on P3D, seriously. I am having great expectations of P3D, hope the victory of the versus with FSX will reveal in a year.
    For now, I’ll just fly on P3D few times a week. My computer is mid-ended, and I have little tweaks on FSX.

  33. Frank Xu says:

    very detailed test, but i wanna tell u that: P3DV2 DO NOT need any sort of tweak. if u tweak P3D as what u do as FSX, it wont get any faster but slower

    • Kosta says:

      Frank,
      First – take a look at the date. Then yes – that is P3D v1. And yes, it can be tweaked as FSX (still not talking about P3D v2 in this test!).

  34. Frank Xu says:

    Merry X’MAS BTW

  35. I am so grateful for your article.Really thank you! Will read on…

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s