[Official] DX9 vs DX11 performance difference

Discussion in 'Technical & Support' started by 4thworld, May 4, 2017.

Thread Status:
Not open for further replies.
  1. dadaboomda

    dadaboomda Registered

    Joined:
    May 15, 2016
    Messages:
    712
    Likes Received:
    202
    Me too I think that, but I can not imagine a single second that they will release their official release DX11 WITHOUT AT LEAST HAVE PROVIDED PERFORMANCE IMPROVEMENTS, even partial.

    I managed to have an acceptable graphic quality with 90% of the time 60 constant FPS with quite "original" settings.

    In recent days, I take a liking to play with AI, which was not the case before :

    Result, no way to have 60 FPS in the rain, the wet floor, 4 AI only, and possibly the headlights of AIs ON.

    Guess what with settings?

    No mirrors, automatic reduction mode set to 60 FPS, max visible vehiculed at the minimum, rain drops enabled, ABSOLUTELY ALL REMAINED OFF OR LOW (yes, yes, yes ........), vsync video of the game enabled, none post process, no AA, some improved player.json lines (headlights on cars = false, max headlights = 2, backfire = 15, etc.), and just 3K resolution to compensate for AA at OFF, the parameters of the Default GPU, latest driver update, all drivers up to date including BIOS, LIGHT overclocking CPU and GPU with increased fans, fully open and ultra ventilated PC tower with a home fan cooling all components. By cons, PCIe 2.0 X16. Windows ultra lightened and optimized, anti-virus and almost all disabled etc ... Oh and I forgot: time scale on OFF to relieve the processor (realroad a X10), headphysics has 220% (2.2). Replay totally disabled etc.

    BUT WHAT IS THIS HORRIBLE COMPUTER ???

    SSD, intel step disabled in the BIOS, I5 2500K OC + 15%, GPU GTX 780 OC + 20% (totally stable). 16 GB of RAM at about 1370 MHz. All components ultra cooled. Ultra optimized Windows. RF2 beta DX11 / only car content ISI / S397 and fullscreen.


    At some point, we must not mess ... I'm getting fed up with this disastrous performance. An overclocked GTX 780 is NOT a bad card.

    Please S397 take this feedback into account....... @Marcel Offermans
     
    Amanda Santini likes this.
  2. Suzukinol

    Suzukinol Registered

    Joined:
    Nov 28, 2017
    Messages:
    148
    Likes Received:
    54
    I just hope that if they release DX11 as the standard version will be because its a real non-beta build. If its still beta and giving poor perfomance / problems, then I think a lot of people will complain because its not acceptable that a 2015 graphic card designed to work with DX11 is getting worse perfomance than DX9 which isn't his dessign and also we bought a product that was working good with our computers. I believe in Studio 397 and I think they will release it because they did a lot of tests and they are sure its a non-beta build. If this is wrong... then a lot of hate will be in the forums.
     
  3. dadaboomda

    dadaboomda Registered

    Joined:
    May 15, 2016
    Messages:
    712
    Likes Received:
    202
    In this case, not hate. We are not kids!

    On the other hand, a REAL detailed support on the forum or via a detailed manual will be necessary to help the user to have a graphical experience and especially of PERFORMANCE really correct.

    If there is not enough performance optimization of the graphics engine, then alternative means must be found.

    Do not have to do hundreds of graphic tests to try to find a suitable result.
     
  4. ebeninca

    ebeninca Registered

    Joined:
    Sep 7, 2016
    Messages:
    738
    Likes Received:
    524
    Definitely what we got now isn't what we will have after de next update..

    Look at this...

     
  5. davehenrie

    davehenrie Registered

    Joined:
    Jul 6, 2016
    Messages:
    7,454
    Likes Received:
    4,369
    Why are folks expecting a sudden performance increase once the BETA tag is removed? I think, if we look back, the main jumps in DX11 performance came at the cost of the extra detailed skies that were first previewed. DX11 is very likely already operating at or near it's full rF2 speed.
     
    stonec likes this.
  6. Suzukinol

    Suzukinol Registered

    Joined:
    Nov 28, 2017
    Messages:
    148
    Likes Received:
    54
    If BETA tag is removed should be because its not in beta stage and that means that the DX11 engine is stable and finished but... if its the same -25 fps perfomance, then it shouldn't be called non-beta build because it will make some people angry.
     
    ericRacer likes this.
  7. stonec

    stonec Registered

    Joined:
    Jun 19, 2012
    Messages:
    3,399
    Likes Received:
    1,488
    The same around -10% difference from DX9 to DX11 has existed since July for my GTX 1000 series card and -20 to -25% for GTX 600/700 series. The gap is admittedly quite big for older Nvidia arcitechture, but I would consider it a surprise if they suddenly found some magic switch to improve it six months later.
     
  8. dadaboomda

    dadaboomda Registered

    Joined:
    May 15, 2016
    Messages:
    712
    Likes Received:
    202
    The key could be to have many more graphical parameters available in the UI and especially for each of them to have more possibilities to refine each parameter.

    It would be longer and tedious to find the ideal "global" setting, but this one would ultimately have a very good visual VS graphic quality ratio even for older computers.

    In addition, S397 knows perfectly their graphic engine and all the small optimizations not available in the UI (backfire, mirror distance, headlights on cars etc.) : please give us a manual PDF, which would not need good blood NOT 100 hours of work, should be provides.

    He would cite all the parameters that could have an impact on the performance by specifying :
    - Impact mainly on the processor, on the graphics card or both;
    - By giving the values "recommended but only indicative" according to his PC (very good processor / medium graphics card, old processor / medium graphics card, old processor / powerful graphics card, explanations of secondary parameters and how to optimize them (by example settings in the player.json mirrors that me eat 15 FPS currently ... just an example)
    - A lot of suggestions / advise to optimise performance etc.

    And yes, we can even expect that guys want / hope for a performance optimization for this release !!!! A partial optimization....

    Just be a little "emphatic", it's not because you have a good recent PC that you do not understand the real problems of others :

    You have understood that for the graphic engine of a simulation, the worst is absolutely not the "graphics of another age" : the real MEGA problem for a simulation is the performance.

    In particular the performance unstable and discontinuous : constant FPS, but soon that the AI passes you in front, that you activate your mirrors, that the AI put their headlights, that it starts to rain, that the road is wet or that the sun becomes low in the sky ........ BOOM: 25 FPS max, RF2 is no longer usable ...

    Very sincerely is it acceptable, we who supported for 12 years RFactor?

    Thanks to S397 for taking into account all the feedback, not only those who have a GTX 1080 TI 2 times more expensive than my first youth car.
     
    ericRacer and Rui Santos like this.
  9. MarcG

    MarcG Registered

    Joined:
    Oct 4, 2010
    Messages:
    6,854
    Likes Received:
    2,234
    Can you post your Benchmark Results please as per Post#1 of this thread, thanks
     
  10. Ari Antero

    Ari Antero Registered

    Joined:
    Jul 27, 2012
    Messages:
    1,882
    Likes Received:
    829
    There is one thing you should know about performance to 1080Ti in rFactor 2. If you are running rFactor 2 single monitor 60hz/60fps you are fine but if you have 120/144hz/120/144fps single monitor you must start to make compromises and lower you settings a lot and if you don`t have the knowledge which gfx options you should lower you may not be able to have descent game experience or play this game at all with your 1080Ti.
     
    ericRacer, dadaboomda and Prodigy like this.
  11. Suzukinol

    Suzukinol Registered

    Joined:
    Nov 28, 2017
    Messages:
    148
    Likes Received:
    54
    If they just changed the DX11 engine, then the game must have more FPS and not less fps. If all people are getting 15-25 fps less then I think its a serious perfomance problem.
     
  12. lagg

    lagg Registered

    Joined:
    Oct 1, 2012
    Messages:
    3,043
    Likes Received:
    1,958
    If this were true it would be a serious perfomance problem, but depends on many things.
    Not all people have less fps. I don't have less fps. I've the same.
    I had less fps with my old pc (I5, GTX560TI) but not with the new one (I7, GTX1080).
    I'm not saying that the problem doesn't exist. I say that the problem doesnt affect all people.
     
  13. dadaboomda

    dadaboomda Registered

    Joined:
    May 15, 2016
    Messages:
    712
    Likes Received:
    202
    Warning :

    I specified in what circumstances the problem was greatly amplified :

    Use of AI (the more there are, worse obviously)
    AND / OR
    Time of day or the sun is low (but also when the circuit is pretty ...)
    AND / OR
    When the rain starts (it rains at least 30% of the time in my weather scripts ...)
    AND / OR
    When the track is wet and especially when the water is "raised" by the AI
    AND / OR
    When the AI lights are ON or even mine
    AND / OR
    When I use the virtual mirrors even with only 60 meters of sight and without the smoke and particles (worse for the real mirrors).

    I forget things necessarily.

    But just imagine 10 AI in front of me, with their headlights, the rain, the water of the track "flying" behind 10 AI, the morning at 8 o'clock (low sun) and the need to put at least 1 mirror for avoid contacts ... there really in this condition not so infrequent as that, where I could have 60 FPS constant with absolutely everything off or in LOW, but in 4K / full circuit details / full texture details / anisotropic X16, I have catastrophic performance even in 3K without AA and really ALL of OFF even the circuit details etc.).

    For an i5 2500K OC, 16 GB of RAM, GTX 780 OC + 20% : it's really unacceptable (I think).

    I had found a correct solution before ... but that was before using AI. I do not want to drive without AI, which is legitimate.

    Let's wait for the release, let Christmas pass :

    On the other hand I would seek the help of S397 after. The help of the people in the forum is really great, but I think it's the people who work on the graphics engine at S397 to answer with a maximum of advice ... even if a manual well done that would not take them dozens of hours of work would be better ..........

    Let them send a release and enjoy the end of year celebrations. I would ask them after ;) .
     
  14. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    It affects to too many people. Maybe not all but enough to be considered a serious performance problem. The fact that some are unaffected is irrelevant.

    Furthermore, what annoys me more is that when Studio 397 speaks about DX11 becoming standard they act as if these problems didn't exist and no solution is being planned.
     
  15. patchedupdemon

    patchedupdemon Registered

    Joined:
    Jan 3, 2017
    Messages:
    1,602
    Likes Received:
    1,151
    If I can remember properly when iracing introduced dx11 they had better fps,but they didn’t actually improve the graphics at first,where rf2 did.
    I may be wrong as I usually am but with with better graphics one will eventually have to upgrade hardware.
     
  16. lagg

    lagg Registered

    Joined:
    Oct 1, 2012
    Messages:
    3,043
    Likes Received:
    1,958
    We don't know how many people are affected. Only a few users post in the forum. And it's more normal to post when something goes wrong that when goes well.
    I don't say that the problem doesn't exist. What i've said is that not all people have less fps.
     
  17. stonec

    stonec Registered

    Joined:
    Jun 19, 2012
    Messages:
    3,399
    Likes Received:
    1,488
    All people who made the detailed tests (by that I mean a graph or absolute number, not just a hunch) in this thread have reported less FPS in DX11 (some are very close to DX9, but still less). The only report I have seen of more FPS in this thread is from Ari running in SLI, which isn't surprising considering that SLI was 100% broken in DX9 rF2. There were a couple of users like @patchedupdemon who claimed they got much better results in DX11, but when they re-ran with proper benchmarking they found it not to be the case.
     
    Emery, demerzel and Ho3n3r like this.
  18. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    This thread contains many benchmarks from the people that write regularly in the forum. This is not a technical support thread but a benchmark one. It should be considered representative from an average user point of view. If you read through it as I have since I have regularly posted my benchmarks every new beta build, you will realize that it is more people losing fps than those who don't. I wonder if S397 has taken the time to collect the available data and judge perfomance.
     
  19. dadaboomda

    dadaboomda Registered

    Joined:
    May 15, 2016
    Messages:
    712
    Likes Received:
    202
    I do not know if you noticed, but 90% of the people who post on this forum are real regulars. It's almost always the same 100 people. And my impression that I think is good is that all these people are :

    - Regulars of this forum RF2 and often other forums of sim-racing.
    - Hardcore sim-racers in the sense that this hobby occupies a very important place in their life.
    - And where I wanted to come : they are in my opinion almost all well-equipped hardware PC (steering wheel etc.). This is not the case in my opinion of more traditional sim-racers who play less frequently, and who represent the majority of RF2 users. You must be among those who have a good recent PC, whose problems of FPS are very limited.

    When I tested AC for about 7 hours before uninstalling it (game used "as a demo"), AC gave me a benchmark. I did not take into account these automatic settings and I absolutly put everything in ultra. On the same circuit as RF2 Nordschliefe. Result ? 7 hours with graphics above RF2 even max without ever going below 60 FPS. CONCLUSION: the problem does not come from my hardware, not windows optimization, no drivers, no OC, no RF2 settings ................ it comes ............ from the graphic engine extremely poorly optimized. I would prefer less good graphics but performance FPS : simulation or not?
     
  20. lagg

    lagg Registered

    Joined:
    Oct 1, 2012
    Messages:
    3,043
    Likes Received:
    1,958
    Hey guys, calm down!
    I don't say that there is not a problem.
    One guy has said that ALL PEOPLE have the problem, and this is not true.
    Thats all.
    I had that problem in the past, but fortunately i could buy a new pc and i don't have this problem nowadays.
    I respect all people who have work to post all data to show the problem.
    I apologise If have offended someone.
     
Thread Status:
Not open for further replies.

Share This Page