Latest Release 1109 Rain Effects AND MORE!

Discussion in 'News & Notifications' started by Christopher Elliott, Dec 21, 2017.

  1. DrivingFast

    DrivingFast Registered

    Joined:
    Mar 1, 2018
    Messages:
    1,638
    Likes Received:
    1,083
    ....... have fun with 30 AI for 25000 meters .....

    1 AI every 850 meters, what fun !!

    In real life the 24h nurburgring welcomes more than 100 cars: simulation or not simulation ?
     
  2. patchedupdemon

    patchedupdemon Registered

    Joined:
    Jan 3, 2017
    Messages:
    1,602
    Likes Received:
    1,151
    Racing ai isn’t fun I agree,and isn’t simulation just following a train of cars, but seriously, which sim with the level of physics that rf2 has, allows the use of 80ai, what’s the performance like.
     
  3. mesfigas

    mesfigas Registered

    Joined:
    Oct 15, 2015
    Messages:
    1,722
    Likes Received:
    830
    @DrivingFast
    i have a feeling that 61 AI is just traffic and not racing but thats just my opinion
    as for the results of your test i v tried your settings and had different results in rain on clear sky
    i believe shadows is the key when speaking about performance
    i manage to have 21 minimum fps 6.30 in morning at rain 12 hours sebring track with your settings.
    but it was playable
    when looking at minimum fps doesnt means that we always run in minimum
    minimum is mostly the fast corners
    moments in race that game demands insane power to give you smooth fps
     
    patchedupdemon likes this.
  4. ceecee

    ceecee Registered

    Joined:
    Jan 10, 2012
    Messages:
    691
    Likes Received:
    286
    60-100 AI ? My question is why ?
    If you want to see cars on the track, start from the back of the grid and race forward.
    AtlantaMP is only 2 miles (3.2186 km), 50 AI could get kind of busy !
    Entry grid numbers for the following race was :
    2018 12 Hours of Sebring 43 cars
     
    patchedupdemon likes this.
  5. DrivingFast

    DrivingFast Registered

    Joined:
    Mar 1, 2018
    Messages:
    1,638
    Likes Received:
    1,083
    Do not focus on the number of AI in my test :

    I do not have the exact number in mind, but I did the same test today with only 30 AI, almost the same settings or maybe exactly the same settings, and my minimum FPS was as I recall at 45 or 48 FPS maximum.

    Frankly this is not enough, it equates to about 53 FPS about average FPS, from what I could see. This is not enough for me.

    I persist in saying that there is a major performance problem in rF2. It is incomprehensible that some users do not realize it.

    More users will defend the alleged "good performance" or "acceptable performance" of rF2, minus the direction of S397 will want to cash back their graphics engine.

    Keep up guys, this is not in our interest.
     
  6. 2ndLastJedi

    2ndLastJedi Registered

    Joined:
    Nov 3, 2016
    Messages:
    1,873
    Likes Received:
    1,198
    I only really need to have a car in front and a car behind so I if I set the AI strength properly I get away with 19 Ai cars and never not have the car in front and behind for any race on any track! Having 80Ai seems excessive imo.
     
    patchedupdemon likes this.
  7. patchedupdemon

    patchedupdemon Registered

    Joined:
    Jan 3, 2017
    Messages:
    1,602
    Likes Received:
    1,151
    If you check my posts I’ve never said rf2 hasn’t got performance issues, butt aa level 5 with the rest at high or max with 30-100 ai in rain also, is expecting too much.

    You know we know, everyone knows rain and ai are FPS killers in rf2, maybe if S397 dumb down the physics and tyres then you could have your magic expected performance

    I’m still waiting for reply to which sim allows 80-100 ai at night with rain and has good performance

    I’d like to make my bmw 330ci make 1200bhp, without rebuilding or swapping the internals,but I know it just won’t happen, being realistic about expectations is the key here that’s all I’m sayin
     
    Last edited: Dec 1, 2018
    Alessio Feletto likes this.
  8. stonec

    stonec Registered

    Joined:
    Jun 19, 2012
    Messages:
    3,399
    Likes Received:
    1,488
    I think it's fair to say that rF2 has performance issues with 60-100 AI, but then again there is a reason why most sims won't let you add more than 20 AI. With these high-poly car models, the number of draw calls etc. increase rapidly with every AI car you add. At certain number of draw calls, the performance starts plummeting even more rapidly, so you might lose relatively more by going from 30 to 40 than 20 to 30 AI. PCI-E limitations might come to play here as well.

    A simple test to understand the huge impact of AI cars is to run an older open wheel model like Formula ISI, which has much less detail per car. In order to run 100 AI, I think either quality of car cockpit or FPS has to suffer. Maybe rF2 should allow an option like in Raceroom to run the AI cars without cockpits.
     
  9. DrivingFast

    DrivingFast Registered

    Joined:
    Mar 1, 2018
    Messages:
    1,638
    Likes Received:
    1,083
    There is an option on the player.json! I will test
     
  10. Will Mazeo

    Will Mazeo Registered

    Joined:
    Sep 12, 2015
    Messages:
    2,220
    Likes Received:
    1,578
    Yeah that's a very old option in isimotor, unless you are going to do a driver swap you can turn cockpits off. Would be nice if they made this work with swap, if cars had a list of drivers that are part of the entry so when we get in the server only that cockpit would load.
    Not sure how this helps fps in rF2. I remember it had a impact in older games.
     
  11. Emery

    Emery Registered

    Joined:
    Oct 24, 2010
    Messages:
    3,035
    Likes Received:
    1,654
    I've been trying to figure out exactly what it is about rain that cuts fps in half. So far the only conclusion I came to is that the harder it is raining, the worse the fps get. When it stops raining, fps come back up to normal, even with lots of puddles.

    Whether track has puddle map or not, it's the same percentage drop. The graphic setting for rain at Off, Low, or Max had very little effect, so using Low at least lets you see the raindrops. fps scales with number of AI at the same rate whether it's raining or not.
     
    Last edited: Dec 2, 2018
  12. Marcel Offermans

    Marcel Offermans Registered

    Joined:
    Oct 4, 2010
    Messages:
    646
    Likes Received:
    2,930
    You do realize that the minimum reported fps is the worst case scenario, right, but since you don't provide full settings nor provide average and maximum values there is no way for me to validate or recommend changes.
     
  13. Filip

    Filip Registered

    Joined:
    Feb 9, 2017
    Messages:
    1,208
    Likes Received:
    937
    Not on my system. I see big difference between rain off and low even on dry weather
     
  14. Emery

    Emery Registered

    Joined:
    Oct 24, 2010
    Messages:
    3,035
    Likes Received:
    1,654
    GTX 1080 here.
     
  15. DrivingFast

    DrivingFast Registered

    Joined:
    Mar 1, 2018
    Messages:
    1,638
    Likes Received:
    1,083
    I also have a GTX 1080.

    For information there is a really noticeable difference between rain high and rain ultra in term of FPS performance.

    It is easily verifiable.

    And I do not agree at all with people who say there is no difference between rain high and rain ultra, and PP high and ultra PP in terms of graphics quality.

    I see a real difference in cockpit view and yet I'm only in full HD ........
     
  16. Emery

    Emery Registered

    Joined:
    Oct 24, 2010
    Messages:
    3,035
    Likes Received:
    1,654
    Woo-hoo! I improved the minimum framerate by over 30%!

    Accidently discovered I was being thermal limited on the GPU, so created a custom fan curve and got a significant improvement with no thermal limiting. Still, my minimum framerate in rain with GT3 AI is unplayably low.
     
  17. DrivingFast

    DrivingFast Registered

    Joined:
    Mar 1, 2018
    Messages:
    1,638
    Likes Received:
    1,083
    GTX 1080, single monitor 60 hz, full HD

    rF2 does not allow to get in all circumstances 60 FPS with a GTX 1080, when the graphics are in FULL ULTRA, with a single screen installation and in Full HD.
    What I call "in any circumstance" is all the cumulative race conditions: a lot of AI, heavy rain and lots of clouds, day/night transitions especially sunrise/sunset/night/low sun.

    Given the price needed to buy this hardware, I thank it is abnormal, and I always think it.

    I said a lot that rF2 was very poorly optimized, that's true and I still think so.
    *BUT* IT IS TO QUALIFY.

    Indeed I did not know a game with almost identical graphics quality, and changing weather conditions.

    I tested Project Cars 2 recently, and I realized that the graphical rendering between rF2 and PC2 was very similar (to my astonishment because I thought PC2 would be very much above), and that Project Cars 2 also did not allow to have 60 FPS stable (in Full HD, single monitor, Full Ultra, unfavorable race conditions [heavy rain, sunset/sunrise ...]).
    I really thought that a game so "big budget" was better optimized and much more beautiful than rF2.

    I followed @Marcel Offermans advice concerning the AVERAGE FPS for running my rF2 performance test :
    I only watched the FPS MINIMUM before, and I think I was wrong, so thank you Marcel.

    I adapted my graphic settings with the following goals :

    1. Really good performance, and at least 60 AVERAGE FPS in the worst game circumstances :
    rain/sunrise-sunset-night-low sun racing conditions, Sebring 12H, 61 AI [max AI on Sebring]).
    2. The most beautiful graphic settings possible when it is not raining and the road is dry:
    (Graphical ingame and post-process settings + Nvidia inspector graphical settings)
    3. Keep a good graphic quality in the rain and on wet roads very good, but make two compromises to keep extremely good settings in all conditions except rain, these settings are also beneficial for the graphics in the rain.
    4. To avoid unpleasant surprises when driving, "a lot" of AI visible.

    TEST (done 3 times to be sure):

    - Sebring 12H
    - 61 AI
    - Grid position: 45
    - 8:00 AM ---> ~11: 00, time scale x5
    - My car: Mclaren GT3, other cars: S397 GTE and 397 GT3
    - Mostly cloudy and 100% rain
    - Important : during the whole test, I always positioned my car in the middle of a group of cars, therefore in the worst performance conditions.

    NVIDIA INSPECTOR GRAPHICS SETTINGS:

    - Maximum pre-rendered frames: 1 (low input lag)
    - AA: x2 Multisampling associated with the very powerful AA x2 Sparse Grid Supersampling (Enhance the application setting): it is thanks to that that the image is always beautiful = important.
    - Anisotropic filter optimization: OFF
    - Trilinear optimization: OFF
    - Texture filtering: HIGH QUALITY (very important for global visual quality)

    INGAME GRAPHICS SETTINGS:

    - Post-Process: ULTRA
    - Full HD
    - AA: forced by nvidia inspector
    - FullScreen mode

    - visible AI: 12
    - AUTO FPS: 60 (really help)
    - Mirrors: ON + 1 virtual rearview
    - FOV: 47

    https://www.dropbox.com/s/3phb8axhehk35r2/ingame rF2 settings.PNG?dl=0
    [​IMG]

    - ***IMPORTANT, THE GRAPHIC SETTINGS THAT I LOWERED:***
    Opponent: medium, Soft Particles: low, rain: medium


    GTX 1080 OVERCLOCKING

    Allowing to win about 5.5% of FPS.

    RESULT OF THE TEST (3 times the same test):

    1. A good performance even in difficult race conditions, and perfect in all other race conditions.
    - AVERAGE FPS after 30 minutes: 64-67 (according to the 3 tests carried out).
    - MINIMUM FPS: 44-45
    - MAX FPS: 94-96.
    2. Excellent image quality in all circumstances (AA x2 sparse grid, etc.)
    3. Only in the rain rF2 does not have ultra graphics.
    During wet and rainy conditions, the loss in visual quality is actually acceptable, because the difference is not very visible when driving.

    I strongly recommend these settings to owners of GTX 1070/1070 ti/1080/1080 ti and people who have a single Full HD screen.
    Everything in these settings to its importance.
     
    Last edited: Dec 8, 2018
    Capeta, pkelly and Juergen-BY like this.
  18. 2ndLastJedi

    2ndLastJedi Registered

    Joined:
    Nov 3, 2016
    Messages:
    1,873
    Likes Received:
    1,198
    What CPU is used here? Also is Sparse grid AA really required? I find in game AA to work really well even just Level 3 is far better than all other Sims maxed out AA imo.
     
  19. DrivingFast

    DrivingFast Registered

    Joined:
    Mar 1, 2018
    Messages:
    1,638
    Likes Received:
    1,083
    With a big screen of 132 cm (52 "), full HD not 4K, located 75 cm from my eyes only, the x2 sparse grid is the beginning of a beautiful image.
    The level 5 is insufficient.

    Especially shadows are NOT Aliased with x2 sparse grid.

    Little i7 4770k NOT OC.
     
  20. Alex72

    Alex72 Registered

    Joined:
    Dec 27, 2016
    Messages:
    1,317
    Likes Received:
    992
    Is GTX980 "mid end"? Also if its on triple screen maybe it turns into "low end"? :D
     

Share This Page