Next Gen Simulation

Discussion in 'General Discussion' started by vegaguy5555, Feb 3, 2018.

  1. vegaguy5555

    vegaguy5555 Registered

    Joined:
    Jan 28, 2017
    Messages:
    960
    Likes Received:
    505
    I am asking this question not to complain but to feel out everyone's expectations, including the programmers here at Studio-397.

    The question is, what is the next gen sim going to look like? If computers had 128 bit processors what added features can we look forward to as far as physics are considered.
    Lets leave GPU's and 4K/VR out of this for now but take a good look at what is missing from physics instead.

    There is two reasons I ask. One is I'm just not sold 100%. And second is remembering back to when AC updated from 32 to 64 bit was really noticeable. There was just so much more I could do in the corners with 64 bit and I can't help think what 128 bit would feel like, hence the question?

    So please, tell us guys at Studio-397, what would you like to do if you had the money, power, time and future cockpit designs to work with?
     
    Jokeri likes this.
  2. Louis

    Louis Registered

    Joined:
    Dec 27, 2016
    Messages:
    1,379
    Likes Received:
    840
    Championship mode in single player and new UI
     
  3. stonec

    stonec Registered

    Joined:
    Jun 19, 2012
    Messages:
    3,399
    Likes Received:
    1,488
    32 vs 64 bit makes no difference on the actual physics that are run, not sure where you got this impression from, because I can assure that's wrong. The app is basically same in both 32-bit and 64-bit, a simple app can even be converted with a few clicks from 32-bit to 64-bit. rF2 still allows running in 32-bit mode today and it doesn't make a difference in lap times. The only reason most apps today are 64-bit is due to 4 GB of addressable memory limit of 32-bit (32-bit rF2 crashes if you try to run it with something like 30 GT3 cars on max graphics). If it weren't for this, rF2 would probably still run 32-bit today.

    Same with 128-bit, no advantage except for some bigger registers in CPU, but this is all on assembler level code. The numer of bits isn't the limiting factor, it adds nothing to the performance. The limiting factor is single core performance in CPU and the degree of parallelism you can achieve in simulation.
     
  4. Will Mazeo

    Will Mazeo Registered

    Joined:
    Sep 12, 2015
    Messages:
    2,220
    Likes Received:
    1,578
    For next gen sim I'd like something that actually simulates the alive environment of a real life race event. On the physic aspec they got many things right, just add gearbox damage and hybrids and you are done.
     
    TOCA2FREAK and The Iron Wolf like this.
  5. momoracer

    momoracer Registered

    Joined:
    Dec 28, 2017
    Messages:
    110
    Likes Received:
    24
    I think it's all about VR.

    If we can get high definition VR with huge field of view, eye tracking etc, we will basically be in a racing car immersion wise.

    I believe we will eventually get gloves that render our hands in VR too, but this could lead to some problems... since wheel sizes and shapes vary. Also trying to use buttons would still be a problem, but at some point VR will be too good to use triples.

    Graphics wise, I would like to see rFactor 2 look as good as AC without compromising on the handling... that's all (of course, with improved lightning system, AC sucks lightning wise, no night mode, no proper shadows in tunnels, what a joke.. but the rest is awesome)

    Improved and affordale motion simulators in the next 2 to 5 years could change the game too.
     
  6. dadaboomda

    dadaboomda Registered

    Joined:
    May 15, 2016
    Messages:
    712
    Likes Received:
    202
    For the next generation I want two things, what I want for RF2 at this time in fact :

    - PERFORMANCE :
    See a report quality graphic / performance very good for 95% of PC users. Users with a low power PC can enjoy acceptable graphics in ALL game conditions (many AI, wet road, rain, headlights on, low sun in the sky, mirrors etc.) : in fact, as in any recent game.

    - REALISM :
    Want that all the aspects of realism are very close to the actual RFPRO, which can be taken as a reference, which means: * Physics very advanced and more precise.
    - Logically, to exploit the physics of high quality, circuits laser scanned with high precision, and transcribed with a lot of details in the simulation.
    - Physics of the rain and especially of the wet road very evolved (aquaplanning, real puddles according to the physics of the track etc.)
    * Car contents : fully integrating all the latest technologies of physics + 100% real data of cars (as for circuits with the laser). * Real road: more evolved than today, much more precise in term of physics, management of the wind with a precise interaction on aerodynamics and realroad, all temperature ranges managed very precisely.

    This will necessarily improve in the next generation, but is totally secondary :

    Graphics / ergonomics (beautiful UI) / visual effects (car damage, rain drops visible on the AI .......) and so on.

    As you can see, what I want for RF2 at the moment is what I want for the next generation :

    Realism realism realism / good quality-performance ratio for JUST driving under normal conditions .. = to be close to RFPRO...

    That is my wish for RF2 or the future RF3.... but yes, RF2 has taken another direction for over a year. So sad.
     
    Suzukinol likes this.
  7. Comante

    Comante Registered

    Joined:
    Nov 20, 2013
    Messages:
    1,694
    Likes Received:
    1,218
    I don't see a feature that require a new hardware architecture, proof is that physic simulation run at a fraction of the performances of actual CPU. Well unless they would start to calculate airflow in real time, or start to simulate all vehicle physic as an actual physic simulation, where suspension working are not numbers on a table but actual moving parts interlinked eachother. But the question is would that level of simulation useful or better than what we have today?
     
  8. patchedupdemon

    patchedupdemon Registered

    Joined:
    Jan 3, 2017
    Messages:
    1,602
    Likes Received:
    1,151
    I doubt the improvements are linear,you get to a point where things become much more difficult to om
    Just to be clear,are you saying the suspension physics isn’t derived from multi body physics
     
  9. Comante

    Comante Registered

    Joined:
    Nov 20, 2013
    Messages:
    1,694
    Likes Received:
    1,218
    For sure they are not calculated in real time piece by piece, as there is no purpose in doing it, for the simple reason that suspension movement are easily predictable : for each load condition there is a very precise position of the tire in respect to the car body, no need to calculate continuously to achieve a predictable result.
     
    patchedupdemon likes this.
  10. patchedupdemon

    patchedupdemon Registered

    Joined:
    Jan 3, 2017
    Messages:
    1,602
    Likes Received:
    1,151
    Thanks for replying,i only asked because someone on another sim forum said he thought rf2 was a joke because of what you described above,he said without multi body physics what’s the point in simulating anything lol.

    He didn’t like it when i said that cars still handle better in rf2 and look more realistic when they interact with the track and curbs etc than that other sim
     
  11. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    The limitation is always in the number of dof that need to be calculated for each iteration that determine the matrix of equations to be solved. A realtime process has to balance time step with number of unknowns.

    You have to carefully balance how you model the different parts of the car. There are some evident dofs like wheel rotation suspension or steering. However when it comes to deforming parts it is not so straightforward. The chassis, aero elements and tires deform in a more complex way.

    In order to determine which area of the simulation would require further advances you would have to compare real and virtual cars telemetry in controlled tests. Those should the engine's strengths and weaknesses. Probably, by now the car and physics team have enough feedback on this. The relationship of that team with rfpro and real teams feedback I understand exists.

    An alternative and more independent approach would be to have a complete and detailed model where even the tyre model was being simulated in the overall car simulation. Its drawbacks would be that it would not run in realtime and the inputs should be either offline (taken from the simplified model to be adjusted) or from a driver that gets used to driving in slow motion. AI intelligence could be useful for this task since it doesn't understand about time.
     
    Last edited: Feb 4, 2018
  12. AMillward

    AMillward Registered

    Joined:
    Apr 4, 2017
    Messages:
    1,879
    Likes Received:
    1,837
    You'll need to fork out for a decent processor then if you want super high res physics.
     
    GrimDad likes this.
  13. dadaboomda

    dadaboomda Registered

    Joined:
    May 15, 2016
    Messages:
    712
    Likes Received:
    202
    Yes.
     
  14. AMillward

    AMillward Registered

    Joined:
    Apr 4, 2017
    Messages:
    1,879
    Likes Received:
    1,837
    I should have said a VERY decent processor.
     
  15. dadaboomda

    dadaboomda Registered

    Joined:
    May 15, 2016
    Messages:
    712
    Likes Received:
    202
    My point, hope you understand more after read this :

    All the other "simulations" do not pose any problems of performance to my PC with often better graphic quality (with my relatively old CPU, PCIe X2.0, my gtx 780).

    For information, that is just a fact (I try many car games this 6 month).

    Really there is a huge gap between rf2 and ... all the other car games.
    It is not optimized, it is not normal. I think it's not up to the user to change all the components of his PC (it's very very expensive, you know ...), because the software is not properly optimized.

    I'm talking about the consequences of the bad graphics engine (sorry S397, that's just right).

    On the other hand, if the physics of the tracks + cars + the realroad / wind / moving ground / aquaplanning / real puddles according to the physics of the track / temperatures etc are strongly improved, precise, detailed = I buy myself everything immediately a very high-end processor with a good motherboard.

    But you understand that, given the extremely high price of a good PC, and the obligation that I have to change everything except the SSD (minimum 1770 euros, more expensive than my 3 cars when I had 18-20 year), I will not pay such a sum because we has an editor incredibly good in the essentials (realism etc.) but extremely bad on other aspects (graphic engine for example).

    I hope you understand my reasoning, it just seems logical (Otherwise it does not matter, we will not understand each other ;) ).

    The 1770 euros, I prefer to keep them for my wife and my son.
     
    ericRacer likes this.
  16. AMillward

    AMillward Registered

    Joined:
    Apr 4, 2017
    Messages:
    1,879
    Likes Received:
    1,837
    And like you've been told several times by other users, including Marcel and Chris: They're actively working on it and it's going to take time. The physics, at this moment in time, don't need work. The graphics optimisation does.

    Your computer will run AC maxed out, but not RF2. And that's like saying "I can run minecraft maxed but not Fallout 4! They're both open world games!"

    But then, my £500 (550 euro) rig is running rf2 just fine (FX6300, 1050ti, 8gb RAM which, at least in terms of processor is being run by @krusti and others) and it's not a space shuttle by any means and I ran 30 cars at Imola at the weekend no issues at 60fps in 1080. And that was recorded too.
     
  17. Daniele Vidimari

    Daniele Vidimari Registered

    Joined:
    Apr 20, 2016
    Messages:
    659
    Likes Received:
    593
    @AMillward i have now an i7 6700k here ready to be mounted :p anyway yes i confirm what you said, seems strange but rF2 run smooth as silk with the FX 6300.
     
    AMillward likes this.
  18. buddhatree

    buddhatree Registered

    Joined:
    Dec 22, 2011
    Messages:
    1,700
    Likes Received:
    249
    ^ This is all I've ever wanted since 2011 ^ :)
     
    Louis likes this.
  19. stonec

    stonec Registered

    Joined:
    Jun 19, 2012
    Messages:
    3,399
    Likes Received:
    1,488
    FX 6300 is limited to PCI-E 2.0, so you should see notable gains with i7 6700k, especially at race starts.
     
  20. AMillward

    AMillward Registered

    Joined:
    Apr 4, 2017
    Messages:
    1,879
    Likes Received:
    1,837
    30ai at Imola at the weekend no issues at 60fps. Seems odd really.

    @krusti let me know how that goes for you, I might have to upgrade mine to a Ryzen or similar by the end of the year.
     

Share This Page