Problem Upgrade from 1050Ti to 1660S and same FPS !

Discussion in 'Community Support' started by TheJackal, May 18, 2021.

  1. TheJackal

    TheJackal Member

    Joined:
    May 4, 2020
    Ratings:
    +12 / 0 / -0
    My old Nvidia 1050Ti used to net around 50-80fps depending on the track and number of cars on circuit at 1080p. Graphics settings were custom of course with shadows turned down all the way and no motion blur which if I'm not mistaken are the things that most hinder FPS.

    I recently upgraded to a Nvidia 1660 Super and I was expecting to be somewhere in the 120+ fps range but, there's barely any difference.

    You're probably thinking "its your CPU that's the bottleneck" but, the Xeon W3680 I've got this paired to averages 30-40% usage.

    Needless to say, fire up Assetto Corsa and FPS is 120+ with everything maxed out.

    Can anything be done ?
     
  2. Vale

    Vale Well-Known Member

    Joined:
    Jul 4, 2019
    Ratings:
    +279 / 0 / -0
    I get between 80 and 120 FPS with my 1070M and 144hz screen and your desktop card should be about 20% faster than mine.
    You could switch off time of day transition in single player mode and put shaders and reflections to lower settings as well as trackside animations and see if that boosts it a little. Vsync off too, obviously if you are running a higher than 60hz monitor. Virtual mirror and any other stuff like 3rd party car telemetry will also cause an FPS hit.

    Also compare Nvidia drivers as the newer ones might not be as optimised for this game or your card.
     
  3. TheJackal

    TheJackal Member

    Joined:
    May 4, 2020
    Ratings:
    +12 / 0 / -0
    Thanks!

    I'm running a 165hz screen, that's why I was aiming to squeeze a bit more FPS out of the whole rig.

    Vsync is off already but I'll try the time of day transition and turn down the shaders and reflections to see what difference it makes. Still can't help thinking R3E has a fundamental graphics engine problem because ACC is in another league when it comes to FPS.

    As for the drivers, how do I know which version is best for the game/card?
     
  4. Maskerader

    Maskerader Well-Known Member

    Joined:
    Oct 6, 2019
    Ratings:
    +356 / 0 / -0
    These numbers aren't really helpful because R3E can only use one core. According to Passmark your CPU is not very powerful in single-core use: https://www.cpubenchmark.net/cpu.php?id=1274

    So your CPU might very well be the main bottleneck.

    Which also means you can actually crank up your graphics setting (those that are GPU-related) without tanking your FPS.
     
    • Agree Agree x 3
  5. TheJackal

    TheJackal Member

    Joined:
    May 4, 2020
    Ratings:
    +12 / 0 / -0
    But if single core CPU occupancy is between 30-40% I don't see how it can be a bottleneck. My CPU is not maxing out.
     
  6. Maskerader

    Maskerader Well-Known Member

    Joined:
    Oct 6, 2019
    Ratings:
    +356 / 0 / -0
    That's single core? Yeah, that doesn't look right.
     
  7. Vale

    Vale Well-Known Member

    Joined:
    Jul 4, 2019
    Ratings:
    +279 / 0 / -0
    That´s interesting. Benchmark single core performance for your chip is indeed worse than even on my old beater laptop so that may well be the issue with RRE.I am guessing you are using DDR3 RAM from the age of the chip so that could be another factor.
    https://www.cpubenchmark.net/compare/Intel-Xeon-W3680-vs-Intel-i5-3210M/1274vs815

    Your multicore performance is much better so that´s maybe why it is not an issue with Assetto. My last 2 core i5 was always getting the in game processor 100% use warning on Assetto but ran RRE adequately.

    How are you measuring single core occupancy?

    You could try the DX9 benchmark at the below thread and compare with the other results we have from different chips and GPUs.
    https://forum.sector3studios.com/index.php?threads/directx-9-cpu-benchmark-thread.13473/

    If single core performance is the issue then there is not much else you can do apart from overclocking or forcing the game to run on another core just to see if one core performs better than the others. IIRC from GTR2 days, process lasso was the app that allowed you to manage that.
     
    Last edited: May 19, 2021
  8. TheJackal

    TheJackal Member

    Joined:
    May 4, 2020
    Ratings:
    +12 / 0 / -0
    Thanks, I'll run the benchmark to see what kind of results I get.

    I wouldn't shy away from a hardware upgrade, what kind of CPU would I need paired to my 1660S to get 140-160fps from RaceRoom? This might be a stupid question considering R3E is DX9 and single-core....
     
  9. Vale

    Vale Well-Known Member

    Joined:
    Jul 4, 2019
    Ratings:
    +279 / 0 / -0
    I wouldn´t think going above a steady 90 or so FPS is going to make much difference to the experience and stability is more beneficial. I get steady 80FPS on AMS2 at higher detail than RRE and the feeling of speed is good except for using chasecam, which is dreadful on Madness engine.
    The list ordered by single core performance will say a recent i9K or i7K or Ryzen 5900/800 but then your new graphics card will be the bottleneck again!
    https://www.cpubenchmark.net/singleThread.html

    If Intel do release the 12th gen this year then prices of the 11th will collapse so it might be worth making do with what you have till September/October.
     
  10. s9khenrik

    s9khenrik New Member

    Joined:
    May 16, 2020
    Ratings:
    +0 / 0 / -0
    To determine if you have a CPU bottleneck is very simple.
    1. Make sure vsync or any other form of FPS limiter is turned off
    2. Measure GPU usage instead of CPU usage. If your GPU usage is below 98-99% you have a CPU bottleneck
     
    • Agree Agree x 1
  11. Ablaze

    Ablaze Well-Known Member

    Joined:
    Mar 16, 2018
    Ratings:
    +120 / 0 / -0
    If unsure, use MSI Afterburner with RTSS to get an overlay ingame and look if your GPU usage is at 100% or not. If yes, you've hit the GPU limit. If not, it's your CPU. Simple as that. :)

    Here's an overview how different CPUs performs under DX9:
    DirectX 9 CPU Benchmark Thread
     
    • Agree Agree x 1
  12. TheJackal

    TheJackal Member

    Joined:
    May 4, 2020
    Ratings:
    +12 / 0 / -0
    Benchmarks are in and the score is.....underwhelming. What gives?

    1050Ti

    dx9_Benchmark_1050ti.PNG


    1660 Super

    dx9_Benchmark_1660S.PNG
     
    Last edited: May 19, 2021
  13. Andi Goodwin

    Andi Goodwin Moderator Beta tester

    Joined:
    Jan 16, 2015
    Ratings:
    +791 / 0 / -0
    I would see if your cpu takes an overclock; just using the multiplier in the bios to 4ghz; this will give you a few extra fps

    i take no responsability for it going boom

    Andi
     
    • Funny Funny x 1
  14. TheJackal

    TheJackal Member

    Joined:
    May 4, 2020
    Ratings:
    +12 / 0 / -0
    I does accept an overclock via Intel's XTU, this is a workstation (HP Z400) no none of those fancy BIOS things :p Does have a self-contained watercooling system though :)

    It can do 4.0Ghz on a single core and I'll test it over the weekend by using process lasso to assign RRE64 to a single core and then overclock that particular core.
     
    • Like Like x 1
  15. Maskerader

    Maskerader Well-Known Member

    Joined:
    Oct 6, 2019
    Ratings:
    +356 / 0 / -0
  16. Maskerader

    Maskerader Well-Known Member

    Joined:
    Oct 6, 2019
    Ratings:
    +356 / 0 / -0
    Well, no matter what hardware you have and what programs you run, there will always be a bottleneck somewhere. I wouldn't say 1660 Super will drag your system down all that much, it's a decent GPU.
     
  17. Vale

    Vale Well-Known Member

    Joined:
    Jul 4, 2019
    Ratings:
    +279 / 0 / -0
    Yes. My point was really that it is a vicious circle as there will always be something capping the system and throwing money at the issue to get a super high FPS, which I doubt would be noticeable to the naked eye anyway, is maybe not worthwhile just for the one game. Especially when there are no guarantees that the game will stay on the same graphics engine anyway.
     
  18. Maskerader

    Maskerader Well-Known Member

    Joined:
    Oct 6, 2019
    Ratings:
    +356 / 0 / -0
    I really don't think it's applicable here.

    1) TheJackal's CPU is really old and is basically on its last legs for modern usage. It's totally reasonable to upgrade it to something newer, no matter the R3E performance.
    2) An improvement from 80+ FPS to something better is neither super-high nor unnoticeable. If you have a 165Hz monitor and a GTX 1660S you can expect and reach much higher numbers and a smoother picture than barely scraping out half of the refresh rate at lower game settings.
    3) A newer CPU that has a single-core performance good enough for R3E isn't incredibly expensive, you aren't throwing money away. Although the current hardware situation makes PC upgrades a bit more complicated...
    4) An upgrade like this is not only for one game, many other games would benefit from that too.
     
    • Agree Agree x 1
    Last edited: May 20, 2021
  19. TheJackal

    TheJackal Member

    Joined:
    May 4, 2020
    Ratings:
    +12 / 0 / -0
    Thanks for everyone's input!

    The next question is, what sort of CPU would be needed to get 120+ fps and even if the 1660S is capable of this in R3E.

    A CPU like an i5 10600K isn't out of reach with the associated other bits (motherboard + ram + cooler + case).
     
  20. Nir Tal

    Nir Tal Well-Known Member

    Joined:
    Aug 29, 2016
    Ratings:
    +53 / 0 / -0
    The latest AMD 5X00 is the most powerful single-core performance CPU. If u can get the 5600X for 300$, then its the best performance for the money.
    I had Intel-i5-2500 (which has same single-core performance as ur XEON) with Nvidia-GTX1060 (which is a bit less powerful than ur 1660) - and i was limited with FPS and graphics settings (was around 60FPS with low-med settings). Then i upgrade to the new AMD 5600X and didnt upgrade the GPU - now i get 150-200FPS with high graphics settings.

    BTW, RRE is using about 2.5 cores, not only one, and thats the reason u see CPU usage as 30-40% - 2.5 out of 6 cores u have.