DirectX 9 CPU Benchmark Thread

Discussion in 'General Discussion' started by Thomas Jansen, Jul 22, 2019.

  1. 007ven

    007ven New Member

    Joined:
    Jan 5, 2022
    Ratings:
    +0 / 0 / -0
    Hi all,
    inspired by this thread and looking for optimization, I run the benchmark for my system and was a little bit shocked by the poor performance I got compared to the other results with similar hardware here. This may also explain the variable results I get in the game. So maybe you guys have an idea what might cause this?

    Here are my specs and results:
    • CPU: Intel Core i7-9700K
    • CPU OC: @4,9GHz all core
    • Memory: 32GB DDR4-3600 MHz
    • Memory timings: CL16-18-18-38
    • Mainboard: Asrock Z390 Phantom Gaming-ITX/ac
    • GPU: Nvidia GeForce RTX 2070 Super (EVGA RTX 2070 Super Black Gaming)
    • Display: 3440x1440@ 144 Hz on Desktop
    • OS: Windows 10 Pro
    • nVidia Driver: 497.29
    DX9-Unigine-result.JPG CPU-Z1.JPG CPU-Z2.JPG CPU-Z3.JPG
    GPU-Z.JPG
     
  2. Wuschel

    Wuschel Member

    Joined:
    Oct 16, 2015
    Ratings:
    +6 / 0 / -0
    Here's my run if it helps to gather some more data. :D
    • CPU: AMD Ryzen 9 3900X
    • CPU OC: Stock (was stable around 4.25 GHz on the used cores)
    • Memory: 32GB DDR4-3600 MHz Dual Channel
    • Memory timings: CL16-19-19-39
    • GPU: Nvidia Geforce RTX 2070 Super
    • Motherboard: MSI B450 Tomahawk Max
    • OS: Win 10 Pro
    • Nvidia Driver: 496.49

    [​IMG] [​IMG]
     
  3. Nir Tal

    Nir Tal Well-Known Member

    Joined:
    Aug 29, 2016
    Ratings:
    +53 / 0 / -0
  4. Vale

    Vale Well-Known Member

    Joined:
    Jul 4, 2019
    Ratings:
    +278 / 0 / -0
    Has anybody got a 12th gen intel chip yet? If so, how does it perform on RRE and other driving games?
     
  5. 007ven

    007ven New Member

    Joined:
    Jan 5, 2022
    Ratings:
    +0 / 0 / -0
    Thanks for the tip, but it looks like it was my own stupidity...
    I checked my results again and again and then suddenly I found out that the max FPS was always around 144, which is the refresh rate of my monitor. As I had G-Sync enabled, I disabled it and run the benchmarks again. Now the results seem to be in the right ballpark. I feel so stupid now :expressionless::neutral::)

    Here are the new results:
    Intel Core i7-9700K @4,9GHz all core
    DX9-Unigine-result4-oc49.JPG

    Intel Core i7-9700K @stock settings
    DX9-Unigine-result4-stock.JPG
     
  6. nolive721

    nolive721 Well-Known Member

    Joined:
    Dec 2, 2018
    Ratings:
    +72 / 0 / -0
    Hello
    I am participating to this Thread for the 1st time despite being long time user of RRE.

    Sorry for the long post but during the weekend I decided to test some CPU/GPU monitoring on a couple of my SIMs titles like RRE
    I play on UW 1440p and being a bit focused on power efficiency(might be because of living in Japan with high Energy bill....), I have CPU and GPU UVed and UCed with the goal to reach min 60fps in Game.
    My monitor is LG with Freesync Pro so freesync range is 48-160Hz and also with LFC feature so I am having butter smooth experience even when frames drop below the 60fps threshold

    I also tend to not only underclock/undevolt my CPU and GPU but also set Windows power mode to Power saver so to maintain the CPU clock around 1.7Ghz and Core voltage around 0.85V

    My GPU is UVed at 1V and UCed at 220Mhz core, its giving me around 35W CPU power usage and 75W for the GPU so my all system pumps only 120-130W maximum

    What I found during these testing is following, I am curious about what you guys could comment here

    1. RRE would ignore the Windows Power saver setting and get the CPU to default 4.1Ghz and Core voltage 1.3V (I don't like that so much....)
    2. I was under the impression that RRE was heavy on single core CPU use but I noticed that actually 3 cores are heavily loaded during Gaming, being practice with 16cars on track or Race with 32 cars on tracks
    specific to the purpose of the Thread, here are my Benchmark results (I might do another run with CPU and GPU OCed just for fun later)
    • CPU: AMD Ryzen 5 3500X (6C/6T)
    • CPU OC: Stock (was stable around 4.1 GHz on the used cores)
    • Memory: 16GB DDR4-3433 MHz Dual Channel
    • Memory timings: CL17-18-18-36
    • GPU: AMD 6600XT Nitro+
    • Motherboard: ASROCK B550M Steel Legend
    • OS: Win 10 Pro
    • AMD Driver: 2022.02

    upload_2022-2-27_19-20-14.png


    upload_2022-2-27_19-21-22.png
     
    Last edited: Feb 27, 2022
  7. Maskerader

    Maskerader Well-Known Member

    Joined:
    Oct 6, 2019
    Ratings:
    +352 / 0 / -0
    What about other games?
     
  8. nolive721

    nolive721 Well-Known Member

    Joined:
    Dec 2, 2018
    Ratings:
    +72 / 0 / -0
    Rf2 and AC without SOL activated leave that setting alone

    Need to check AMS2
     
  9. Balrog

    Balrog Well-Known Member

    Joined:
    Apr 10, 2015
    Ratings:
    +466 / 0 / -0
    Not single core, but single thread, which is not the same thing. The latter means that the tasks required for running RaceRoom cannot be executed in parallel. BUT, that doesn't mean that it will run on a single CPU core, because the Windows scheduler will spread the timechunks of the job on the CPU cores. It just means that for a given time chunk, we need the result of the previous one, so the CPU can not work on them in parallel. (But it can be that the previous chunk was processed on a different core compared to the current one.)
     
    • Informative Informative x 2
    • Like Like x 1
  10. Giovanni Giorgio

    Giovanni Giorgio Member

    Joined:
    Feb 28, 2022
    Ratings:
    +6 / 0 / -0
    Hi, first time posting. I did the benchmark, but I think it's deprecated. I'll explain. I had a Ryzen 5800X before and played Raceroom with it in VR. In certain scenarios it would struggle to keep up in VR if there was too many cars on track.

    I found out about this thread today and decided to share something interesting.
    I did not benchmarked the 5800X using the Unigine software unfortunelly, but I swear the performance of my new CPU, which is the Ryzen 7 5800X3D stomps on any other CPU I've seen running 40+ cars on Nurburgring, specially while watching the race replay. Here is my result:

    • CPU: AMD Ryzen 7 5800X3D
    • CPU OC: 4.450 GHz
    • Memory: 32GB DDR4-3733 MHz Dual
    • Memory timings: CL18-22-22-42
    • GPU: Nvidia Geforce GTX 1080ti
    The extra L3 cache really did something to the sim. It runs much better compared to my previous 5800X@4.7Ghz. That's what I meant with "I think the benchmark is deprecated" because the score I got on Unigine does not reflect the real performance I had with this new CPU compared to my previous in the sim, not even close. This is the first time I can race with I think as many cars as I want it on this game without having performance penalty in VR.
     

    Attached Files:

    • Like Like x 1
  11. Maarten

    Maarten Member

    Joined:
    Apr 15, 2019
    Ratings:
    +23 / 0 / -0
    True, the benchmark performance does not exactly translate to game performance.
    There is no in-game benchmark unfortunately which would make comparing hardware a bit easier.

    I wouldn't expect the new 5800X3d to be a lot better in r3e considering the lower frequency which usually is the deciding factor. Interesting.
     
  12. Giovanni Giorgio

    Giovanni Giorgio Member

    Joined:
    Feb 28, 2022
    Ratings:
    +6 / 0 / -0
    It's much faster. I wish I could post links here, but as my account is new, the forum does not allow me to do it.
     
  13. Giovanni Giorgio

    Giovanni Giorgio Member

    Joined:
    Feb 28, 2022
    Ratings:
    +6 / 0 / -0
    I think R3E results are similar to Starcraft 2 is this regard. Both are DX9 games, very CPU limited, mostly single core, and both make a high ammount of drawcalls. Here is the result of Starcraft 2 running on both the 5800X and the 5800X3D.

    [​IMG]

    I did see other people on reddit posting similar results for other simulators, like MS Flight Simulator 2020, iRacing and also Assetto Corsa Competizione. The difference is night and day.
     
    • Like Like x 1
    Last edited: Apr 29, 2022
  14. nolive721

    nolive721 Well-Known Member

    Joined:
    Dec 2, 2018
    Ratings:
    +72 / 0 / -0
    I
    I moved since I posted here to a 3700x so 8 cores 16threads with featured SMT as opposed to my previous CPU

    I am a bit obèssed with reducing my PC power consumption and I notice a couple of things running the Game in Windows power saver mode

    the cpu clock now remains at 1.7ghz
    I am getting massive stutters with 30Ai cars
    GPU utilization goes from 50 to 70%

    I opened a thread to report this but people mentioned that I reached the limits of what min cpu click can give in this Game so I am realizing finally it’s not very well optimized

    again same scenario in RF2 in AC and I checked also AMS2 allows me to run 30AI races smoothly with minimum power draw
     
  15. Giovanni Giorgio

    Giovanni Giorgio Member

    Joined:
    Feb 28, 2022
    Ratings:
    +6 / 0 / -0
    I just would like to report that I've completed a 3 lap race with my Ryzen 7 5800X3D in VR using an Oculus Rift CV1 on Spa with 52 visible cars on track, and the framerate never dip below the default ASW threshold. Completely smooth performance.

    I did this race because I've read on race department about someone saying that this race was like a CPU benchmark for Raceroom.
    Is there anything that hits harder than that on the CPU, so I could test it?
    On the options menu, 52 visible cars is the maximum allowed. The only track I can think of that would be another candidate for a CPU stress test would be "the green hell" Nordschleife. Someone please correct me if I'm wrong.
     
    • Like Like x 1
  16. Giovanni Giorgio

    Giovanni Giorgio Member

    Joined:
    Feb 28, 2022
    Ratings:
    +6 / 0 / -0
    I followed an advice from two guys that said to me on another thread I should try Nords with 99 cars for the ultimate CPU stress test in this game.

    Well... 24h layout of Nords, 99 cars with 52 visible, in VR! And it never droped below the ASW threshold. No stutter, no freezing or crashing. It was effing amazing!
    And the best part: it didn't cost me a kidney to make the upgrade to the 5800X3D, unlike it would be if I were building let's say an Intel Alder Lake build from scratch. I sold my 5800X and I have zero regrets. This 5800X3D is out of this world.

    Sorry for the double post.
     
    • Informative Informative x 2
  17. Maskerader

    Maskerader Well-Known Member

    Joined:
    Oct 6, 2019
    Ratings:
    +352 / 0 / -0
    This is absolutely impressive.

    BTW, was time progression on or off?
     
  18. Giovanni Giorgio

    Giovanni Giorgio Member

    Joined:
    Feb 28, 2022
    Ratings:
    +6 / 0 / -0
    Time progression was off. I tried to minimize any variables that could interfere even more with the "impossible" race 'cos I thought it was not possible to accomplish it. So I just set the Ai to a fixed 100 difficulty, chose Nords 24h layout, filled it up with 99 cars and off I went.
     
  19. Balrog

    Balrog Well-Known Member

    Joined:
    Apr 10, 2015
    Ratings:
    +466 / 0 / -0
    That sounds too good to be true.:D I've also read that the upcoming Zen 4 generation won't feature 3D V-cache technology, maybe the next one after that, so it will be hard to decide what to upgrade to.
     
  20. nolive721

    nolive721 Well-Known Member

    Joined:
    Dec 2, 2018
    Ratings:
    +72 / 0 / -0
    Not sure because YTers also mentioned a 3D or multilayer architecture could be used on next Gen coming this year

    so there is hope for CPU no?