Live Performance Benchmarking Comparison for rFactor 2

Discussion in 'General Discussion' started by DrR1pper, Sep 29, 2014.

  1. MrPix

    MrPix Registered

    Joined:
    Jan 10, 2012
    Messages:
    249
    Likes Received:
    2
    The results are in for this live benchmark from me:

    Intel i7 3820@ 4.3GHz
    rFactor 2 Build: 860
    Graphics driver version: 344.16

    GTX 980 @ 1500 core / 1752 mem

    PCI-E @ 2.0 x16
    Time: 134407ms - Avg: 121.459 - Min: 100 - Max: 151

    PCI-E @ 3.0 x16
    Time: 134329ms - Avg: 164.142 - Min: 119 - Max: 181

    So at 1080p, (no scaling applied on my 4k monitor, so it's the bit in the middle) PCIe speed makes a difference... 20% on min/max and a whopping 33% on average.

    Right, enough of this futile 'on-air' stuff... my EK waterblock arrived today, so time to add the 980 back in to the chilled loop and get some clocks shifted in the right direction :D mwahahaha
     
  2. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,294
    Likes Received:
    36
    Oh wow! Fantastic! Thank you and your friend so much for taking the time to get us these results, truly! :D

    So this whole time, even my GTX 770 was underperforming, lol.

    I wonder where the cut off line will be, if there is one, lol. Perhaps someone with a GTX 670 can perform the same test and see if it's also affected?
     
  3. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,294
    Likes Received:
    36
    Wow, very nice as well. Thank you for doing the tests MrPix! :D
     
  4. Marek Lesniak

    Marek Lesniak Car Team Staff Member

    Joined:
    Oct 6, 2010
    Messages:
    1,585
    Likes Received:
    101
    @DrR1pper
    Which line is considered as the "finish line"? The chequered one, just after the last corner or the latter, located at about middle of the main straight? Difference between them is about 5% in min fps and about 1fps in avg fps (earlier line gets better fps).
     
  5. Spinelli

    Spinelli Banned

    Joined:
    Jan 28, 2012
    Messages:
    5,290
    Likes Received:
    32
    The latter. If you are using a 100% default player file for the benchmark then you will be running with 3 HUDs on your screen. 1 of them shows circles going from left to right. As soon as it's at the end of the right, it will disappear and start from the far left again, the moment this happens is the moment you started a new lap. This will show you that the 2nd, further-up line is the start line :)
     
  6. MrPix

    MrPix Registered

    Joined:
    Jan 10, 2012
    Messages:
    249
    Likes Received:
    2
    @DrR1pper

    Just notice my time are for full laps... e.g. out of the pits... past the old finish line (after woodcote) and before you go through maggot and becketts and across the line... which is just over 2 mins.... your times are just over 1 min... which is out of the pits and across the old finish line (half a lap to go).... not sure if it's within the 10 pages about this... as I haven't the time to read it all... but LesiU has the same half lap times too.
     
  7. Marek Lesniak

    Marek Lesniak Car Team Staff Member

    Joined:
    Oct 6, 2010
    Messages:
    1,585
    Likes Received:
    101
    @MrPix
    You must use the short layout (International).
     
  8. MrPix

    MrPix Registered

    Joined:
    Jan 10, 2012
    Messages:
    249
    Likes Received:
    2
    Oh yeah ..lol I'll re-run LMAO
     
  9. MJP

    MJP Registered

    Joined:
    Oct 5, 2010
    Messages:
    988
    Likes Received:
    21
    The benchmark time should be ~67sec.
     
  10. Spinelli

    Spinelli Banned

    Joined:
    Jan 28, 2012
    Messages:
    5,290
    Likes Received:
    32
    Pay close attention to all the settings and pics on DrR's first post. It's easy to miss some stuff :)
     
  11. MrPix

    MrPix Registered

    Joined:
    Jan 10, 2012
    Messages:
    249
    Likes Received:
    2
    SO... once I used the correct layout for this bench....

    Intel i7 3820@ 4.3GHz
    rFactor 2 Build: 860
    Graphics driver version: 344.16

    GTX 980 @ 1500 core / 1752 mem

    PCI-E @ 2.0 x16
    Time: 67703ms - Avg: 125.090 - Min: 105 - Max: 160

    PCI-E @ 3.0 x16
    Time: 67578ms - Avg: 165.660 - Min: 119 - Max: 181

    so not much difference to the GT layout... I'll post in the other thread about the difference between the resolutions higher up with this same bench.
     
  12. Spinelli

    Spinelli Banned

    Joined:
    Jan 28, 2012
    Messages:
    5,290
    Likes Received:
    32
    Wow insane.

    Something just isn't right with rF2. Either that, or rF2 uses a humungous amount more PCI-E bandwidth than pretty much any other PC videogame in existence, lol.
     
  13. Marek Lesniak

    Marek Lesniak Car Team Staff Member

    Joined:
    Oct 6, 2010
    Messages:
    1,585
    Likes Received:
    101
    @MrPix
    You are getting much lower performance than overclocked GTX 970. There's still something not right with your PC.
     
  14. MrPix

    MrPix Registered

    Joined:
    Jan 10, 2012
    Messages:
    249
    Likes Received:
    2
    yeah, but my clocks are lower... the card I have throttles at anything more than 121 on the GPU and 0 on the memory... even at 125% power max
     
  15. Spinelli

    Spinelli Banned

    Joined:
    Jan 28, 2012
    Messages:
    5,290
    Likes Received:
    32
    Set the throttling "Prioritize" to temp instead of power. Then set temp target to something well above your max temp, like 95 degrees or something. That should help you with the throttling. EVGA Precision has the "priority" feature, I don't know about MSI Afterburner or ASUS GPUTweak.

    If your temps are fine but you're just getting too close to the power limit then water cooling won't do too much for that. You would still most likely need to increase the power target. Edit the card's BIOS, takes 10 minutes. Just change a value from 125000 to whatever you want, lets say 200000, then bam, you can raise the power target to 200% and avoid power target-related throttling. You won't use 200% power obiously, it just raises the threshhold for throttling.

    If you need water cooling for temps, though, then that's a different story. However, even then, you still need to get that power target raised if you're hitting it or getting within 3 or 4percent of it. 10 minutes is all it takes - edit a number, save it, flash the BIOS in Windows, done.

    I can help you with BiOS flashing if you want, literally takes 3 minutes.
     
  16. rogue22

    rogue22 Registered

    Joined:
    Jan 14, 2012
    Messages:
    261
    Likes Received:
    18
    As awful as it looks (no offense ISI), you think it wouldn't. I run into a bottleneck with Asseto corsa especially in the reflection department. Could be DX11 though.
     
  17. Timpie Claessens

    Timpie Claessens Registered

    Joined:
    Jul 26, 2011
    Messages:
    491
    Likes Received:
    1
    It's not just rF2, I think its just modern driving sims since they have to stream alot of data all the time. After being amazed from the results in rF2 gaining 25% performance on single an triple screen performance going from 3.0 X8 to 3.0 X16 I immedeatly fired up iRacing and Spa to go around the La Source hairpin to get the framedrop you always get there (from my stable 121) to around 85FPS and it was gone aswell. Same thing with the runup to Pouhon where the FPS always dropped it stuck to a stable 121 now.
    Havent tested AC yet but I guess there's probably a difference there aswell.
     
  18. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,294
    Likes Received:
    36
    Something doesn't seem right. Here are the GTX 970 scores that LesiU reported from his friend:

    Even with a 6% higher gpu and 14% higher mem overclock on a GTX 970 vs GTX 980, it should not be performing slightly better than the GTX 980.

    Are you sure you have a 980 and not a 970 MrPix?
     
  19. MrPix

    MrPix Registered

    Joined:
    Jan 10, 2012
    Messages:
    249
    Likes Received:
    2
    LOL
     
  20. Spinelli

    Spinelli Banned

    Joined:
    Jan 28, 2012
    Messages:
    5,290
    Likes Received:
    32
    Ya, those are real nice results, quite a bit faster than even a 980, wow, very nice Mr.Pix. Are you using the 344.16 drivers? Exact settings from the OP (alt-tab and check carefully), 100% default Nvidia control panel settings (except max pre-rendered frames 1)? Nothing changed in your player file that would affect framerates? If yes to everything then that is a very, very nice result.
     
    Last edited by a moderator: Oct 2, 2014

Share This Page