1. Intel iGPU (6xx series) crashes
    Fixed drivers available!
    Instructions here

    Dismiss Notice

BANANA Benchmark Results Comparison

Discussion in 'General Discussion' started by garyjpaterson, Jun 23, 2014.

  1. Specht77

    Specht77
    Expand Collapse

    Joined:
    Jan 16, 2013
    Messages:
    78
    I'm very curious to see how would an overclocked Ryzen 7 1800x perform on bananabench

    Looking at the results of the Ryzen 5 above, i believe Ryzen 7 1800x would be able to reach the 6950X's 355MBeams/s, while costing 1/4 the price
     
  2. Justy4WDTURBO

    Justy4WDTURBO
    Expand Collapse

    Joined:
    May 14, 2016
    Messages:
    604


    Done on my laptop, a Toughbook CF-31 Mk3 with a GTX 1050 eGPU running at x1 Opt 2.0
     
  3. BeatPlay

    BeatPlay
    Expand Collapse

    Joined:
    Apr 11, 2015
    Messages:
    35
    That might be a bit unrealistic, the Ryzen 7 1800X should be compared to the 5960X, because both are 8C/16T.

    Even if the cores themself might be a bit faster, there would be a significant drop every 8 vehicles (the moment when the first core starts to calculate 2 cars) and a even worse drop at 16 vehicles (the moment theres no seperate threads left for cars).
    On the 6950X those drops only happen every 10/20 cars, because of the 2C/4T more. In the end this would lead to the 1800X may being faster at the start, but decreasing way faster than the 6950X.

    And also looking at the raw power of those CPUs the 1800X is fast, but a step behind still. I know raw performance does not say much across companies generally, but in this case per clock and per core performance is only a bit faster on side of the 1800X. And with both CPUs running 4Ghz, it again only comes down to the 2C/4T more on side of the 6950X.

    Also a little note: the 6950X has more room while overclocking compared to the 1800X, so when comparing max OC benchs, the 6950X will make the clear win.

    Edit: Also looking at price where im at it would be 1/2 to 1/3, not 1/4
     
    #763 BeatPlay, May 8, 2017
    Last edited: May 8, 2017
  4. NoxiousFumes

    NoxiousFumes
    Expand Collapse

    Joined:
    Apr 24, 2016
    Messages:
    786
    When you forget powerplan settings are 1% cpu speed
    upload_2017-5-15_20-1-50.png
     
  5. Bob PD830

    Bob PD830
    Expand Collapse

    Joined:
    Oct 25, 2013
    Messages:
    22
    24/7 Clocks Capture.jpg
     
  6. DBeee

    DBeee
    Expand Collapse

    Joined:
    Apr 14, 2016
    Messages:
    24
    1600X on my 24/7 OC
     

    Attached Files:

    • test2.jpg
  7. DD-Indeed

    DD-Indeed
    Expand Collapse

    Joined:
    Jun 22, 2016
    Messages:
    410

    R5 1600 here. How many times did you try out the bench ? I noticed, that when doing it couple times in a row, you can squeeze out more performance and keep the processor running in the top levels. I did the test like 10 times in a row and managed to break the 200 JBeam/s barrier. 3.8 Ghz speed, 2933 RAM speed with 15-15-15-35 latencies.
     
  8. DBeee

    DBeee
    Expand Collapse

    Joined:
    Apr 14, 2016
    Messages:
    24
    I ran it just the once. I can OC to 4.2 on the CPU and 3200 on the RAM and I got around 220, but my cooling seriously caused thermal throttling and it dropped to below 150. So that was my stock 24/7 OC... until I get my water cooling fitted anyway.
     
  9. DD-Indeed

    DD-Indeed
    Expand Collapse

    Joined:
    Jun 22, 2016
    Messages:
    410

    So you went over that throttling limit of 75 C, that's the information I got about that, correct ? And interesting to see, that you managed to get over 3000 with RAM speed, that also brings significant difference to these CPU's.
     
  10. jboptical

    jboptical
    Expand Collapse

    Joined:
    Aug 4, 2013
    Messages:
    28
    I've had my 1700 3.7ghz at 90c on the stock cooler :/ no throttling got hyper 212 LED turbo now 60c
     
  11. firestarspelt

    firestarspelt
    Expand Collapse

    Joined:
    Aug 4, 2013
    Messages:
    98
  12. DD-Indeed

    DD-Indeed
    Expand Collapse

    Joined:
    Jun 22, 2016
    Messages:
    410

    Ouch, that's pretty critical. I read somewhere, that is near the shutdown/breaking point of 95 C and the unstability begins at 75 C. My R5 1600 goes up to 60-65 C in gaming, but Prime95 tests shoots it up to 75 C. So the stock cooler is not capable enough on the highest stress, especially the smaller one.
     
  13. DBeee

    DBeee
    Expand Collapse

    Joined:
    Apr 14, 2016
    Messages:
    24
    That's far too high and scary. I'm using my very old AM3 cooler and when it hits 75C it immediately throttles. I'm assuming your early BIOS and voltage caused that? Was going to say, there was a 20C issue that AMD implemented to prevent thermal throttling, but afaik, that was on the early X series CPU's. Roll on Vega though, once I have that, I'm going full custom water loop.
     
  14. jboptical

    jboptical
    Expand Collapse

    Joined:
    Aug 4, 2013
    Messages:
    28
    1.25v :/ any bios still tried to cook itself almost convinced my 1700 has the 20c offset
    also don't bother with a full loop just get a air cooler overclocking is not really these chips forte
     
  15. DBeee

    DBeee
    Expand Collapse

    Joined:
    Apr 14, 2016
    Messages:
    24
    1.25? Mine is set to 1.3V (95W) but using Ryzen Master and CPUZ shows it at 1.375, but that's the BIOS trying to avoid crashing. But yeah, I can get 4.2 I recon, I just need to cool it better, so a water cooling solution is needed. I already own it all anyway, just don't have a GPU to put in the loop yet, and I want to do it all in one go.
     
  16. Blueshift777

    Blueshift777
    Expand Collapse

    Joined:
    Jul 9, 2017
    Messages:
    1
    Late 2016 MacBook Pro running BootCamp
     

    Attached Files:

    • Capture.png
  17. nolotank

    nolotank
    Expand Collapse

    Joined:
    Dec 26, 2014
    Messages:
    11
    @4.2 GHz

    EDIT: Better photo.
     

    Attached Files:

    • banana.jpg
  18. Drivver

    Drivver
    Expand Collapse

    Joined:
    May 1, 2015
    Messages:
    232
    BeamNG.drive version 0.10.0
    CPU: i5 6600k @ 4.5GHz
    RAM: 16GB 3100MHz CL15
    Untitled.png
    SAME PC:
    0.9 - 149 Mbeams/s
    0.10 - 126.5Mbeam/s

    I'm interested what exactly has been changed so that performance went down for ~15% - IMO crashes looks better in .10 than in .9 + added volumetric pressure to wheels, improved aero = better handling, also sounds calculations, but 15% is a quite big loss.
     
    • Informative Informative x 1
  19. BeatPlay

    BeatPlay
    Expand Collapse

    Joined:
    Apr 11, 2015
    Messages:
    35
    Well not really a big loss considering most PCs are allready fast enough to still do 6+ vehicles and that the last update was a 15% plus
     
  20. Specht77

    Specht77
    Expand Collapse

    Joined:
    Jan 16, 2013
    Messages:
    78
    Decided to test 0.10 here also...

    Same PC:
    0.9 - 161.8Mbeams/s
    0.10 - 149.8MBeams/s

    So my performance went down for ~7%, not 15% like in your system
     

    Attached Files:

    • BananaBench3.png
  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice