MSI GTX 970 mini review with rFactor 2 performance

Discussion in 'General Discussion' started by DrR1pper, Sep 25, 2014.

  1. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,294
    Likes Received:
    36
    That makes no sense. That would mean either you R9 290 is now performing 50% better than a GTX 780Ti with maximum overclocks in the replay benchmark or that your R9 290 is performing worse than GTX 470. I can't believe that it's either of them.
     
  2. TechAde

    TechAde Registered

    Joined:
    Oct 13, 2010
    Messages:
    606
    Likes Received:
    38
    Not just CPU clock speeds but bus speeds, RAM frequency & timings, mobo chipset, driver revisions, all sorts of things can and do make a measurable difference in my experience.

    Sent from my SM-G900F using Tapatalk
     
  3. speed1

    speed1 Banned

    Joined:
    Jul 26, 2012
    Messages:
    1,858
    Likes Received:
    0
    I see i'm confusing you somewhat. :) The replay performance of my sys wasn't equal or similar to a live session with the same config and requirements of the replay benchmark before.

    Yes.
     
    Last edited by a moderator: Sep 26, 2014
  4. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,294
    Likes Received:
    36
    Hmm...

    At least with 3d mark, i was comparing my gpu scores to the data base and the gpu score with same gpu and cpu clocks didn't change based on varying ram frequency's, one was 667mhz and the other 800mhz. I then tested 667 and 800 on my own and indeed, no change. I did not however check if the cpu score changed.

    But even if all the above points you raised made a significant difference, how does that translate to a replay vs live benchmark. Do you think those can affect the replay vs the live more? And if not, then were screwed either way....be it comparing replay or live performance.

    On drivers, i can understand completely but on the others, i would have thought even the most standard motherboards would not bottleneck the gpu in a graphics test. In a cpu test, i can see that happening.
     
    Last edited by a moderator: Sep 26, 2014
  5. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,294
    Likes Received:
    36
    Sorry speed, i still don't follow.

    Maybe if you through in some numbers, etc.
     
  6. speed1

    speed1 Banned

    Joined:
    Jul 26, 2012
    Messages:
    1,858
    Likes Received:
    0
    This has nothing to do with the current result of our little test in the other thread, so you can take those results serious.

    What am i trying to say is, as example, until today when i was using the replay benchmark, the result wasn't the same when i was using the config of the replay benchmark for a live session. I had big differences between a replay and a live session before and this is not the case anymore. The replay performance was terrible and now isn't.
     
  7. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,294
    Likes Received:
    36
    Right, ok. So the replay performance was worse before? How can that be considering the replay gfx settings are (i presume) higher than what you normally use (if i'm correct).
     
  8. speed1

    speed1 Banned

    Joined:
    Jul 26, 2012
    Messages:
    1,858
    Likes Received:
    0
    Really don't know, the settings i've always double checked, maybe it was a driver thing or something messed up before with my rf2 installation. The current driver in use is the 14.7 RC3 beta and the rf2 installation a clean lite install.
     
  9. Marek Lesniak

    Marek Lesniak Car Team Staff Member

    Joined:
    Oct 6, 2010
    Messages:
    1,585
    Likes Received:
    101
    That's on a single or SLI?
    Can't wait to see your scores from the benchmark. Of course, please double check you set all the details according to the description of the benchmark :)
     
  10. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,294
    Likes Received:
    36
    I've done a few different benchmark tests to compare the GTX 770 to the GTX 970, in both their nvidia rated stock clocks and the overclocks i was able to achieve on these cards. Each graph holds a different card as it's constant, representing the 100% performance reference point.

    [​IMG]

    [​IMG]

    [​IMG]

    The second one is what interests me most ofc and based on Sentri's stock GTX 970 performance, with my MSI overclock, i should be able to score about 84.2 fps in the civic + silverstone replay benchmark (a 46% increase over my Palit GTX 770 Jetstream).

    Now, regarding price-to-perofrmance's, for the £275 MSI GTX 970 card (with the easy overclocks), you're looking at about 60% more performance than a factory stock clocked GTX 770. To match it in price-to-perofrmance, you'd have to find a stock GTX 770 for £170 or my Palit GTX 770 Jetstream for £190. The cheapest i could find was an new MSI GTX 770 for £200 which performs roughly the same as my Palit. Ebay second hand GTX 770 cards however are selling at the same price-to-performance ratio.

    I would have included rFactor 2 into the results however i've still been unable to resolve the performance issue i'm having with rf2. It's still messed up with either cards. I'm beginning to fear a reformat is my only option.
     
    Last edited by a moderator: Sep 27, 2014
  11. Spinelli

    Spinelli Banned

    Joined:
    Jan 28, 2012
    Messages:
    5,290
    Likes Received:
    32
    I still don't understand why replay framerates drop so severely. In older ISI gmotor based sims, replays would be - at worse - the same as live.

    Someone once mentioned it's because of the "continue from replay" feature. That's the only thing I can think of as well. If that is true, then, damn......how much power does that feature need? I mean, we're not just talking about a 5 or 10 % difference here.
     
  12. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,294
    Likes Received:
    36
    Like i said in the other thread, don't compare the difference between live and replay benchmarks we did here and in the dedicated benchmark thread. The live replay was without any other additional cars.
     
  13. Spinelli

    Spinelli Banned

    Joined:
    Jan 28, 2012
    Messages:
    5,290
    Likes Received:
    32
    Then the comparison is pointless, no?

    Why not just record framerates of live gameplay and then record framerates of the replay of that exact live session that was just played?

    Even so, I remember when I did the comparison myself - 6 or so months ago - that the replay of the exact live session I also benchmarked had lower framerates than the live gameplay...
     
  14. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,294
    Likes Received:
    36
    Not really. Using a reply avoids problems with difference in race condition settings for example (e.g. whether, real-road, tod, etc). It just means one less thing to worry about when benchmarking comparing.

    Well, you can but with only yourself on the track and letting the AI drive without mistakes (to ensure consistent times around the circuit, etc., during repeated benchmark runs). Adding more makes the scene very subject to change which effects benchmarking result consistency.

    I don't even know if there actually is a large difference between a live and replay.

    edit:

    ok, so there definatly is. Just did a quick test, 25AI civic at silverstone, 40 second bench, average live was 65 fps and the replay was 44. So 32% less.
     
    Last edited by a moderator: Sep 27, 2014
  15. Spinelli

    Spinelli Banned

    Joined:
    Jan 28, 2012
    Messages:
    5,290
    Likes Received:
    32
  16. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,294
    Likes Received:
    36
    Ah, ok, gotcha. We did that comparison (me and spped1) to see if the performance ratio between cards would translate proportionately between a replay benchmark or solo live benchmark (which we found it does).

    Nothing more.
     
  17. speed1

    speed1 Banned

    Joined:
    Jul 26, 2012
    Messages:
    1,858
    Likes Received:
    0
    This was what i had experienced all the time until the latest tests, but for me there is no difference anymore, for whatever reason. I assume that my sys is now running on it's limits and the R290 is bottlenecked by, because it also makes no difference when i OC the GPU. Of course it is obvious that driving it with a Phenom on a PCI 2.0 base can't push the R290 to it's limits, but i would be intrested to see if your results between a replay and a live session with the same config of the replay would still differ in the result.

    I see in this thread ( http://isiforums.net/f/showthread.p...hen-shouldn-t-it-be-the-opposite-)?highlight= ) you also experienced a difference. Would be intresting to see if it is still the case.
     
  18. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,294
    Likes Received:
    36
    New graphs including rf2 performance with the live benchmark method:

    [​IMG]

    [​IMG]

    [​IMG]
     
  19. adaptable1

    adaptable1 Registered

    Joined:
    Feb 14, 2013
    Messages:
    112
    Likes Received:
    0
    Finally got around to testing again, and with reflections on high & everything max'd, I only got between 50-60% GPU utilization (avg FPS 34). Just changing reflections to low, it jumped 90-100% utilization (avg FPS 90). Something seems off about my SLI configuration. Also, with reflections on high, there were frequently black sections of track. I get between between 90-100% utilization in Assetto Corsa and other titles.
     
  20. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,294
    Likes Received:
    36
    This actually rings a bell with what i've heard from others so it's not an isolated problem to just your system.

    I can't actually remember what settings i used when i did have my 2x GTX 460's so i can't confirm it was indeed reflections on low that was why i found SLI working well for me. Actaully, i'm not even sure if those settings were available at the beginning of the beta.
     

Share This Page