Massive FPS gains in rf2 using PCI-e 3.0 x16 with higher end cards!

Discussion in 'General Discussion' started by DrR1pper, Sep 30, 2014.

  1. DurgeDriven

    DurgeDriven Banned

    Joined:
    Mar 20, 2012
    Messages:
    6,320
    Likes Received:
    42

    I would like to see the bench-test format extended to include 1, 12 and 24 cars. :)

    I think we could get some interesting data , when it comes to better CPU's and faster clockspeeds vs same GPU.
     
  2. speed1

    speed1 Banned

    Joined:
    Jul 26, 2012
    Messages:
    1,858
    Likes Received:
    0
    I see some are still not bored to highlight Irrelevant things and to discourage new customers because supercomputers are needed to run rf2.
     
  3. Connor Caple

    Connor Caple Registered

    Joined:
    Oct 4, 2010
    Messages:
    28
    Likes Received:
    0
    I don't have a supercomputer and rF2 runs at a steady 60fps (VSync) with everything maxed.

    Some people just like to push things over the top. ;)
     
  4. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,286
    Likes Received:
    27
    Not being arsey, sincere question but do you consider yourself a glass half empty or glass half full type of person speed?
     
  5. speed1

    speed1 Banned

    Joined:
    Jul 26, 2012
    Messages:
    1,858
    Likes Received:
    0
    What ?
     
  6. speed1

    speed1 Banned

    Joined:
    Jul 26, 2012
    Messages:
    1,858
    Likes Received:
    0
    Can anybody helping me please what he is asking me ?
     
  7. speed1

    speed1 Banned

    Joined:
    Jul 26, 2012
    Messages:
    1,858
    Likes Received:
    0
    Good Night DrR1pper, what ever i'am in your view, have a good night.
     
  8. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,286
    Likes Received:
    27
    Sorry for delayed reply. It's a common expression in english that i've just realised you may have not heard before. It's an idiom for how people sort of generally perceive things. So being "glass half empty" means pessimistic and "glass half full" means optimistic and i was wondering which you would consider yourself leaning towards more. :)

    So for example, I consider myself to be a "glass half full" type of person but how we perceive ourselves does not always agree with other peoples point of view of us.
     
    Last edited by a moderator: Nov 10, 2014
  9. Marc Collins

    Marc Collins Registered

    Joined:
    Jan 11, 2012
    Messages:
    3,168
    Likes Received:
    162
    Just upgraded my motherboard and CPU over the weekend, moving from PCI-e 2.0 to 3.0 in the process, keeping my GTX 770 and all else identical. No re-install of Windows (used Sysprep).

    Drum roll, please....

    Almost no performance change to rF2 whatsoever. And what did change better be the effect of the i7 4.0 GHz chip I just laid out a bunch of cash for, because if that didn't result in the minor FPS improvement (10% max) then wow, the CPU has almost no effect for rF2.

    My conclusion is that after a controlled experiment where only two factors changed, PCI-e 2.0 vs. 3.0 accounted for zero.
     
  10. DurgeDriven

    DurgeDriven Banned

    Joined:
    Mar 20, 2012
    Messages:
    6,320
    Likes Received:
    42

    I can't agree less mate I had good gains with my specs switching from Sandy@4.5GHz and Z77 to Haswell@3.9GHz and H97

    But as I have said before there is so many variables to consider.

    How much of that 20% was ultimately down to my drivers. settings, build changes and OS conditions over time I could not even hazard a guess.


    If you want me to run identical fresh OS and drivers on the 2 setups forget it. lol :)

    Equally me running Pcie3.0 vs Pcie2.0 on my new system is irrelevant and pretty pointless when comparing from a upgrade perspective.



    I would really do a fresh OS though Marc........... :)

    Good opportunity to zero-write a SSD with Parted Magic and prep in diskpart to eliminate the "extra" partition. ;)

    aka:
    create partition primary align=1024
    format quick fs=ntfs
     
    Last edited by a moderator: Nov 10, 2014
  11. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,286
    Likes Received:
    27
    Marc, did you use the live benchmark?

    Are you sure you are using the exact same gpu clock and memory speeds as you were using on the old base system?

    I found a 7-9% fps increase going from 2.0 to 3.0 with my gtx 770 when running the live benchmark test and other 770 owners have reported the same.

    Are you 100% sure your card is running in 3.0 x16 mode when rendering something 3d? You can use GPUZ to check this.

    Provided your observation of no change in performance is made when using a well controlled and repeatable test environment such as the one used in the rf2 live benchmark thread, i would be inclined to say something is not working right on your system (since we have found a consistently noticeable increase in performance from everyone who has tested 2.0x16 vs 3.0x16, especially with a gtx 770 or greater. My immediate suspicions would be that your card is not running at pci-e 3.0 x16 mode if your observations are true.
     
  12. DurgeDriven

    DurgeDriven Banned

    Joined:
    Mar 20, 2012
    Messages:
    6,320
    Likes Received:
    42

    as 99% of time I agree with you .......lopl


    Still he needs to do a fresh OS you ask me coming from Sandy to Haswell ( I assume that is what he has done )
    Chipset and management even with 3rd party tools to rollback is not acceptable baseline to start from. ........imho

    Like I highly recommend TrueImage BUT strictly for full image backups , I would never ever use its "upgrade a motherboard" tool.

    A new generation CPU and Motherboard combo deserves a brand new partition.

    If doing that is always too much drama you been managing OSes wrong. ;)

    Not that I am saying that is why his results are the same, just saying. lol ;)


    edit:

    All I know between the 2 CPU/Mobos combos run on the same day with identical settings in
    inspector and rF2 Spa in pits jumped 30 fps.

    With a fresh OS I use saved NV Inspector profiles and my rF2 is on a separate HDD so no chance of changing or messing up any settings.

    Did not even have to re-run resolution as I used the same driver then, 344.48
     
    Last edited by a moderator: Nov 11, 2014
  13. DurgeDriven

    DurgeDriven Banned

    Joined:
    Mar 20, 2012
    Messages:
    6,320
    Likes Received:
    42
    Poor man's 4K .......lol

    Did a little test with reduced graphics settings and levels for DSR4.0 just to see if I could get acceptable feel.

    Result was pretty decent, ie: fps vs settings, stutter was very minor, good wheel latency, no screen tearing. Never did a PB but good laptimes stuck with 120% ai np.


    PC Specs in Sig
    Settings in Pics

    Spa , 4 pm , light cloud

    1 CAR

    2014-11-11 08:30:36 - rFactor2
    Frames: 28222 - Time: 470795ms - Avg: 59.945 - Min: 48 - Max: 70


    8 CARS OFFLINE

    2014-11-11 08:47:44 - rFactor2
    Frames: 5330 - Time: 96268ms - Avg: 55.366 - Min: 45 - Max: 75


    8 CARS ONLINE ( which is what I mostly do )

    2014-11-11 09:17:06 - rFactor2
    Frames: 6738 - Time: 120136ms - Avg: 56.086 - Min: 44 - Max: 70



    Just a test won't be using anytime soon in F3-Rookies obviously. lol
     
    Last edited by a moderator: Nov 11, 2014
  14. Ricknau

    Ricknau Registered

    Joined:
    Nov 12, 2011
    Messages:
    760
    Likes Received:
    31
    Really speed1? Do you expect us to end discussion because you think it is irrelevant? Why do you keep reading this thread? Can you not resist pissing in our beer? (Another English saying that I'm sure has a German equivalent!)
     
  15. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,286
    Likes Received:
    27
    Thing is Durge, i simply migrated my harddrive to my new mobo (Asus to MSI) and 2500k to 4690k and new gpu and I still observed the performance gain with my now 3-4 year old os. I also scored perfectly identical performances with other 970 owners.

    Before it was found out that it was the pci-e gen that was the cause for performance discrepancy between people with the same cards, i thought my install might have been the problem and came very close to breaking my lazy streak of 3-4 years without a reformat. Luckily it was discovered in time before i could break that "still going strong" streak. :)

    I'll tell you what was quite a bizarre yet pleasant surprise though. When i migrated this raid OS to the new MSI motherboard, the slow boot up time that had developed over time with my older 2500k was reduced dramatically. On the 2500k, booting into windows into a usable (i.e. you could open programs and windows etc) was dramatically reduced. It used to take like at least 60 seconds because the system was remotely usable to now being completely usable within 10-15 seconds from the first instance you see the desktop. It was a very nice suprise indeed, especially since this was one of the annoyances that made me also want to break my 3-4 year streak but it somehow fixed/removed itself.

    I'm not sure if it's the cpu or the motherboard that did this but if i had to guess i would say it's more likely something do with the motherboard. Perhaps the newer intel chipsets have improved the raid0 performance or something. But not so confident that is so since it works as quickly to get me into new mutliplayer maps equally as quick as SSD users (which i thought was interesting and know because we tested it and it was quite a bit quicker loading us into new maps online in sometihng like bf3 vs other non raid0 or ssd users.)
     
    Last edited by a moderator: Nov 11, 2014
  16. Spinelli

    Spinelli Banned

    Joined:
    Jan 28, 2012
    Messages:
    5,290
    Likes Received:
    30
    Hey Durge, ya, I'd be interested in doing some CPU tests. It would kind of be depressing though, it sucks when a game only makes use of 2 threads yet you have 12 (6 cores w/HT) :( personal sadness aside (lol) I'm definitely interested in CPU tests :).

    Marc Collins, you're case seems to be an exception considering the vast amount of users who have reported between 7 and 35+ percent fps improvements when going to PCI-E 3.0 @ 16x.

    Speed1, no one is discouraging anybody, and of course you don't need extreme high-end gear to run rF2. We simply noticed a weird phenomenon with rFactor 2, which is that - unlike almost any other game (in any genre) - there is a mysteriously, relatively large increase in framerates in rFactor 2 when going to PCI-E 3.0 @ 16x, and this is simply a thread discussing that phenomenon and sharing results - simple as that :)
     
  17. speed1

    speed1 Banned

    Joined:
    Jul 26, 2012
    Messages:
    1,858
    Likes Received:
    0
    No, yesterday I couldn't resist, sorry. :rolleyes:

    See my avatar, perhaps it explains my personality better.

    Excuse me, as it seems the dark side had the power over me, and I found myself in a kind of sarcasm mode, driven me in any needless problems.
     
  18. Ricknau

    Ricknau Registered

    Joined:
    Nov 12, 2011
    Messages:
    760
    Likes Received:
    31
    Looking forward to better tasting beer! :D
     
  19. Marc Collins

    Marc Collins Registered

    Joined:
    Jan 11, 2012
    Messages:
    3,168
    Likes Received:
    162
    OK, first thanks for being so smart! I installed GPU-Z and there seems to be a problem.

    At first, I checked GPU-Z and it said PCI-E 1.1x16@x16 1.1. I thought WTF?!?!? Then I read a bit and found out about the automatic power-down. Ran the rendering program and nothing. Stuck at 1.1.

    I pulled and re-seated my card and tried again. This time, when I first loaded GPU-Z it briefly flickered 3.0 before settling at 2.0. Now, the behaviour is that it shows 2.0; rests at 1.1 and switches to 2.0 when I run the render test. Very predictable. I have not been able to get it to show 3.0 again.

    My mainboard supposedly supports 3.0x16 (in the specs) and I am using just one card in the correct slot (closest to CPU), so !?!?!?!?!?!

    Checked performance in rF2 since the re-seating and 1.1 to 2.0 GPU-Z display switching and no impact whatsoever.

    Is it possible it is switching up to 3.0 when under load? Why wouldn't GPU-Z show 3.0 all the time? What the heck caused it to say 1.1 initially? Is there a common thread that results in either GPU-Z display errors and/or cards switching up and down automatically (and GPU-Z is accurate)?
     
  20. Jamie Shorting

    Jamie Shorting Registered

    Joined:
    Sep 11, 2013
    Messages:
    2,628
    Likes Received:
    1
    called power saving features. It will go to 3.0 when under load. if you're using latest version of gpu z click on the question mark beside the bus interface tab and you'll be able to put the card under load to see if it changes to 3.0. :p

    3rd gen and up intel chip with a compatible mobo and in the right slot it will be 3.0 ;)

    not familiar with amd chips though.
     

Share This Page