Alex said that his GPU-z showed PCI-E 2.0 8x while he had FPS issues. None is arguing about that. Did you run the benchmark? Some people mentioned less than 10% improvement, which could be not noticeable if no consistent test runs were performed. If pci-e bandwidth is responsible for FPS differences then factors like resolution, textures, GPU(more powerfull - more impact) itself or something else would make a difference.
Nothing to understand it is running at 2.0x Play a video , then watch the PCIe reading it will be 2.0x16 under load. I find rF2 lacking in W10 even with older drivers it just don't have the same spring to it. I went back to W7 Anyone who thinks pcie2 vs 3 is imagination or placebo is a idiot. lool
Are you sure that you couldn't have made a mistake and that when you thought you were testing in 2.0 x16 mode that you were in fact not but still in 3.0 x16 mode? It's just that there is around 20+ accounts confirming a difference between pci-e 2.0 x16/3.0 x8 and 3.0 x16. Ranging from the GTX 770 card and up, your 780 card included. Did you check with gpuz that you were in fact running in 2.0 x16 (or 3.0 x8) mode? With the amount of well established evidence we have supporting the claim that pci-e mode makes a tangible difference, even between the 2.0 x16/3.0 x8 vs 3.0 x16 mode level....makes me question your sanity a bit that you think your one test automatically trumps 20+ others who have all come up with the same consistent results and answer. Whilst you being right and 20 others wrong is certainly not impossible, don't you think at the very least it's less likely to be true that yours i right and thus thought maybe it be a good idea to double, tripple or even quadruple check your results before being so sure everyone else has it wrong? Or maybe even asking people what they think could be causing the anomalous result vs everyone else? You know, skepticism of everything including yourself is a wonderful thing especially when the overwelming body of evidence goes against your not as indepth/thorough time spent testing/experimenting and results.
I use an added 150mhz oc in afterburner to my GTX 970 gaming 4g card. Net total is just over 1500 mhz during load. I've managed up to 1595 mhz stable in game without driver crash. But it's not necessary and the backing down gives peace of mind.
With a 780, he should see at least a 10% increase in average fps from pci-e 3.0x8/2.0x16 to 3.0x16 mode.
Yes, I did run benchmark Heaven. I have also read this : http://isiforums.net/f/showthread.p...f2-using-PCI-e-3-0-x16-with-higher-end-cards! . rFactor2 Gmotor2 is updated a lot since 09-30-14, 09:01 PM , cpu core updates etc. I don't think thread is valid any more. I actually feel that more powerfull > less impact at least it is like that with my old MBO/CPU vs my new MBO/CPU with same cfx card (780 ti)
Why did you run the heaven benchmark? You won't see any (or any significant) difference in fps there between pci-e modes. It only affects performance in rf2 (to the best of our knowledge) and that is the only claim that has ever been made here. I thought this was obvious given that has been stated a number of times here and in the other thread. Or did you compare it in rf2 as well? If so, i apologies. But even if so....it's statistically far more likely that your results are wrong vs 20+ others being wrong. Maybe instead of just arguing who's wrong...we can try to collectively figure out who is wrong and why pragmatically? So i ask again...did you check with GPUZ that you were definately able to get your card running in 2.0 x16 mode during load?
Technically the higher the bandwitdh of the GPU the more impact it should make (if pci-e bandwitdh is the issue).
Alex set fullscreen window in rF2 Run MSI afterburner Click the "Windows Logo" button, click Afterburner on taskbar and bring to front. Drive rF2 and watch MHz / power / heat in realtime 1329MHz boost is correct
If you worried about expense it need not cost a lot If you listen to someone says you must have overclocked i7 Z97 and 16GB of Ram they have a lend of themselves. Of course it would be nice! no disputing that ........ but you simply don't need it for smooth fluid 1920x1080 ....shucks I get that with a GTX670 ! Stock 4690 + H97 Gaming + 8GB 1600MHz like mine is all you need as there is no need to overclock anything. I mean serious, if a 3.9GHz quad can't run it what is extra 2mb cache and 600+ MHz going to do in the big scheme of things...... ? Great bang for buck ? Bloody Hardly ! My if I had a GTX970 in my setup ... oh my ! .......... I can't imagine lol
Yeah. I remember testing anything above 2Ghz on an i5 2500k made zero difference to the performance in rf2 with and i5 2500k. (ofc though there is an improvement to be had with a 3000 series or newer cpu independent of clock speed due to the pci-e 3.0 x16 mode.)
Yeah mate I really want to like it I really do but it is not giving me anything I need. If rF2 ran as good I would use it. Drivers aside seems a bug with W10 and Kepler. If I get i7 , 16GB and Fast Card like you I will try it again but I not going to waste anymore on it otherwise. W7 is bullet proof for me. Even if it was as smooth as W7 , I can't stand either W10 startmenu and Classic Shell is just more resources Then you got to run blockers for the MS stuff. Yeah if not for Old sims and rF2 performance I would use it, so if you test on you tower tj and it runs identical or better then what it did you would keep it ...obviously. Just: Do a trueimage of W7 Keep MBR partiton so you don;t lose all your stuff on ( other partitions ) hardrive. With MBR you lose FastBoot but it runs just as good in every other way, trust me I spent months with both. Then if you don't like it you do Repair Windows from USB/DVD , run elevated command and Diskpart and "Clean" C drive then boot with Trueimage and restore W7 including the MBR table.
As long as others understand I speak only about 1920 X 1080 running "most" high settings with good smoothness.
I think it was independent of graphics settings as well. But probs not independent of AI numbers which i can't remember how many i tested with. So i guess it will be dependent on what you like to do with the sim then, i.e. AI count and whatnot. So i'm wrong.
Well D I``ll defo give it a try, haven`t yet busy this last while but I will for sure. Will let you know how it goes mate. Pity it`s not going great for you...
Well is was hard to go back, like I said rF2 was dropping fps even though I got it running what appeared smooth. Alas, Latency , timing, drivers , whatever it is something just does not feel right and my laptimes show it ie: opposed to W7. In a nutshell that is it, I can tolerate the rest. P.S. Upgrade you tower, activate ( it will use MBR table ) Get Product Key Make USB and DVD images Leave Bios as it is. Do fresh install , don't select a boot device says UEFI in front of it. In W10 hardrive setup go custom and DELETE C:\ and the 100Mb space. Then click "Next" DON'T click "new" DON'T click format. When you get a SSD I would first install W7 on it in MBR so you can connect the Sata as 2nd drive and then delete C drive and the 100mb space and format the unused space .......... or you can do it in Parted Magic. IE: You won't be able to do it from W10 once its running a GPT partition. Best way obviously assuming you have the space somewhere is to first transfer all your "stuff" off the 1TB and resize and format the lot to GPT.
Sorry to go off topic here but thought i'd share this for those who also own 970's or looking to get one. You know that last bit of 512mb of VRAM in the 970's that is only able to run at 1/8th the speed? Well if you didn't, you do now. It was the cause of some controversy back when this fact came to light a few months after it's initial release. So i've been playing batman arkham knight for the last couple of days and have been trying to figure out the cause of constantly random stutter ingame that as you can imagine is rather annoying and makes the playing not so enjoyable if not down right unplayable. I tried testing all different setting combinations except for the texture resolution setting. It never crossed my mind to test this variable because according to the ingame description, "Normal" (the highest available texture resolution setting) is recommended for any card with 3GB VRAM or more. Well, my GTX 970 has 4GB of VRAM, so surely i should be in the clear....right? Well, i finally checked my VRAM usage whilst playing the game and saw that it was always just over 3.6GB but never over 4GB. Not over the 4GB my card has....all should be good then right? But then in got me thinking, if it's over 3.5Gb then some of that 512mb 1/8th speed VRAM is going to have to be used of course. Could that be causing the issue i'm observing? So i tested the "Low" texture resolution which dropped the VRAM usage (reported by afterburner) to just over 3GB constant and the outcome from this? 100% silky smooth stutter free performance. **** me.....all those sources claiming that using any of that last bit of 512mb of VRAM that is only being able to run at 1/8th the speed as the rest would not affect performance in any game in practice.....Bullsh**! Now in practice does this affect most games you'll play....most likely not. But if you use more than 3.5GB of VRAM....the answers seems to be yes it will.
