Latest Release 1109 Rain Effects AND MORE!

pcie limited? Of course will be. GPU is a super fast 1080ti and the games utilities everything. pcie 2.0 is simply not enough. Where is the weird stuff?

 

The point being made is that rF2 is more limited by PCIE 2.0 than many other games.

Whenever it gets mentioned as a potential issue someone posts a link with a general comparison of 2.0 and 3.0 in games; these general comparisons don't correlate to what happens in rF2. So that's why the PCIE limitation is being talked about, not that it's something special that only rF2 does, but that rF2 exhibits more sensitivity to it.

For the record I think it's dangerous to rely too much on a comparison to exactly perfect GPU clock scaling with FPS differences of only a few percent, but there have been numerous indications of this 2.0/3.0 bottleneck over time in this forum.

 

Ya you are right @Lazza, problem is so many set ups hardware wise and somehow every systems seems to be reacting different. So i am glad i'm not a dev.
Happy new year

 

Maybe you are right i don't know. But i remember the hassle it gave when a deadline wasn't made some months ago. So my guess is S397 try to get every "bug" beta testers report sorted. But also keep in mind the deadline they have made for themselves. In the end no product released will ever be perfect and that not only goes for rF2 but is a general thing.

 

All the comparisons are from many years ago where gpus were way slower and didn't highlight the lack of pcie 2 protocol.
Nevertheless there is absolutely no point referring to pcie 2.
It's dead for a long time. No issue at all with rf2. It's just a 100% gpu bound game that will simply highlight a dead protocol's weakness when has to deal with the monster 1080ti.
Move on please.

 

It's 2018 and DX11.
No one should deal with pcie 2.0. these days. Especially with strong GPUs. It's common sense. And it's not S397's job to fix something that it's not there. It's like complaining why you can't run with a LaFerrari in a country with narrow roads.

 

Does anyone know what s397 is thinking about this pci2.0/3.0 problem?
I have a 2500k with pci2.0 and i would like very much to see performance improvements. All other games perform a lot better with my PC (assetto, pcars2, BF1 for example).
I can't buy a new pc just for one game.
I have nearly same fps with max settings or medium settings.

 

i think you will find that the reason it is talked about if because there is no game out there that bottlenecks PCIE as much as rfactor 2 does.
Even with PCIE3.0 rfactor2 uses far more then other games. i think i observed a 10fps difference when PCIE3.0 was utilizing 30% instead of 20% (id have to check my old post to confirm), so even though not a bottleneck it still manages to produce odd performance when heavily loaded. the high usages is mainly down to the fact that it uses something like 12GB paging file which suggest its copying VRAM into paging like old DX9 games. so i think its still relevant to talk about. perhaps studio397 can shed light on why exactly usage is so high.

 

You simply don`t have any clue what you are talking about regarding PCIe. You know my friend that PCIe 2.0 x16 has about same performance as PCIe 3.0 x8. Nvidia SLI has been around more then 10 years and most of SLI users are don`t have MOBO`s which are supporting PCIe 3.0 x16 x16 and when you are forced to use PCIe 3.0 x8 x8 you simply loose 30% of performance which means that you get only 10% performance improvement with SLI.
I have Asus X99 WS-E/10G > I7 6950x > 2x1080Ti and don`t have any problems with PCIe but still my ego isn't the level that I shout to community - it is 2018
rFactor 2 is PCIe limited so simple is that and I hope that S397 is going to fix this issue ASP

 

Again : rf2 IS a GPU bound game.
In my system my i5 4460 is at 50% maximum and GPU is at 100% if not at vsync.
It's totally normal that a GPU bound game will suffer more when on the slow PCIe 2.0 with strong GPU's.
Who ever says otherwise , needs to read more what pcie is.

 

Few games aren't GPU bound. The PCIE 2.0/3.0 performance boost was discussed (and researched) to death here in 2014, when other games had no difference between the two. And today nearly all other games still don't (though, coincidentally enough, those built on gmotor do).

Again - it's not just rF2, but it's unusually so. But yes, when everyone moves to PCIE 3.0 everything will be great - right? Because this anomaly can't be a symptom of anything specific to rF2...

 

I've never seen memory usage over 4 Gb in rF2, how do you guys manage to get that 12 Gb?

 

Even PCIe 3.0 x16 is near the limit currently. With an RX570 I am getting peaks of over 90%.



Knowing the way of working of ISI / S397, I suppose that next year anyone who can afford it will see a great improvement in this problem.

http://www.tomshardware.com/news/pcie-4.0-5.0-pci-sig-specfication,35325.html

 

Talking about page file usage or when rfactor is writing into memory to SSD/HDD showed using Afterburner. Normal ram usage with RF2 Ctrl + f is about 4gb I think after burning shows for rfactor around 12gb for virtual memory. I have 32gb 3000mhz so ram usage for me is nothing. Perhaps I'll turn of page file again and see how things work but even so that's a huge amount of data being sent around.

 

I use a fx 8350 with 2.0 and a gtx 1060 gives me mas of 100 fps, simply the person who uses a 1050 ti, which shows that it is not medium quality, it is low

they are overestimating a low performance of pcie 2.0, which does not mean that 3.0 is not better, but it shows that a 1050ti is low quality not medium

the fx is old so forget the history of pcie 2.0 vs 3.0, which is not more fundamental, than a graph of medium quality

I do not doubt the difference of a pcie 2.0 vs 3.0, but only with that does not feed

we can always get a higher quality hardware, and upload hardware-based fps, but it's also an economical issue ... if I bought a 1080ti even if the bottleneck would have more fps, but economically for me it is not possible, what if I believe that a 1050ti is low quality, together with a very old hardware that facilitated the dialogue about it, it shows that it is the result of its low fps, and the same person with all that wanted to play 4k on a tv...

today and since he has been studio 397 in front of the project rfactor 2 the improvement is impressive, obviously there are things to improve and things wrong, no one doubts that but the improvement is differentially huge, errors and mistakes will continue correcting and improving, happy 2018

 

I don't see this pcie bottleneck as a bug or an easily fixable problem. I view it as a design choice made by Cjon and ISI over half a decade ago. If this were an easily fixable problem, it would have been addressed when 3 or 4 dx11 updates a week were popping out.
My suggestion would be this. Everyone who even thinks they have a performance problem should be starting with the lowest settings suggested by S397. Start at the bottom and work up. I think most of us, myself included, choose settings we think we can live with and then degrade them if necessary. I'm suggesting reversing that. Go for the absolute bare minimum, and then step up an item or two SLOWLY til you find an fps/graphic quality compromise you can race with.

 

true but if I do a comparison of a gtx 1050 ti and a 1060, someone comes and says that the problem of low fps of the 1050 ti is because of pcie 2.0, for me it is a question of having a better graphic card and not necessarily of much more money, for example 1070 or 1080 or 1080ti.
if the hardware is old to be able to reach a 3.0 motherboard, in the end it is more money, than buying a better graphic card

 

I have a Gigabyte 1050 TI OC 4G in pcie 3.0 and no problems...only whit rain,and is not a old card,i purchase it new about 2 months,almost 200€,but after 2 days i realize that whit more 43€ i could bring the 1060 TI 3G....weel,now i have to deal whit a rushed purchase hehe,now new year and i need a new monitor my old Acer X243W is dieing

 

Nobody has attempted to pair a 1080 Ti with a PCI-E 2.0 motherboard, I agree that would be moronic. I run with 1050 Ti and it is PCI-E 2.0 limited (in rF2 only) and 1050 Ti is almost 4 times slower than 1080 Ti. If you take some time and re-read my calculations on the previous page, you will see that Ari's 1080 Ti is slightly PCI-E 3.0 limited (this under assumption that there is no other bottleneck and that performance should scale with GPU, which it normally does in rF2 because CPU usage is minimal).