Performance issues? Read more here

[Brian Clancy]

Okay, its pretty general knowledge that rF2 is heavy on resources, with even killer PC's having some pretty low FPS. I wanted to add a few details and put together a little info on the whys and what if's

Firstly, rF2 does NOT need or use the +Fullproc command anymore as rF1 did, it currently makes use of 2 cores. However, this is an issue for many people who run multicore (3-4-6 or more cores). Many people think that really good frame rates are almost totally down to the GPU (gfx card) but in reality, the CPU does handle a massive amount of the data/work for the GPU so the current limitation of 2 cores will and does effect some PC's performance The good news is that ISI may well increase this number later

Next up is GPU Drivers, this is something else that can have a dramatic impact on the performance of your PC. ATI/Radeon cards have proven to show big increases in FPS in rF2 by using the newest Preview Drivers (you can find the article Here http://www.racedepartment.com/forum/threads/rf2-fixes-and-adjustments-list.45695/) and you can be sure that newer drivers later will further improve performance on most cards.

The next thing to consider is that this is very much a BETA, the code has yet to be optimised for performance and little tweeks that can add to FPS rates have not been made yet. I did some testing today on my rig:

Intel Q8200 quad core 2.33Mhz O/Clocked to 3.5+ Mhz
Asus P5QB Deluxe MoBo
XFX HD6970 (Overclocked) running 3 screens in Eyefinity ( 3840 x 1024)
8Gig Corsair Dominator Ram

Now, this rig, running rF2 with all setting at Max/Full with 16xAA (No HDR) at Mills with the Megane averaged around 120 FPS. Very playable and really quite nice looking, but well below par compared to other games. Once I had finished my tests, it was apparent the GPU peaked at a VERY LOW 62% occupancy, leaving almost 40% of the cards capability unused. This will improve as the software is improved, but added to the 50% of cores unused, its not too hard to see that we can expect some very good increases in performance in the future before the gold edition.

I hope this helps to make users a bit more aware of the reasons for the current FPS issues..... At the end of the day, this is a BETA test and its really too easy to forget that with all the excitement

We will continue to add new fixes/updates and improved settings etc as we get them in the 'Fixes and adjustments list' sticky thread here: http://www.racedepartment.com/forum/threads/rf2-fixes-and-adjustments-list.45695/

 

[Eric. K]

Yes, Tim Wheatley did state

My good old friend Tim...

 

[Ole Marius Myrvold]

Yey! Go Hexa-Core!

 

[Mark Birney]

No mark it is not using half the CPU... it is using half of the CPUs (you have 4)

Should have known Mr. Intel would see such an error

 

[EIF]

Now, this rig, running rF2 with all setting at Max/Full with 16xAA (No HDR) at Mills with the Megane averaged around 120 FPS. Very playable and really quite nice looking, but well below par compared to other games

This makes no sense to me. 120 fps is playable but below par??? 120 or even half that depending on your monitor refresh rate is all the fps you ever need! I'm amazed you eget 120 on your spec and I for one would be delighted with that sort of performance

 

[Brian Clancy]

This makes no sense to me. 120 fps is playable but below par??? 120 or even half that depending on your monitor refresh rate is all the fps you ever need! I'm amazed you eget 120 on your spec and I for one would be delighted with that sort of performance

It does play really well (as I stated) but a game like this in DX9 I would normally see at least 50% more FPS. I am suprised that so many people are shocked at the good fps for my sytem. I was going to upgrade the Mobo and CPU soon....Not so sure now

 

[Teemu Toikka]

rF1 I get 200+ fps with full settings, with rF2 I get max 50. Still it is playable for me. And I guess I need to get some memory, rF2 takes 1,7 Gb when in game and I only have 4...

 

[EIF]

Anything above 60 or 120 fps is just a waste and only creates extra load and heat on the GPU . . . epeen and nothing more.

I can definately perceive the difference between 30 and 60 but in the hundreds? C'mon!

 

[Brian Clancy]

Anything above 60 or 120 fps is just a waste and only creates extra load and heat on the GPU . . . epeen and nothing more.

I can definately perceive the difference between 30 and 60 but in the hundreds? C'mon!

Disagree With the newer driver I saw a gain of about 15-35 FPS depending on track to around the 120 mark and I could tell the difference. I have had this argument before, I also believed the old tales of 30 fps being the limit for the human eye, but its just not true. For sure there are diminishing returns after say 60-70 FPS, but you can still see or 'feel' them. I'm not an expert on this, but my gut feeling is that we may not percieve much change above 80-90 FPS but we do notice the minuscule drops in frame rate that happen many times a second. By running the FPS at over 100 even these 'drops' remain above 80-90FPS and as such are far harder to notice, giving the 'feel' of an improvement Only a personal theory, but I can see the changes in FPS even though they are small, although much above 120 fps, i think would be really hard to see

 

[Ivo Simons]

And with more FPS you also have less Latency in your steering wheel.
Thats why almost everyone drives with vsync off.

 

[Brian Clancy]

And with more FPS you also have less Latency in your steering wheel.
Thats why almost everyone drives with vsync off.

In almost all sims that is 100% right But I believe from what I have read, that the wheel and FFB effects in rF2 are NOT tied to the screen / programme re-fresh rate :O Thats pretty good news

 

[Mark Birney]

The human eye cannot see more than 60fps, I heard somewhere.

 

[Brian Clancy]

The human eye cannot see more than 60fps, I heard somewhere.

From what I understand and my own findings, thats not 100% true

 

[johnnyw]

In almost all sims that is 100% right But I believe from what I have read, that the wheel and FFB effects in rF2 are NOT tied to the screen / programme re-fresh rate :O Thats pretty good news

Where did yo got this info? My personal feeling is that its not correct. Tested different settings myself today and steering definately feeled better/smoother with mix of high/med settings and fps around 170 than with maxed out and fps in 80-100 range.

 

[UMC 22]

It was probably gleaned from the internet, the same place many people have 'heard' things that simply aren't true. This discussion is one that comes up frequently on gaming forums. Unfortunately, there are never any optical scientists involved to point out how wrong people are. No-one has proved one way or another just how much information we can process through our eyes in terms of 'FPS'.

One thing worth pointing out is that our eyes are not video cameras. They process a constant stream of information, and do not capture individual 'frames' at a specific rate. This information is then processed by our brain, which is capable of applying some interesting affects of it's own to make the final image that we 'see' look better. Those affects are why TVs, monitors and cinema screens work at all for displaying motion, rather than looking like a set rapidly flashing still images.

Put very simply, the closer something on a monitor is to the real world, the better it will appear to our eyes because less of those affects need to be applied by our brains. In other words, the higher the FPS (and therefore closer to a constant 'stream' it is), the better it will look.

 

[Brian Clancy]

Where did yo got this info? My personal feeling is that its not correct. Tested different settings myself today and steering definately feeled better/smoother with mix of high/med settings and fps around 170 than with maxed out and fps in 80-100 range.

Direct from ISI Its stated in the forums....

I have found no change to the FFB/steering in the FPS range you are talking about, strange... But an awful lot of this is perception and to some degree, may be also be influenced by hardware

 

[James Scanlan]

id love to know why my system is being such a massive pain in the but Core i7 920 6gb corsair dominator ram asus P6T deluxe V2 and an ATI 4870 is struggaling to run with all the on off options set to off and all detail settings set to medium except for shadows which is on low. i know the 4870 is getting on a bit but rF2 is still DX9 and this is a DX 10 card this system runs crysis 1 on full everything and crysis 2 as well. F1 2011 struggles a little but again its a DX 11 game. i realise this is still in beta and performance issues are a given but still this system shouldnt be struggaling on MEDIUM !! surely ?

 

[UMC 22]

I don't know what the problem is on your system, but it should be running better than that. I have a Core2Duo, 2GB or RAM and a 4870, and I can run rF2 smoothly with everything set to 'high'. Perhaps check your card settings in CCC, try with Catalyst AI enabled/disabled, turn off AA and AF if they're on, and make sure you're using the latest preview drivers detailed in Brian's OP.

 

[Adam Vaughan]

I was told by a College tutor that if the human eye can see x amount of frames, then the more FPS you do get will be very noticable as the frames that you see will never ever be synchronised with your eyes "frames". He also said the human eye isnt just as simple as that. Hope that make sense.......

 

[Mark Birney]

Interesting observation today. Reinstalled drivers (had problems with newer beta ones) and then got some bad artefacts in the game. I reinstalled the game and was dumbfounded to find that the performance was much smoother.

Worth a try to reinstall the game, heard others mention it.

 

[Gary Corby]

Performance gain from parallel CPUs is always constrained by whatever amount of the code must be run in sequence, which varies per program. Typically you see a linear improvement curve with small numbers of CPUs, which slowly flattens out to no improvement at all at some point which has to be found by experiment (communication overhead between the threads is also a factor). Since driving a car down the road is very linear (by definition!), I'd be surprised if 32 cores showed any significant improvement over, maybe, 8. That's just a guess, of course.