Rfactor2 The VR thread -settings & tips

Should I use AA or Supersampling?

 

From what I've noticed aggressive threading seems to put high usage on core 1 and then core 2 (the virtual core I think) gets utilised a lot too (see pic above). Maybe it helps latency? With it turned off usage seems to get spread out among all the cores.

I could be talking complete bollox tho' - like I said I'm no expert.

 

CPU Cache don`t give you one FPS more or better game experience and it is an area of fast memory located on the processor. Intel® Smart Cache refers to the architecture that allows all cores to dynamically share access to the last level cache.
i5 8600k > Cores 6 > threads > 6 > 9 MB
i7 8700k > Cores 6 > threads > 12 > 12 MB

 

I have also i7-6950X which has 10 cores and 20 threads which I have tested vs i7 8700k and i7 4770k. What I can see is that all cores and threads are used in rF2 no mater which CPU it is but when I run benchmark all CPU`s has equal performance and fps in rF2 and other games I am playing.

I asked once from S397 same question as you about CPU usage and this is answer I got: "We are trying to use more CPU cores to help speed up graphics and we are trying to make sure that physics always have priority of that. Those two things combined can mean that the physics core ends up "helping out with graphics" a bit more if it has idle time."

 

I don't know if anyone here has had an issue with their Rift going black, with sound still on, and then being chucked out to the desktop. I had been using my Rift for some time without issue, and then recently I had this issue. In the last couple of weeks it increased in frequency and could happen after 10 mins or half an hour.

After the last time, I did some research and found a couple of threads in the Oculus forum about this type of issue. I followed several pieces of advice and post here what has worked for me. This appears to relate to the sensor and the power from the USB 3.0 socket, perhaps after the last Windows 10 update. Anyway:

1. Completely remove the current Nvidia driver and do a fresh install of driver 388.71 [do not use a newer version].
2. Separate the headset USB 3 and the sensor USB 3 - in my case I had already purchased a USB 3 PCIe card. I now have the headset in this card and the sensor in a USB 3 port on the front of the computer.
3. Windows power setting - I had this on the max power setting, but have now changed this to the balanced power setting.

Having done the above three things, I have not had a reoccurrence of the black screen issue and have used rF2 for many hours [practice and in races]. My system is - i5 6690K running @ 4.2 GHz, Strix 980 Ti and 16 GB RAM.

 

What you are saying in your post is that i5 6600k at 4.8GHz is bottleneck in pC2 and other games There is noway task manager is showing i5 6600k >100% CPU usage in any game. However I remember different story from your previous posts in this forum about you not using equal size of RAM etc. with i5 6600k which could indicate that 100% CPU usage was caused by something else

 

That appears to contradict your earlier statement that rF2 only uses 2 cores and hyperthreading is not needed?

 

rFactor2 uses one core to physics and one to graphics. There are only few games which has true quad core support and rFactor2 is not one of them.

 

@GTClub_wajdi As I understand your PC does not support PCIe X16 3.0 (Gen 3) If that is the case read this post: https://forum.studio-397.com/index....dmap-update-may-2018.60196/page-9#post-947713

 

I don't think that's true anymore since last perf patch. That's why the more core/threads system has the better. This will likely become even more important once higher FOV HMDs arrive. https://www.studio-397.com/2018/04/release-of-build-1110/

"Thanks to DX11, we are now able to take advantage of multiple graphics threads and leverage the modern CPUs to render the scene faster. Internally we had to make quite a lot of changes to the code to make this possible and really take advantage. One area where we got a significant improvement is in our shadow rendering. Using multiple threads speeds that up, and another thing we did was to no longer render the shadows separately for each screen"

Also, turning off HT on modern processors is not productive, and modern here applies to any Core series processor. Roughly, HT processor is capable of running ~1.5x more threads simultaneously compared to no HT. HT had problems in early CPU implementations, now turning it off is just throwing away a bit of multithreading capacity.

Additionally, multiple cores/threads help in VR (Oculus runs separate service process, Steam VR too) and for sim tools running in background.

 

If build 1110 was released with true quad core support it is so huge news that we could read about that in the first sentence in the roadmap.
"we are now able to take advantage of multiple graphics threads and leverage the modern CPUs to render the scene faster etc." Does not mean that game has true quad core support.

 

"..One area where we got a significant improvement is in our shadow rendering. USING MULTIPLE THREADS SPEEDS THAT UP, and another thing we did was to no longer render the shadows separately for each screen"

That says to me hyper-threading is an advantage? Unless I'm misinterpreting it ?

 

Yes. More physical cores is better but HT is a plus as well. On my i7 8700K rF2 is no longer CPU bound in VR, it is held back by my 1080Ti. Except for rain of course, something isn't ok there.

 

Sorry what does that mean? My cpu isn’t good with rfactor2?

 

Up to a point, yes. But scalability (gain from more cores/threads) will always get less the more cores/threads you add and this applies for all games. So you will likely see a big difference between a dual and quad thread CPU. But 6 vs 12 threads, you are not likely to see any real benefit in most of today's games.

 

Sorry it is my mistake. I thought you have i7 2600k as Adrian. You are just fine with i7-8700, you have PCIe 3.0 x16 (Gen 3) support.

 

The RAM has since been replaced with the CPU so that may have been the cause ! But I ain't lying about my 6600k being at 100 % in pC2 , WD2 and ROTTR ! But cant test anymore since I've sold both RAM and 6600k !
But after my RAM to 2x8GB 3200MHz and 7700k upgrade ALL my games no longer show past 50%in task manager where before it was red and sitting in 95-100% all the time and stutter would be present !

 

Is it worth going for 3200mhz ram?

I’ve read conflicting opinions but I guess that’s due to the application

 

That is something i never tested in games as I swapped out my RAM and CPU at the same time . I had a mix match set and upgraded more for the sake of having matching system ,GPU,CPU and RAM all of similar level .
Digital Foundry did a good vid showing performance increases with RAM speeds when they reviewed the 7700k .


Jump to around 4:50
@Ari Antero might like this vid also , it shows gains from i5 to i7 jump to 3:00 or just watch the whole thing , its pretty informative !

 

This guy says that i5 is similarly clocked as 17 but it is not. i5 and i7 CPU ratio are at 4,8 GHz but he don`t OC the CPU internal bus speed to same ratio which you must do if you like to get equal performance i5 vs i7.
Just stay tuned to youtube and believe everything those youtube guys are telling to you and you are soon the proud owner of the i9 7980xe.