Mirror Usage and FPS

[fiddlinjim]

I have a new computer rig and fly the the YAK-1b and today I implemented the mirror and FPS went down dramatically-very dramatically in 1 on 1 combat.    Would like to fly using the mirror but can't use it with the impact I am seeing.

I fly with ULTRA graphic settings, clarity is great and prefer to remain that way so if anyone has a solution outside of decreasing my graphic settings in VR would appreciate hearing how to do it.

 

Win 10, i7-7700K, 32GB, GTX 1080Ti, 1 TB SSD, 4 TB HDD, Oculus Rift CV1, TM Warthog with extension, Saitek pedals

Edited October 22, 2017 by fiddlinjim

[THERION]

I'm a bit astonished to see you've got troubles with fps using mirrors. You have a very decent rig and far better then mine.

 

I use mirrors too, but run them in simple mode (which is absolutely enough for situational awareness) and I don't have

significant fps drops - even on the very demanding Kuban map.

 

But since I do not have VR system I don't know how this is affected. My monitor runs 60 mHz so I simply locked my frame rate

in the game to 60 and I never ever get any fps drops nor stuttering.

 

My system: Win10, i7-4770, 16GB, GTX970

 

Cheers

[chiliwili69]

Do you overclock?

 

If not, you should at least try to go for 4.7Ghz (with good cooling and monitoring temps). It improves significantly the fps.

 

Try to run the VR performance test to be sure you achieve the expected VR performance for your rig.

[dburne]

Make sure you are not running with HUD enabled, and also no SSAO or HDR. Also no cinematic camera.

 

Nvidia settings - bout all you can do in there is set to " Prefer Maximum Performance ". Make sure it is not on optimal or adaptive. Everything else leave at default of '"application controlled".

 

Are you running any SS in either the Oculus software or Steam VR? Anything above 1.5 in Oculus or 3.0 in Steam VR usually has diminishing returns at a great performance penalty. Definitely do not use both, just one or the other.

 

Also when you installed the Nvidia drivers, did you install only the drivers and PhysX? That is all I ever install, definitely not any of that other stuff especially GeForce Experience.

 

Even at your cpu's default speed I would think you should get very good performance even with mirrors on.

Edited October 22, 2017 by dburne

[SCG_Fenris_Wolf]

Dburne, what settings do you play on again? I am happy with mine, but the mirror performance is terrible, and want to try yours out as well when returning home tomorrow evening. Maybe there is some certain secondary setting causing "mirror on" to cook performance.

[dburne]

Dburne, what settings do you play on again? I am happy with mine, but the mirror performance is terrible, and want to try yours out as well when returning home tomorrow evening. Maybe there is some certain secondary setting causing "mirror on" to cook performance.

 

I run on High settings, no SSAO and no HDR.

Landscape detail x4.

4x AA

Sharpen textures

Shadows on Ultra

HUD Off for most part

 

Note though I am currently flying the Yak-1 in PWCG campaign, so no mirrors for me at this time. 

[TG-55Panthercules]

Hmm - it would be nice to have a version of the benchmark track that uses a plane with mirrors - I didn't see much impact when I ran the benchmark track with and without mirrors activated, but I later realized that the plane in that track doesn't seem to have a mirror so the results may not mean much WRT the mirror FPS question.

[chiliwili69]

 

 

Hmm - it would be nice to have a version of the benchmark track that uses a plane with mirrors - I didn't see much impact when I ran the benchmark track with and without mirrors activated, but I later realized that the plane in that track doesn't seem to have a mirror so the results may not mean much WRT the mirror FPS question.

 

Yes, when Balapan and m0n0type created the recorded track they used the Yak-1 and at that time they didn´t know mirrors were going to be implemented.

 

If we had to create a new benchmark from scratch we would chose one plane with mirrors, so we could measure the performance hit.

There are others factors that influence the performance like the plane itself, the clouds, the smoke, altitude, the mountains, etc.

 

I was thinking to create a test procedure rather than a recorded track (where plane, scenario, time or weather is fixed). For example doing a Quick Campaign with autopilot for 5 minutes. But the amount of info to log would be huge and I have not too much time and could confuse new users with the current benchmark results.

 

If any of you would like to create a more complete new recorded track we could discuss that in the related thread.

[Dave]

I have subjectively noticed an obvious hit in VR with the mirror fitted.