Rise of Flight settings...what's everyone using?

[TX-EcoDragon]

I just updated RoF and see many new graphics options. What are you guys using?

[GOZR]

All maxed!! on Crossfire  ..  I should fire it up  and go for a spin!

[TX-EcoDragon]

OK, after much fumbling around with things like RadeonPro and seeing horizon shimmers again and generally not the best ever image quality with some of the higher settings I've my best overal image quality and performance just doing what I did in the beginning:

Using only the Catalyst Control Center and setting it to allow most of the settings to be established in RoF's control panel.

I was running maxxed out settings in RoF but with SuperSampling off - and it looks like it's worth it to run supersampling even if other settings are lowered. I'm still experimenting with things but here's what I'm currently pretty happy with:




Note that to properly use Supersampling you should only use 2X AA and set filtering to Bilinear.

[GOZR]

Why special medium shadows? try gamma at 40 just look a tight better

[TX-Chukar]

Thanks Eco, I needed a good starting point.

[GOZR]

Now if you want performance FPS those can be tweaked a bit more with out loosing much in quality.

[TX-TOAD]

I got 15fps by checking MultiGPU Eco You should try.

[TX-EcoDragon]

Yes, I know about the Multi-GPU option, but it introduces a little stutter, and my fps are good enough...I'm sure you see better fps because the Multi-GPU button turns on optimizations that are normally off, but check for yourself if the gameplay is actually smoother better, ignoring the fps.

[TX-EcoDragon]

I heard that the new Fantasy Island Island was cause Steve some serious fps hits so I've run a quick benchmarking session on that map during a multiplay session with 8 players.

I departed the tree surrounded mountain base to the South East (the hardest fps hitter I think)
I took off, headed north to the bay, engaged a DVII and flamed him, then landed at the SW base after a flyby of the castle

2012-12-27 14:29:21 - rof
Frames: 25928 - Time: 426179ms - Avg: 60.838 - Min: 32 - Max: 123
I saw the 32 fps at engine startup while in the tree bowl
though the 32 was smooth, if you didnt' have fraps on you wouldn't see any issues...no stutters...

[TX-Gunslinger]

Working to test and contribute Nvidea settings.

Ever notice that I'm always the one (well Todd is with me in this) that has to opposite vendor?

Soon...

S!

Gunny

[TX-TOAD]

In this track from original developers I'm getting this kind of performance.
Just use fraps to benchmark as the screen prompts.

http://www.filedropper.com/perfomancecheck-resave-me2012-12-2722-19-0200 Grab it soon it won't stay there forever.


My machine Specs
Mobo MSI 890FXA-GD70 AM3+
CPU AMD Phenom™ II X4 965 Processor @ 3.8
CPU Cooler ZALMAN CNPS9900MAX-B 135mm Long life bearing CPU Cooler Blue LED
SDRAM G.SKILL Ripjaws X Series 16GB (2 x 8GB) 240-Pin DDR3 SDRAM DDR3 1600 (PC3

12800)
GPU EVGA 03G-P4-3668-KR GeForce GTX 660 Ti FTW+ 3GB
HD are a Mix of SATA Drives WD and Seagate None SSD
CASE COOLER MASTER HAF X RC-942-KKN1 Black Steel/ Plastic ATX Full Tower Computer

Case

In-game Settings are on this run are

1600x1200
Full screen
Vertical Sync
fps limit 60

Post effects

Super Sampling
Raindrops



Quality Settings

lanscape High
Grass med
Trees high
Shadows Common High
Reflection High
Textures High
Filter Bilinear
Forest Vis Very Far
Shader Quality High
Light Sources 15

Advanced Settings

Antialiasing 2
Gamma correction 30%
landscape mesh 100%
Objects visibilty 80
Multi-GPU Checked

Outside of game GPU settings

AF application controlled
AA off
AA Gama Off
AA Mode aplication COntrolled
AA setting aplication COntrolled
AA transparency Off
Cuda All
maximum pre-render Use the 3d application
Multi display single display performance
texture filtering AA  off
texture filtering  Neg LOD Clamp
texture filtering Quality
texture filtering trilinear On
theaded Optimization Auto
tripple buffering of
vertical sync Use 3d application



Frames, Time (ms), Min, Max, Avg
  1115,     59825,  16,  21, 18.638


---------------------------------------------------------

Further testing

from this point I'd like to get a normal Online session track about 3-5 min long from Eco's smooth as silk machine. So we will know the track has no hicups induced form a poor performance machine. That way we could tweak our machines to a setting to get smooth replay out of his track. Or just buy more hardware LOL

[TX-EcoDragon]

TOAD...we should probably turn off vSynch and set fps to unlimited for benchmarking....

Also, everyone should be using the *Special* shadows unless they are using a 4xx series nVidia card.

http://riseofflight.com/forum/viewtopic.php?t=12384

[TX-TOAD]

OK Joe said the same thing about vsync.

I've always tested with it on because that's how I play???

I can't recall ever being told to do it otherwise short the notion of just getting higher numbers for bragging sake, from the old IL2 days of tweaking.


That post is from 2010 and I'm using a much newer Nvidia card than the 4xx series. I'm a 660.

I'll try it both ways though for testing.

Here are numbers after setting to no vsync, unlimited, and special shaders

Frames, Time (ms), Min, Max, Avg
  1236,     60078,  12,  24, 20.573

Same settings but with Common High FTW  

Frames, Time (ms), Min, Max, Avg
  1249,     59996,  17,  24, 20.818

Same settings but OC my GPu to boost settings base clock is 1045
GPU OC Clock 1136 temp was 49c

Frames, Time (ms), Min, Max, Avg
  1461,     60176,  20,  28, 24.279

[TX-EcoDragon]

Right, the "Common" Shadows were written specifically for the 4xx series nVidia cards because those cards did not have certain instruction sets that other cards have. For this reason, the only people that will benefit from the "Common" setting are those card owners. It may be more a matter of quality than it is performance, as we see your numbers are essentially the same either way.

As far as vSynch goes, well....benchmarking is about comparitive performance and seeing changes from one set of settings to another, and vSynch will force a max fps of 60, so it will invalidate your testing unless you are well below 60 fps on average and never see higher than 60 fps, so the convention in benchmarking is not to restrict the max fps. That said, vSynch does reduce performance somewhat, and not just because it limits max fps to your refresh rate, but also because it adds some rendering and processing time, but generally not so much as to outweigh the image quality gains.

[TX-EcoDragon]

Also, apparently the Tree setting of High apparently really hits fps after the latest patch (perhaps because of the new fog?) so consider lowering that value.