CQC Game performance - optimization - hardware & technique

how do you suggest i improve my play ?


on 4Q 1xMultiSampling the game caps out at 59.9Hz (refresh speed lock)


at 4Q 2xMultiSampling the game caps out at 29.9Hz (sometimes a little higher but not often)


do 1xMultiSampling is best in my opinion.... and 4Q...


people are reporting not properly configured Crossfire config with stuttering (sort this please) + help the developer; you have tools !


the timing may be related to buffer clearing ? is that a problem..
memory usage on settings 4Q 2x supersampling consume 2998mb ram and 2500mb on 4Q 1x supersampling with full effects....
so 4Q 1x supersampling recommended i consider for a 3gb card and 16x Tessellation (16x is 300% faster on the tessellation benchmark than 32X in my testing)


Yours Sincerely me


links :


Frame Rate and stats are displayed thoughout both videos #arena #Deathmatch #PvP


https://youtu.be/RoI22Qg8HT0 (games best performance (way the best on 4Q 1x supersampling)


https://youtu.be/nYG9TUS5GYM Elite Dangerous 64Bit - PvP Arena DeathMatch - 4Q Anti-aliasing 2x super-sampling - 2016 04


feel free to analyse both videos - id like you to


1920x1080 native resolution used
 
Elite Dangerous uses deferred rendering so multi sampling AA doesn't really do anything except increase draw time?
 
SNIP
Actually scratch that, I am half asleep, can blame the morphine sulphate; you need shader based AA for compute shaders and other deferred shading techniques, forcing traditional AA through drivers has 0 impact on image quality
 
Last edited:
the supersampling is set ingame aka 2 or 1 ... 2 has half the frame rate of 1

First of all your inital post mentions multisampling, there is no multisampling available in game, which is why both insomnia and I assumed you must be talking about forcing AA through drivers, which does not work in ED.

If you mean in game MLAA that is a postprocess shader it stands for Morphological Anti Aliasing and it is applied after the scene has been rendered, both FXAA and SMAA work in much the same way and all are used where a game engine uses deferred rendering and traditional anti aliasing will not work. They are listed in game in order of performance hit and image quality with FXAA having the least impact but also producing the worst image quality of the three methods.

The in game supersampling affects the size of the frame buffer that the scene is rendered to; so for example, X2 supersampling renders your 1920 X 1080 scene at 3840 X 2160 then down samples that rendered image to the chosen resolution before it is sent to your monitor; X1 supersampling means render scene at the chosen display resolution, X0.75 means render the scene at 75% of your chosen screen resolution, so your 1920 X 1080 scene is rendered at 1440 X 810, and so on.

There is no direct corelation between changing to X2 supersampling and halving the frame rate, that is an effect of your having chosen to use vsync when the card cannot render the increased frame size at the native refresh rate. When the card is not able to render the screen at 60 fps it is halving it, if your card cannot render at 30 fps it will attempt to render the scene at 20 fps, then 15 fps, and lower until it can render the scene at a fixed frame rate that can be synchronised with the screen. If you want a decent explanation of how Vsync works I recommend reading https://hardforum.com/threads/how-vsync-works-and-why-people-loathe-it.928593/

In most cases you should be leaving in game supersampling at X1 unless you have a higher end card as it is computationally expensive. If your card is struggling to render the game you can try using .85 .75 .65 to have the game render the scene at a lower resolution while still displaying at your monitors native resolution. FWIW, If you have an Nvidia card, rather than using in game supersampling, you are better off enabling Nvidia DSR, and selecting a higher render resolution in game. Because of the optimisations on the newer Nvidia cards, assuming the same render resolution the performance impact of using DSR will be much lower than using in game supersampling and the anti aliasing effect will be exactly the same.
 
so you are saying HexCaliber that the setting doubles the size rendered to ? ... why does the amd driver alter that rather than the VSR then ? but cheers for the decent explanations...
 
Back
Top Bottom