FPS throttling (?) - Improving phone update / draw mode?

I am creating an Android phone application and am experiencing a strange "throttling". I believe that this is due to the fact that the semaphore is called to stop what the application does to process something else on the phone. Although I'm not sure.

I am curious if there will be a way to distribute these breakthroughs or something else to make it less noticeable to users that the application has slight delays in speaking.

Edit: some additional information I'm currently working with is the 2nd array of images β€” only about 80 drawings during instanced ~ 8000. They only draw if their color cast is not RGB0 (black). The update run cycle checks which images are closest to the player and gives them a minimum minimum RGB illumination of 0.2f. In addition, the basic event handlers and move / view-port cycles in the update. Please note that I am using the Libgdx platform, not the native one. So, OpenGL, etc.

Edit: I want to note that the problem is not what you think. I sent vector2 about 3,800 times of rendering, but not just β€œsent” one, but announced a new Vector2 and sent the parameters this way. THE COLLECTOR ASSEMBLY DOES NOT ALLOW THIS SLOPE. Running smoothly now that I am sending just 2 floats. My bad ._.

+3
source share
2 answers

If several active processes do not work at the same time, this should not happen.

My guess (without seeing the code) is that the garbage collector kicks too much. Do you initialize objects in a loop? If so, can you reuse objects instead?

, Views. , , .

, , - : , ? (.. view.invalidate(rect); view.invalidate();)

, ? . , RelativeLayout LinearLayouts.

: " " , Google I/O 2009. , .

, , 8K . , , , GC , . , . , , Q & A 53-55 . , , HashMap , GC , . . , .

+4

/ , , , - (, , , ) .

. AsyncTask .

0

All Articles