I have some questions related to my application that does video decoding (written in C) and displays a decoded frame (I use NDK):
1- On Android (runs on a Cortex A-8 processor with a frequency of 1 GHz), the application runs about 10 times slower than windows (runs on a 3 GHz processor). I expect 3-4 times slower, but not 9-10 times slower. (Mostly frames come at a very low speed)
2- I thought I could not compile for the right purpose (Cortex A 8). Therefore, I edited the Application.mk file.
APP_ABI := armeabi armeabi-v7a mips
Is there any other .mk file that I need to edit?
Is this inclusion enough to suggest that I am compiling for the right purpose?
(I ask for this since I am not getting any visible improvements even after turning this on. Is there a way that I can be sure that I am compiling for the right purpose (Cortex A8).
What could be the reason for getting such a huge performance difference between Android and Windows, although Android is a real-time operating system, Cortex is a dual-processor processor with Neon support, etc.
Additional information and my analysis:
1-I should also mention that I use the main thread to update the interface (calling Ondraw ()) and another thread to decode.
2 - , , . , , ; , , , . windows , , . ( , .)