MediaDecoderStateMachine finds that the frame should be rendered, it sends to Compositor(b2g process) when the frame is decoded, mediaserver sends the data back to video appĬ. mediaserver gets the frame from video appī. On FxOS, we get the decoded data from HW codec, keep it in a state machine, and decide when we can display it. FxOS has internal synchronization mechanism.įor Android, the encoded data is sent to codec then render it. ![]() Here are some reasons about why FxOS uses more CPU time than Android.Ģ. So it means that when playing 20 seconds video clip, FxOS uses 203.186 ms more CPU time than Android. Then you average them, is it right? And the result shows that FxOS uses more 203.186 ms of CPU time than Android. Even I set the pid to set_ftrace_pid, set 1 to tracing_on.įrom comment 0, you mentioned that you get the result by testing 5 times and every time capturing 20 seconds of data. ![]() But it seems that we should change some kernel configs to enable the tracers? I cannot get any data from my nexus 5.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |