Hi Guys,
I'm trying to take the elementary stream from the codec engine output,
and pass it through libavformat for packetising / multiplexing and
further output.
However, by passing this data through libavformat, I end up utilizing
100% cpu usage to achieve around 18fps. Not particularly adequate :(
To start to analyse this usage, I wanted to profile the dvsdk demo code,
but compiling the dvsdk-demos' with -pg for gprof using montavista's
toolchain is causing bus-errors and not getting me very far.
Has anyone used any other profiling tools on this platform ? I think
I'll give oprofile a go next, but I was wondering if anyone else has
actually done all this before.
What are people using to process the elementary streams from the codec
engine outputs to make them more useful - or do you just leave the video
as m2v / m4v?
--
Regards
Kieran Bingham
_______________________________________________
Davinci-linux-open-source mailing list
[email protected]
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
