I have a single-thread application which does 3D rendering; it is a moderate-heavy user of the single core it uses (~50%). In my quick tests, recording the rendered output to video using libav is fastest when I use AVCodecContext::thread_count == 3 on my quad-core PC, but it's almost identical with it set to 0.
As well as any general advice you can offer, a couple of specific questions... 1)What does setting AVCodecContext::thread_count = 0 do (on Windows x86/64 hardware if it is platform-dependent)? Does it default to num-cores or nothing that simple? 2)In a situation similar to mine with a single-thread app using most of one core on a N-core CPU, what would you set AVCodecContext::thread_count to, and why? Thanks, John. _______________________________________________ Libav-user mailing list [email protected] http://ffmpeg.org/mailman/listinfo/libav-user
