Hi,
Sorry for replying late but I was in hospital. Regarding the GPU performance question:
I have a PII X4 (comparable CPU performance to A8-3850) and a GCN core GPU with much more shaders than a8-3850. In my case The GPU accelerated code path is at most 50% faster and is very parameter/codec dependent. Sometimes the performance difference is within 10%. So with a much weaker GPU I can believe there are situations /parameter combinations when the GPU code path is slower.
Subtitles:
As I have written in the readme my program is only a frontend to ATI/AMD directshow filters and the functions are fixed in these filters.
|
