The application is using opencv and ffmpeg, we want to build the opencv, such that it will leverage all the parallel processing and GPU hardware available under that particular os. can you please list all compiler level flags which can be used to improve the performance of the opencv while compiling.