how to control OMP Thread number between libraries
I meet a problem when using iomp.
I have a dynamic library A, and export some functions within it, this functions may be use openmp for parallel compute.
then I implement a executable application, and call some functions from library A. this executable application will be called by others by multi threads(also with iomp). so, when it runs, the number of threads are too much.
how can i control iomp thread numbers functions in A used,but without modify library A?