Hi,
I am currently evaluating whether we will/can/want to use OpenMP in a new project on a multicore DSP. As I understood so far, if we use OpenMP, the DSP is running the OpenMP Runtime and not Sysbios - or is the Runtime on top of Sysbios? However, when using a parallel region, the runtime will use one thread per core to share the work, thus I do not have to worry about running and syncing several instances of Sysbios. I further understand that OpenMP usually creates pthread-code internally - is that true for the TI implementation?
My main question is as follows: We want to use some rather large libraries which were designed to be multithreaded, but not with/for OpenMP. So can I still create pthreads when using OpenMP? And if I can, will the runtime spread those pthreads across the cores?