HMLP: High-performance Machine Learning Primitives
|
Classes | |
class | Comm |
class | Context |
Functions | |
template<typename FUNC , typename... Args> | |
void | Parallelize (tci::Comm *comm, FUNC func, Args &&...args) |
end class Lock
void hmlp::tci::Parallelize | ( | tci::Comm * | comm, |
FUNC | func, | ||
Args &&... | args | ||
) |
end class Comm
Create a shared context pointer for communication.
Create a parallel section with omp_get_num_threads().
Create a global communicator with the shared context.
Now call the function in parallel.
end pragma omp parallel