Using OpenMP to parallelize small programs


I’m working in a project in the University of Buenos Aires to use a new architecture for improve fine grain TLP. Our idea is to take the llvm bitcode and convert it into our new architecture by using the omp annotations.

I would like to know if you could bring documentation about the openmp bitcode for llvm? or about the annotations?

Thanks in advance.

Hello David,

What do you mean by “omp annotations”?

There is an implementation of OpenMP support in clang ( but it inserts OMP runtime library calls right into front-end. LLVM IR doesn’t have any special support for any kind of OMP annotations.


(disclaimer: I work for Intel; the OpenMP implementation linked above is created mostly by Intel.)