Walberla with Openlb and Cuda Walberla/walberlav6.1/apps/tutorials/lbm/03_LBLidDrivenCavity.cpp
Good afternoon y'all. So i am wondering how can i easily convert the Walberla/walberlav6.1/apps/tutorials/lbm/03_LBLidDrivenCavity.cpp application so support both Openmp and Cuda. I have played around with the same application trying to use more threads in the enviroment with the use of export OMP_NUM_THREADS = [number of threads] seems that i get worse performance if the number of threads is more than one. So my question is how can i impliment threading in walberla? My goal is to run walberla in HYBRID (Openmp with MPI) mode for the lid-driven app and with CUDA. I have seen this function for Openmp, do i think around the same lines? In normal Openmp pragma omp{} starts the parallel section what would be the equivalent in Walberla? | #define WALBERLA_FOR_ALL_CELLS_XYZ WALBERLA_FOR_ALL_CELLS_XYZ_OMP( field, omp parallel for schedule(static), CODE ) And for cuda do i have to write something new completely?