ERF_ChopGrids.cpp File Reference

#include <ERF_Utils.H>

Include dependency graph for ERF_ChopGrids.cpp:

Functions
BoxArray	ERFPostProcessBaseGrids (const Box &domain, bool decompose_in_z)
void	ChopGrids2D (BoxArray &ba, const Box &domain, int target_size)

Function Documentation

ChopGrids2D()

void ChopGrids2D	(	BoxArray &	ba,
		const Box &	domain,
		int	target_size
	)

{
    IntVect chunk = domain.length();
 
    while (ba.size() < target_size)
    {
        IntVect chunk_prev = chunk;
 
        std::array<std::pair<int,int>,AMREX_SPACEDIM>
            chunk_dir{std::make_pair(chunk[0],int(0)),
                      std::make_pair(chunk[1],int(1))};
        std::sort(chunk_dir.begin(), chunk_dir.end());
 
        // We only decompose in and y
        for (int idx = 1; idx >= 0; idx--) {
            int idim = chunk_dir[idx].second;
            int new_chunk_size = chunk[idim] / 2;
            if (new_chunk_size != 0)
            {
                chunk[idim] = new_chunk_size;
                ba.maxSize(chunk);
                break;
            }
        }
 
        if (chunk == chunk_prev) {
            break;
        }
    }
}

Referenced by ERF::initHSE(), and make_terrain_fitted_coords().

Here is the call graph for this function:

Here is the caller graph for this function:

ERFPostProcessBaseGrids()

BoxArray ERFPostProcessBaseGrids	(	const Box &	domain,
		bool	decompose_in_z
	)

{
    //
    // This is used to avoid the case where the native amrex decomposition makes
    //     too many grids for the number of processors.
    //
    // The idea is to not override the user preference if expressed by max_grid_size
    //     but instead to ensure that the default behavior is what we want.
    //
    BoxArray ba0 = amrex::decompose(domain, ParallelDescriptor::NProcs(),
                                    {true,true,decompose_in_z});
    return ba0;
}

Referenced by ERF::MakeNewLevelFromScratch().

Here is the caller graph for this function: