In Geoinformation Systems for Mining a wide range of computational methods for creating polygonal and voxel models, including methods of geostatistics, triangulation, interpolation and optimization of resource extraction boundaries, are used. For the solution of the mining task requires the construction of detailed voxel or block models for operational planning of discrete volumes for rocks under excavation. The article shows the possibilities of implementing GPU acceleration to perform parallel calculations on voxelization of the surfaces of the media and volume geological bodies with the use of CUDA and OpenCL programming technology. In the Gexoblock system the library NVIDIA voxelization based on parallel computing technology CUDA gvdb-voxels is used. The proposed hybrid approach includes methods for parallelizing triangulation, interpolation, and optimization of the sequence of extraction steps in the search for the best extraction strategy for ore reserves.
In Geoinformation Systems for Mining a wide range of computational methods for creating polygonal and voxel models, including methods of geostatistics, triangulation, interpolation and optimization of resource extraction boundaries, are used. For the solution of the mining task requires the construction of detailed voxel or block models for operational planning of discrete volumes for rocks under excavation. The article shows the possibilities of implementing GPU acceleration to perform parallel calculations on voxelization of the surfaces of the media and volume geological bodies with the use of CUDA and OpenCL programming technology. In the Gexoblock system the library NVIDIA voxelization based on parallel computing technology CUDA gvdb-voxels is used. The proposed hybrid approach includes methods for parallelizing triangulation, interpolation, and optimization of the sequence of extraction steps in the search for the best extraction strategy for ore reserves.