CUDAGIS Design
 |
We report the design and realization of a high-performance parallel GIS, i.e., CudaGIS, based on the General Purpose computing on Graphics Processing Units (GPGPU) technologies. Still under active developments, CudaGIS currently supports major types of geospatial data (point, polyline, polygon and raster) and provides modules for spatial indexing, spatial join and other types of geospatial operations on such geospatial data types. Experiments have demonstrated 20-40X and 1000-10000X speedups over serial CPU implementations on main-memory and disk-resident systems, respectively. |
Jianting Zhang and Simin You. CudaGIS: Report on the Design and Realization of a Massive Data Parallel GIS on GPUs. To appear in ACM SIGSPATIAL IWGS Workshop. [PDF]