GPU-based parallel map slicing method and system

Abstract

The invention discloses a GPU-based parallel map slicing method and system. The method includes the steps of S100. performing dynamic mesh division on an original raster map according to the core number of a GPU to obtain cells whose number is not larger than N, i.e., slice regions, one slice region corresponding to a slicing task; S200. allocating and submitting the slicing tasks to the GPU side,specifically including mapping a slicing task to a thread, a preset number of adjacent threads forming a thread block, sliding task allocation using thread block as a basic unit, and threads contained by each thread bock being allocated to streaming processors in the GPU; and S300. parallelly executing slicing on the slice regions by the streaming processors in the GPU side. The GPU-based parallel mapping slicing method provided by the invention is efficient and is easy to realize, maintenance cost and complexity of parallel computing of the GPU are relatively low, the computing cost performance is high, and operation cost of an enterprise can be greatly saved. The method is applied to a WebGIS, and rapid seamless browse of images can be realized.

Description

technical field [0001] The invention belongs to the technical field of map slicing, and in particular relates to a parallel map slicing method and system based on a GPU (Graphics Processing Unit). Background technique [0002] In recent years, with the rapid development of geographic information technology and the continuous improvement of the resolution of space remote sensing instruments, the field of map application has continued to expand, and the amount of raster map data has increased dramatically. With the rapid growth of raster map data, WebGIS (Web Geographic Information System) usually needs to slice raster map data in advance and organize raster map data in the form of tile pyramids. When the client accesses WebGIS to obtain map data, the server returns the pre-sliced ​​tile data to the client. Through the above process, the speed of browsing massive raster maps on the web is guaranteed. [0003] How to efficiently slice these raster map data is the key problem ...

AI Technical Summary

Problems solved by technology

The CPU-based multi-threaded parallel slicing method, due to the limitation of the parallel computing capability of the CPU, a general-purpose processor, in terms of the number of cores, threads, and data throughput, the efficiency of parallel slicing is also very low

Parallel slices based on high-performance computer clusters or high-end distributed environments generally perform parallel operations based on message passing interfaces. During the parallel work process, a large amount of communication will be generated between nodes, resulting in reduced efficiency, and the cost of calculation and the complexity of maintenance. Extremely high, not suitable for large-scale promotion and use

Images