用OpenCL与OpenGL互操作实现雷达终端的高性能绘制

Realization of high-performance rendering of radar terminal based on OpenCL and OpenGL interoperability

随着雷达终端数据的增长,对图形处理器(GPU)采用OpenGL等绘制工作很难满足当前雷达终端显示需求.针对现有GPU绘制中处理大量目标数据存在资源利用率低以及带宽延迟大的问题,基于通用平台的GPU架构,提出一种OpenCL与OpenGL互相操作的加速方法.该方法采用GPU计算和绘制同时进行,通过设计回波数据的GPU计算,实现P、B、A显的回波计算与纹理绘制.应用结果表明,该方法高效地实现了GPU计算和绘制性能,从而实现了雷达终端高性能绘制,提高了终端显示的流畅性.

With the growth of radar terminal data, it is difficult to meet current radar terminal display requirements for graphic processing unit （GPU） rendering by using open graphic language （OpenGL）. Aiming at the problem of low resource utilization and large bandwidth delay in processing a large amount of target data in the existing GPU rendering, and based upon GPU architecture of a common platform, this paper proposes an acceleration method for interoperating open computing language （OpenCL） and OpenGL. The proposed method adopts GPU computing and rendering simultaneously, and performs the GPU calculation of the echo data to realize echo calculations and texture rendering of PPI, B-scope and A-scope display. The application results show that the proposed method efficiently achieves GPU computing and rendering performance, which enables high-performance radar terminal rendering and improves the terminal display fluency.