大视场综合孔径成像优化的研究

In recent years, synthetic aperture telescope showing a wide field, high resolution and wide frequency range, large area characteristics, and then rapidly developed, and the SKA is brought about by the mass production of astronomical data, so we need an efficient data processing software to meet the vast amounts of data.GPU is a kind of multiple cell processor, it is more suitable for complete large-scale repeat operation, has been in the some of the synthetic aperture radar imaging algorithm application, and also got the efficiency optimization of multiplication.So we need a comprehensive optimization of the software for synthetic aperture radar imaging algorithm, and the application of this software to the super performance calculation on a cluster, has reached the software and hardware and optimal analysis of data, and we have 21 cma data and experience for many years.

近年来，综合孔径望远镜呈现出大视场、高分辨率、超宽的频率范围、超大的接受面积的特点，继而平方公里阵SKA的快速开展，带来的是海量的天文数据的产生，因此我们需要一种高效的数据处理软件来迎接海量的数据。图形处理器GPU是一种多计算单元的处理器，它更加适用于完成大规模的重复的运算，已经在综合孔径成像中的一些算法中得到了应用，而且也得到了效率倍增的优化。因此我们需要一种对综合孔径成像算法全面优化的软件，并把此软件应用到超性能计算集群上，已达到软件和硬件并优的数据分析，而且我们有21CMA多年的数据积累和经验。