改进的模拟植物生长算法及其在大型辐照装置自动排源中的实用化

Improved Plant Growth Simulation Algorithm and Its Practical Application for Source Bars Arrangement at Big Irradiation Facility

  • 摘要: 在相关研究工作的基础上,分析得到了基于模拟植物生长算法的排源算法的性能热点,并对其进行了并行化改进,具体是将原有串行执行过程转化为统一计算设备架构(CUDA)的线程块并行执行过程,从而可有效利用GPU的大规模并行能力,实现程序执行效率的大幅加速。若干计算实例结果表明,在保持排源算法原有优点的基础上,新算法的整体性能有最低30倍的提高,使当前的排源算法具有高度的竞争力。

     

    Abstract: Based on the previous research work, performance hotspot of source bars deployment algorithm derived from plant growth simulation algorithm was obtained. It was improved with parallelization principle, in which the original sequential process was converted into that of parallel compute unified device architecture (CUDA) thread blocks, and the massively parallel processing ability of GPU can be effectively used, which makes the performance of algorithm greatly speed up. Several examples show that, based on the advantage of new algorithm, integral efficiency is increased by 30 times at least, so the new algorithm is highly competitive.

     

/

返回文章
返回