基于GPU的蒙特卡洛放疗剂量并行计算

摘要

目的:蒙特卡洛模拟在放疗剂量计算领域被广泛视为最精确的计算方法,但对于日常的临床应用,其效率仍有较大提升需求和空间.方法:本文会呈现放疗剂量计算领域的最新成果-维持相同的粒子输运原理的同时,使用CUDA语言,利用显卡的GPU(Graphic Processing Unit)并行处理蒙特卡洛计算中的主要过程,计算光子剂量沉积.这样既可以保证不失去蒙卡模拟的精度,又可以极大地提高运算速度.结果:实践表明在使用NVIDIA GTX460 1G DDR5 plus INTEL i52300的硬件设备,在GPU上并行计算蒙特卡洛放疗剂量沉积时,计算100万个光子剂量沉积时加速因子达到116.6,处理1000万光子入射,加速因子可达127.5.结论:本文中利用显卡GPU运行CUDA语言对放疗剂量计算进行模拟,是一种可以大幅有效提高剂量计算效率方法.%Objective: Monte Carlo simulation is commonly considered to be the most accurate dose calculation method in radiotherapy. However, its efficiency still requires improvement for many routine clinical applications.Methods: This paper will present recent progresses in GPU-based Monte Carlo dose calculation. We utilizes the parallel computation ability of a GPU to achieve high efficiency, while maintaining the same particle transport physics as in the original Monte Carlo simulation code and therefore obtains the same level of simulation accuracy. Results: Our research results show that using an NVIDIA GTX460 GPU card against an INTEL i5 2300 in computing a one-million sample with all 336 processor cores working together,speed-up factors can be as high as 116.6,as for a ten-million situation,even obtain a result as high as 127.5. Conclusions:Using GPU and CUDA to process a Monte Carlo simulation can highly improve the efficiency of dose calculation.