期刊
JOURNAL OF BIOMEDICAL OPTICS
卷 17, 期 9, 页码 -出版社
SPIE-SOC PHOTO-OPTICAL INSTRUMENTATION ENGINEERS
DOI: 10.1117/1.JBO.17.9.090504
关键词
peer-to-peer; online, web applications; object oriented; Monte Carlo; parallel computing; graphics processing units; compute unified device architecture; photon migration; diffuse reflectance; fluence rate distribution
In the framework of further development of the unified approach of photon migration in complex turbid media, such as biological tissues we present a peer-to-peer (P2P) Monte Carlo (MC) code. The object-oriented programming is used for generalization of MC model for multipurpose use in various applications of biomedical optics. The online user interface providing multiuser access is developed using modern web technologies, such as Microsoft Silverlight, ASP. NET. The emerging P2P network utilizing computers with different types of compute unified device architecture-capable graphics processing units (GPUs) is applied for acceleration and to overcome the limitations, imposed by multiuser access in the online MC computational tool. The developed P2P MC was validated by comparing the results of simulation of diffuse reflectance and fluence rate distribution for semi-infinite scattering medium with known analytical results, results of adding-doubling method, and with other GPU-based MC techniques developed in the past. The best speedup of processing multiuser requests in a range of 4 to 35 s was achieved using single-precision computing, and the double-precision computing for floating-point arithmetic operations provides higher accuracy. (C) 2012 Society of Photo-Optical Instrumentation Engineers (SPIE). [DOI: 10.1117/1.JBO.17.9.090504]
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据