Postdoc: Parallel Finite-Difference Time-Domain Computation of Electromagnetic Absorption in the Human Body

9504074 Gandhi本研究将比较几种用于大规模人体模型研究的并行化EM FDTD代码的方法。异构工作站集群将用于研究代码的可移植性、效率和可伸缩性。这将与高性能Fortran （HPF）和大规模并行机器的使用进行比较。大规模输出数据的可视化和FDTD晶格的子网格划分也将并行化。利用并行代码实现的高分辨率模型，将研究电力线与松果体和脑垂体的耦合，这可能涉及EM生物相互作用。下一代1900兆赫的移动电话将被测试，并将开发一个脉冲响应库，允许分析人体与高达3ghz的任何波形的耦合。

9504074 Gandhi This research will compare several methods of parallelizing the EM FDTD code for large-scale man moll studies. Heterogeneous workstation clusters will be used to study code portability, efficiency, and scalability. This will be compared to use of High Performance Fortran (HPF) and also massively parallel machines. Visualization of the large-scale output data and subgridding of the FDTD lattice will also be parallelized. With the high-resolution model made possible by the parallel code, coupling of power lines to the pineal and pituitary glands, which may be implicated in EM biointeraction, will be studied. Next-generation cellular telephones at 1900 MHz will be examined, and a library of impulse responses will be developed, allowing analysis of coupling of the body to any waveform up to 3 GHz.