{"title":"Parallel Implementation of Boltzmann Transport Simulation of Carbon Nanotubes","authors":"Z. Akšamija, Mohamed Mohamed, Umberto Ravaioli","doi":"10.1109/IWCE.2009.5091137","DOIUrl":null,"url":null,"abstract":"This works treats electronic transport in SWNTs in the Boltzmann Transport equation (BTE) formalism. The BTE is solved self-consistently with the Poisson equation and iterated in time using an upwinding finite-difference scheme until a steady-state is reached. Phonon scattering is included through a relaxation time based on experimental values reported in the literature. The problem is parallelized by dividing the real space into strips, where each strip is assigned to one processing element to minimize communication overhead. The implementation was tested on a many-processor cluster and shows good speed-up over the serial code. This demonstrates that the code is capable of excellent scaling to large supercomputing machines for large-scale parallel simulation of nanotubes, as well as other similar 1-dimensional materials like nanoribbons and nanowires.","PeriodicalId":443119,"journal":{"name":"2009 13th International Workshop on Computational Electronics","volume":"38 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2009-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2009 13th International Workshop on Computational Electronics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/IWCE.2009.5091137","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
Abstract
This works treats electronic transport in SWNTs in the Boltzmann Transport equation (BTE) formalism. The BTE is solved self-consistently with the Poisson equation and iterated in time using an upwinding finite-difference scheme until a steady-state is reached. Phonon scattering is included through a relaxation time based on experimental values reported in the literature. The problem is parallelized by dividing the real space into strips, where each strip is assigned to one processing element to minimize communication overhead. The implementation was tested on a many-processor cluster and shows good speed-up over the serial code. This demonstrates that the code is capable of excellent scaling to large supercomputing machines for large-scale parallel simulation of nanotubes, as well as other similar 1-dimensional materials like nanoribbons and nanowires.