Quantum transport in NEMO5: Algorithm improvements and high performance implementation

Author

Yu He ; Kubis, Tillmann ; Povolotskyi, Michael ; Fonseca, J. ; Klimeck, Gerhard

Author_Institution

Network for Comput. Nanotechnol., Purdue Univ., West Lafayette, IN, USA

fYear

2014

fDate

9-11 Sept. 2014

Firstpage

361

Lastpage

364

Abstract

Quantum transport algorithms such as QTBM and NEGF/RGF have been efficiently implemented in the multi-scale simulation tool NEMO5 by taking advantage of the Hamiltonian´s characteristics of nanowires without explicit spinorbit coupling in the tight binding representation. Benchmarks in a 3nm diameter, 20 nm length Si nanowire in atomistic 10 band tight binding representation demonstrate 3-5 times performance improvement over the current state of the literatures.

Keywords

Green´s function methods; elemental semiconductors; matrix algebra; nanowires; semiconductor quantum wires; silicon; Hamiltonian´s characteristics; NEGF algorithm; NEMO5; QTBM algorithm; RGF algorithm; Si; algorithm improvements; atomistic 10 band tight binding representation; high performance implementation; multiscale simulation tool; nonequilibrium Green´s functions method; performance improvement; quantum transmitting boundary method; quantum transport algorithm; silicon nanowire; size 20 nm; size 3 nm; Algorithm design and analysis; Eigenvalues and eigenfunctions; Equations; Mathematical model; Nanowires; Slabs; Sparse matrices; NEGF; QTBM; RGF; high performance implementation; quantum transport; self-energy;

fLanguage

English

Publisher

ieee

Conference_Titel

Simulation of Semiconductor Processes and Devices (SISPAD), 2014 International Conference on

Conference_Location

Yokohama

ISSN

1946-1569

Print_ISBN

978-1-4799-5287-8

Type

conf

DOI

10.1109/SISPAD.2014.6931638

Filename

6931638