The performance potential for simulating quantum electron transport on graphical processing units (GPUs) is studied. Using graphene ribbons of realistic sizes as an example it is shown that GPUs provide significant speed-ups in comparison to central processing units as the transverse dimension of the ribbon grows. The recursive Greenʼs function algorithm is employed and implementation details on GPUs are discussed. Calculated conductances were found to accumulate significant numerical error due to single-precision floating-point arithmetic at energies close to the charge neutrality point of the graphene.
The present study showed that electron quantum transport simulations might be effectively done using the GPU architecture that outperforms the CPU even for a low-cost build-in hardware solution. Computational speed-ups might be higher by large factors for “cost-comparable” hardware. The recursive Green’s function algorithm can be straightforwardly implemented in the CUDA framework. Highly optimized CUBLAS and CULA libraries allow one-to-one replacement of BLAS and LAPACK libraries. However, because the GPU architecture operates on single-precision floating-point data substantial numerical error might accumulate.
S. Ihnatsenka. Computation of electron quantum transport in graphene nanoribbons using GPU. Computer Physics Communications. Volume 183, Issue 3, March 2012, Pages 543–546. arXiv:1107.5300v1 [physics.comp-ph] [doi: 10.1016/j.cpc.2011.11.019] [Free PDF]