GPUQT: An efficient linear-scaling quantum transport code fully implemented on graphics processing units

摘要

We present GPUQT, a quantum transport code fully implemented on graphicsprocessing units. Using this code, one can obtain intrinsic electronictransport properties of large systems described by a real-space tight-bindingHamiltonian together with one or more types of disorder. The DC Kuboconductivity is represented as a time integral of the velocity auto-correlationor a time derivative of the mean square displacement. Linear scaling (withrespect to the total number of orbitals in the system) computation time andmemory usage are achieved by using various numerical techniques, includingsparse matrix-vector multiplication, random phase approximation of trace,Chebyshev expansion of quantum evolution operator, and kernel polynomial methodfor quantum resolution operator. We describe the inputs and outputs of GPUQTand give two examples to demonstrate its usage, paying attention to theinterpretations of the results.