This folder contains two Python functions for GPU-accelerated simulation, which implements the recursive algorithm in the non-equilibrium Green’s function (NEGF) formalism.

Compared to the matlab implementation [1], the GPU version allows massive parallel running over many cores on GPU hardware. For example, calculation of the Green’s function on energy grids can be paralleled over GPU cores.

This folder contains two functions:

recuresealg3d.py

- which works for 3-diagonal matrices

recuresealgblock3d.py

- which works for 3-block-diagonal matrices

The explanations of the arguments and the calling sequence are in file headers.

Adding command “cuda.current_context().reset()” in the end of your main program is necessary to clear GPU memory after usage.

Researchers should cite this work as follows:

• [1] Dmitri Nikonov, Siyu Koswatta (2006), "recursive algorithm for NEGF in Matlab," https://nanohub.org/resources/1983

• Ning Yang, Tong Wu, Jing Guo (2021), "Recursive algorithm for NEGF in Python GPU version," https://nanohub.org/resources/34777.

  BibTex | EndNote

1. algorithms
2. GPU
3. nanoelectronics
4. nanotransistors
5. NEGF
6. parallel programming
7. programming techniques
8. python
9. quantum transport
10. zuki