https://hgpu.org/?p=4220
Acceleration of LOD-FDTD Method Using Fundamental Scheme on Graphics Processor Units