A high-performance GPU implementation of the electron-phonon Wannier interpolation and the related transport properties

The electron-phonon Wannier interpolation (EPWI) method is an efficient way to compute the properties of electron-phonon interactions (EPIs) accurately. This study presents a GPU-accelerated implementation of the EPWI method for computing transport properties, followed by a performance analysis. The implementation is tested on common systems such as aluminum andsilicon. The results show complete consistency with those obtained through CPU computations. The proposed algorithm has the capability of computing the conductivity of aluminum in 20 minutes on a single NVIDIA Tesla V100 GPU, adopting a 2003 electron and phonon sampling grid. This speed is 54 times higher than the CPU-based algorithm, running on two nodes of the Intel Xeon Platinum 8260 CPU. Such impressive acceleration is achieved by carefully designing the algorithm to exploit the GPU’s specific features. The idea of this approach can also be applied to the GPU code for other EPI-related properties. This record includes the input/output files of this work. The code snippets of the core GPU implementation can be found in https://github.com/zheliu137/EPWI_BTE_GPU