Massively parallel non-stationary EEG data processing on GPGPU platforms with Morlet continuous wavelet transform

Abstract

Morlet continuous wavelet transform (MCWT) has been widely used to process non-stationary electroencephalogram (EEG) data. Nowadays, the MCWT application for processing EEG data is time-sensitive and data-intensive due to quickly increasing problem domain sizes and advancing experimental techniques. In this paper, we proposed a massively parallel MCWT approach based on GPGPU to address this research challenge. The proposed approach treats MCWT as four main computing sub-procedures and parallelizes them with CUDA correspondingly. We focused on optimizing FFT on GPUs to improve the performance of MCWT. Extensive experiments have been carried out on Fermi and Kepler GPUs and a Fermi GPU cluster. The results indicate that (1) the proposed approach (especially on Kepler GPU) can ensure encouraging runtime performance of processing non-stationary EEG data in contrast to CPU-based MCWT, (2) the performance can further be improved on the GPU cluster but performance bottleneck exists when running multiple GPGPUs on one

Using a GPU to Accelerate a Longwave Radiative Transfer Model with Efficient CUDA-Based Methods

“…(5) The RRTMG_LW includes aerosol absorption capability. (6) The RRTMG_LW can be used as a callable subroutine and can be adapted for use within global or regional models. 7The RRTMG_LW can optionally read the required input data either from a netCDF file or from the original RRTM_LW source data statements.…”

“…Earth system models (ESMs) have a large amount of calculation and high resolution, so HPC is widely used to accelerate their computing and simulation [4]. In the past few years, the modern graphics processing unit (GPU), which combines features of high parallelism, multi-threaded multicore processor, high-memory bandwidth, general-purpose computing, low cost, and compact size far beyond a graphics engine, has substantially outpaced its central processing unit (CPU) counterparts in dealing with data-intensive, computing-intensive, and time-intensive problems [5][6][7][8][9]. In the era of pursuing green computing, the booming GPU capability has attracted more and more scientists and engineers to use GPUs instead of CPUs to accelerate climate system models or ESMs [10,11].…”

A Novel GPU-Based Acceleration Algorithm for a Longwave Radiative Transfer Model

“…To greatly improve the computational performance, it is beneficial to use high-performance computing (HPC) technology to accelerate the RRTMG.At present, HPC is widely employed in earth climate system models [13][14][15]. With the rapid development of HPC technology, due to the features of multithreaded many-core processor, high parallelism, high memory bandwidth, and low cost, the modern graphics processing unit (GPU) has substantially outpaced its central processing unit (CPU) counterparts in dealing with data-and computing-intensive problems [16][17][18][19]. Currently, increasing numbers of atmospheric applications were accelerated by the GPUs [20,21].…”