A hybrid MPI-CUDA approach for nonequispaced discrete Fourier transformation

Yang, Sheng-Chun; Wang, Yong‐Lei

doi:10.1016/j.cpc.2020.107513

Cited by 4 publications

(4 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, we proposed a hybrid parallel scheme combining multiple CPU and GPU devices to upgrade the CU-ENUF method, which is described as HP-ENUF method. [44,45] Similar to the CU-ENUF method, [43]…”

Section: Architecture Of the Hp-enuf Methodsmentioning

confidence: 99%

“…[8,40,41] In addition, several (hybrid) parallelization strategies based on gridding [42] and Near-Distance algorithms [43] have been developed to accelerate the evaluation of electrostatic energies and forces using GPU and CUDA technology. [44,45] These derivatives of the ENUF and ENUF-DPD methods exhibit distinct computational efficiencies in handling long range electrostatic interactions between charged particles and charge density distributions at multiple spatiotemporal scales.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

TheENUFmethod—Ewald summation based onnonuniformfast Fourier transform: Implementation, parallelization, and application

Yang

Zhu

et al. 2020

J Comput Chem

Self Cite

View full text Add to dashboard Cite

Computer simulations of model systems are widely used to explore striking phenomena in promising applications spanning from physics, chemistry, biology, to materials science and engineering. The long range electrostatic interactions between charged particles constitute a prominent factor in determining structures and states of model systems. How to efficiently calculate electrostatic interactions in simulation systems subjected to partial or full periodic boundary conditions has been a grand challenging task. In the past decades, a large variety of computational schemes has been proposed, among which the Ewald summation method is the most reliable route to accurately deal with electrostatic interactions between charged particles in simulation systems. In addition, extensive efforts have been done to improve computational efficiencies of the Ewald summation based methods. Representative examples are approaches based on cutoffs, reaction fields, multi‐poles, multi‐grids, and particle‐mesh schemes. We sketched an ENUF method, an abbreviation for the Ewald summation method based on the nonuniform fast Fourier transform technique, and have implemented this method in particle‐based simulation packages to calculate electrostatic energies and forces at micro‐ and mesoscopic levels. Extensive computational studies of conformational properties of polyelectrolytes, dendrimer‐membrane complexes, and ionic fluids demonstrated that the ENUF method and its derivatives conserve both energy and momentum to floating point accuracy, and exhibit a computational complexity of scriptO)(NlogN with optimal physical parameters. These ENUF based methods are attractive alternatives in molecular simulations where high accuracy and efficiency of simulation methods are needed to accelerate calculations of electrostatic interactions at extended spatiotemporal scales.

show abstract

Section: Architecture Of the Hp-enuf Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

TheENUFmethod—Ewald summation based onnonuniformfast Fourier transform: Implementation, parallelization, and application

Yang

Zhu

et al. 2020

J Comput Chem

Self Cite

View full text Add to dashboard Cite

show abstract

“…An other research group [2] also proposes GPU sparse FFT algorithm based on parallel optimization and the authors mention that this algorithm "leads enormous speedups". Moreover, in a very recent research [23], authors propose an hybrid MPI -CUDA implementation for nonequispaced discrete Fourier transformation using parallel threads launched from CPU nodes for managing the thread-level parallelism in multiple GPU devices. The authors prove that using hybrid parallelization, an increased improvement in computational efficiency is obtained without losing the computational precision.…”

Section: Cuda For Fourier Transformmentioning

confidence: 99%

“…The authors prove that using hybrid parallelization, an increased improvement in computational efficiency is obtained without losing the computational precision. Also, their method can balance in a dynamic way the connection between performance and throughput capacity by modifying the number of computer nodes used for parallel computations [23].…”

Section: Cuda For Fourier Transformmentioning

confidence: 99%

Advances in CUDA for computational physics

Spiridon

2023

BUT_Series_III

View full text Add to dashboard Cite

Advances in the graphics processing unit (GPU) development led to the opportunity for software developers to increase the execution speed for their programs by massive parallelization of the algorithms using GPU programming. NVIDIA company developed an arhitecture for parallel computing named Compute Unified Device Architecture (CUDA) which includes a set of CUDA instructions and the hardware for parallel computing. Computational Physics is an interdisciplinary field which is in continuous progress and which studies, develops and optimizes numerical algorithms and computational techniques for their application in solving various physics problems. Computational Physics has applicability in all sub-branches of physics and related fields such as: biophysics, astrophysics, plasma physics, biomechanics, fluid physics, etc. Moreover, with the evolution of technology in the last few decades, this relatively new field has helped to quickly obtain results in these fields, facilitating the connection between theoretical and experimental physics. In this paper, some of the latest researches and results obtained in computational physics by using GPU computing with CUDA architecture are reviewed.

show abstract