Parallel electronic structure calculations using multiple graphics processing units (GPUs)

Authors:
Samuli Hakala;Ville Havu;Jussi Enkovaara;Risto Nieminen
Affiliations:
COMP Center of Excellence, Department of Applied Physics, School of Science, Aalto University, Aalto, Finland;COMP Center of Excellence, Department of Applied Physics, School of Science, Aalto University, Aalto, Finland;COMP Center of Excellence, Department of Applied Physics, School of Science, Aalto University, Aalto, Finland;COMP Center of Excellence, Department of Applied Physics, School of Science, Aalto University, Aalto, Finland
Venue:
PARA'12 Proceedings of the 11th international conference on Applied Parallel and Scientific Computing
Year:
2012

Citing 5
Cited 1

3D finite difference computation on GPUs using CUDA

Proceedings of 2nd Workshop on General Purpose Processing on Graphics Processing Units
Large scale plane wave pseudopotential density functional theory calculations on GPU clusters

Proceedings of 2011 International Conference for High Performance Computing, Networking, Storage and Analysis
PyCUDA and PyOpenCL: A scripting-based approach to GPU run-time code generation

Parallel Computing
phiGEMM: A CPU-GPU Library for Porting Quantum ESPRESSO on Hybrid Systems

PDP '12 Proceedings of the 2012 20th Euromicro International Conference on Parallel, Distributed and Network-based Processing
Computational physics on graphics processing units

PARA'12 Proceedings of the 11th international conference on Applied Parallel and Scientific Computing

Computational physics on graphics processing units

PARA'12 Proceedings of the 11th international conference on Applied Parallel and Scientific Computing

Quantified Score

Hi-index	0.00

Visualization

Abstract

We present an implementation of parallel GPU-accelerated GPAW, a density-functional theory (DFT) code based on grid based projector-augmented wave method. GPAW is suitable for large scale electronic structure calculations and capable of scaling to thousands of cores. We have accelerated the most computationally intensive components of the program with CUDA. We will provide performance and scaling analysis of our multi-GPU-accelerated code staring from small systems up to systems with thousands of atoms running on GPU clusters. We have achieved up to 15 times speed-ups on large systems.