This title appears in the Scientific Report :
2019
Please use the identifier:
http://hdl.handle.net/2128/23603 in citations.
Analytical PAW Projector Functions for Reduced Bandwidth Requirements
Analytical PAW Projector Functions for Reduced Bandwidth Requirements
Large scale electronic structure calculations require modern high performance computing (HPC) resources and, as important, mature HPC applications that can make efficient use of those. Real-space grid-based applications of Density Functional Theory (DFT) using the Projector Augmented Wave method (PA...
Saved in:
Personal Name(s): | Baumeister, Paul F. (Corresponding author) |
---|---|
Tsukamoto, Shigeru | |
Contributing Institute: |
JARA - HPC; JARA-HPC JARA-FIT; JARA-FIT Quanten-Theorie der Materialien; PGI-1 Quanten-Theorie der Materialien; IAS-1 Jülich Supercomputing Center; JSC |
Imprint: |
2019
|
Conference: | Platform for Advanced Scientific Computing Conference, Zurich (Switzerland), 2019-06-12 - 2019-06-14 |
Document Type: |
Conference Presentation |
Research Program: |
Controlling Configuration-Based Phenomena Controlling Spin-Based Phenomena Computational Science and Mathematical Methods |
Link: |
OpenAccess |
Publikationsportal JuSER |
Large scale electronic structure calculations require modern high performance computing (HPC) resources and, as important, mature HPC applications that can make efficient use of those. Real-space grid-based applications of Density Functional Theory (DFT) using the Projector Augmented Wave method (PAW) can give the same accuracy as DFT codes relying on a plane wave basis set but exhibit an improved scalability on distributed memory machines. The projection operations of the PAW Hamiltonian are known to be the performance critical part due to their limitation by the available memory bandwidth. We investigate on the utility of a 3D factorizable basis of Hermite functions for the localized PAW projector functions which allows to reduce the bandwidth requirements for the grid representation of the projector functions in projection operations. Additional on-the-fly sampling of the 1D basis functions eliminates the memory transfer almost entirely. For an quantitative assessment of the expected memory bandwidth savings we show performance results of a first implementation on GPUs. Finally, we suggest a PAW generation scheme adjusted to the analytically given projector functions. |