Parallel Performance and Scalability Experiments with the Danish Eulerian Model on the EPCC Supercomputers

Using single precision floating point representation reduces the size of data and computation time by a factor of two relative to double precision conventionally used in electronic structure programs. For large-scale calculations, such as those encountered in many-body theories, reduced memory footprint alleviates memory and input/output bottlenecks. Reduced size of data can lead to additional gains due to improved parallel performance on CPUs and various accelerators. However, using single precision can potentially reduce the accuracy of computed observables. Here we report an implementation of coupled-cluster and equation-of-motion coupled-cluster methods with single and double excitations in single precision. We consider both standard implementation and one using Cholesky decomposition or resolution-of-the-identity of electron-repulsion integrals. Numerical tests illustrate that when single precision is used in correlated calculations, the loss of accuracy is insignificant and pure single-precision implementation can be used for computing energies, analytic gradients, excited states, and molecular properties. In addition to pure single-precision calculations, our implementation allows one to follow a single-precision calculation by clean-up iterations, fully recovering double-precision results while retaining significant savings.

Download Full-text

An eulerian model for soil dynamics: application to fast slope movements (III) Rheological models

Revue française de génie civil ◽

10.3166/rfgc.7.1037-1051 ◽

2003 ◽

Vol 7 (7-8) ◽

pp. 1037-1051

Author(s):

Manuel Pastor ◽

Manuel Quecedo ◽

Elena Gonzales ◽

Maria Isabel Herreros ◽

José Antonio Fernandez Merodo ◽

...

Keyword(s):

Soil Dynamics ◽

Eulerian Model ◽

Rheological Models ◽

Slope Movements

Download Full-text

A Model of Parallel Performance

10.21236/ada207792 ◽

1989 ◽

Cited By ~ 1

Author(s):

Edward A. Carmona ◽

Michael D. Rice

Keyword(s):

Parallel Performance

Download Full-text

Estimating parallel performance, a skeleton-based approach

Proceedings of the fourth international workshop on High-level parallel programming and applications - HLPP '10 ◽

10.1145/1863482.1863489 ◽

2010 ◽

Cited By ~ 3

Author(s):

Oleg Lobachev ◽

Rita Loogen

Keyword(s):

Parallel Performance

Download Full-text

Modeling of Drag Finishing—Influence of Abrasive Media Shape

Journal of Manufacturing and Materials Processing ◽

10.3390/jmmp5020041 ◽

2021 ◽

Vol 5 (2) ◽

pp. 41

Author(s):

Irati Malkorra ◽

Hanène Souli ◽

Ferdinando Salvatore ◽

Pedro Arrazola ◽

Joel Rech ◽

...

Keyword(s):

Surface Roughness ◽

Material Removal ◽

Direct Shear Test ◽

Shear Test ◽

Shear Resistance ◽

Arbitrary Lagrangian Eulerian ◽

Eulerian Model ◽

Drag Forces ◽

Drag Finishing ◽

Surrounding Liquid

Drag finishing is a widely used superfinishing technique in the industry to polish parts under the action of abrasive media combined with an active surrounding liquid. However, the understanding of this process is not complete. It is known that pyramidal abrasive media are more prone to rapidly improving the surface roughness compared to spherical ones. Thus, this paper aims to model how the shape of abrasive media (spherical vs. pyramidal) influences the material removal mechanisms at the interface. An Arbitrary Lagrangian–Eulerian model of drag finishing is proposed with the purpose of estimating the mechanical loadings (normal stress, shear stress) induced by both abrasive media at the interface. The rheological behavior of both abrasive slurries (media and liquid) has been characterized by means of a Casagrande direct shear test. In parallel, experimental drag finishing tests were carried out with both media to quantify the drag forces. The correlation between the numerical and experimental drag forces highlights that the abrasive media with a pyramidal shape exhibits a higher shear resistance, and this is responsible for inducing higher mechanical loadings on the surfaces and, through this, for a faster decrease of the surface roughness.

Download Full-text

Parallel finite-element method using domain decomposition and Parareal for transient motor starting analysis

COMPEL The International Journal for Computation and Mathematics in Electrical and Electronic Engineering ◽

10.1108/compel-12-2018-0516 ◽

2019 ◽

Vol 38 (5) ◽

pp. 1507-1520 ◽

Cited By ~ 1

Author(s):

Yasuhito Takahashi ◽

Koji Fujiwara ◽

Takeshi Iwashita ◽

Hiroshi Nakashima

Keyword(s):

Finite Element Method ◽

Finite Element ◽

Parallel Computing ◽

Domain Decomposition ◽

Parallel Computation ◽

Domain Decomposition Method ◽

Space Time ◽

Content Type ◽

Parallel Performance ◽

Element Method

Purpose This paper aims to propose a parallel-in-space-time finite-element method (FEM) for transient motor starting analyses. Although the domain decomposition method (DDM) is suitable for solving large-scale problems and the parallel-in-time (PinT) integration method such as Parareal and time domain parallel FEM (TDPFEM) is effective for problems with a large number of time steps, their parallel performances get saturated as the number of processes increases. To overcome the difficulty, the hybrid approach in which both the DDM and PinT integration methods are used is investigated in a highly parallel computing environment. Design/methodology/approach First, the parallel performances of the DDM, Parareal and TDPFEM were compared because the scalability of these methods in highly parallel computation has not been deeply discussed. Then, the combination of the DDM and Parareal was investigated as a parallel-in-space-time FEM. The effectiveness of the developed method was demonstrated in transient starting analyses of induction motors. Findings The combination of Parareal with the DDM can improve the parallel performance in the case where the parallel performance of the DDM, TDPFEM or Parareal is saturated in highly parallel computation. In the case where the number of unknowns is large and the number of available processes is limited, the use of DDM is the most effective from the standpoint of computational cost. Originality/value This paper newly develops the parallel-in-space-time FEM and demonstrates its effectiveness in nonlinear magnetoquasistatic field analyses of electric machines. This finding is significantly important because a new direction of parallel computing techniques and great potential for its further development are clarified.

Download Full-text