In this bimonthly feature, HPCwire highlights newly published research in the high-performance computing community and related domains. From parallel programming to exascale to quantum computing, the details are here.
Honing astrophysical codes on the road to exascale
As larger systems become increasingly complex, code developers are struggling to keep up. These authors, a team of astrophysicists and computer scientists, explore the performance of astrophysics codes on the Intel Knights Landing architecture, which they use as a proxy for future exascale architectures due to its high core count per node and high memory bandwidth. The authors work to optimize several codes on Knights Landing, conveying lessons learned for optimizing codes on complex HPC architectures.
Authors: Salvatore Cielo, Luigi Iapichino, Fabio Baruffa, Matteo Bugli and Christoph Federrath.
Running no-compromise HPC simulations of new fuel cells
Proton-exchange membrane fuel cells are a new form of fuel cell that stand to change how technologies ranging from portable electronics to the space shuttle are powered, and simulations are used throughout the design process to iterate and improve. This paper, written by a team from the French Alternative Energies and Atomic Energy Commission, discusses efforts to perform these complex simulations with “no compromise in terms of geometry and physics” using HPC systems.
Authors: Mathias Gerard, Pascal Schott, Benoit Mathieu, Van-Quang Dinh, Stephane Veys, Adrien Brunton and Pierre Ledac.
Assessing representation of women at HPC conferences
“Women are severely underrepresented in the HPC workforce,” say these researchers, a duo from Reed College, who aim to “provide tangible and reproducible data on the gender gap.” Assessing data from nine HPC-related conferences that took place in 2017, they demonstrate that women comprise only 10 percent of all authors at the conferences, with variations depending on the authors’ countries and experience levels.
Authors: Eitan Frachtenberg and Rhody Kaner
Seeking the best Weather Research and Forecasting model performance
The authors of this paper, from two branches of the University of Castilla-La Mancha in Spain, worked to predict snowfall during the 2018 Winter Olympics using the popular Weather Research and Forecasting model. In order to maximize accuracy for such a concentrated event, the authors tested a variety of compilers, MPI libraries and platforms, cataloguing “substantial performance differences” and developing what they call an “empirical score of special interest to supercomputer maintainers, developers and scientists, which can be useful to obtain the best WRF configuration for their systems.”
Authors: R. Moreno, E. Arias, D. Cazorla, J. J. Pardo and F. J. Tapiador.
Simulating tens of thousands of atoms using millions of cores
The Kohn-Sham density functional theory calculations are important tools in quantum and particle physics. These authors – a team from the University of California Berkeley and three Chinese institutions (including the National Supercomputing Center) – describe how they implemented a density functional theory method on the Sunway TaihuLight supercomputer, allowing them to scale up to over 8.5 million cores to investigate the structures of graphene systems containing tens of thousands of carbon atoms.
Authors: Wei Hu, Xinming Qin, Qingcai Jiang, Junshi Chen, Hong An, Weile Jia, Fang Li, Xin Liu, Dexun Chen and Jinlong Yang.
Describing the Indian experience of HPC education
HPC training and education are necessary to prepare workers in a growing sector, but are sometimes difficult to implement due to limited hands-on access to real systems. These authors from the Centre for Development of Advanced Computing and Savitribai Phule Pune University describe the Indian experience with HPC education, proposing an education strategy to close the gap between the status quo and sufficient training levels.
Authors: V. Venkatesh Shenoi, Vaishali Shah and Sandeep K. Joshi.
Preparing for exascale computing at the Jülich Supercomputing Centre
In this paper, written by a team from Jülich Supercomputing Centre (JSC) in Germany, the authors describe how JSC is preparing for the arrival of exascale in Europe. The paper outlines the efforts (which the authors say revolve around modular supercomputing), including “the deployment and operation of large-scale computing platforms,” the federation of storage infrastructures, and education and training programs.
Authors: Estela Suarez, Wolfgang Frings, Sebastian Achilles, Norbert Attig, Jacopo de Amicis, Edoardo Di Napoli, Thomas Eickermann, Eric B. Gregory, Bjorn Hagemeier, Andreas Herten, et al.
Do you know about research that should be included in next month’s list? If so, send us an email at [email protected]. We look forward to hearing from you.