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.
Hurricane Harvey was one of the most devastating hurricanes to ever hit the United States, and researchers are examining to understand how to prevent similar catastrophes in the future. In this paper, written by a team from the University of Texas at Arlington and the APEC Climate Center, the authors reanalyze the flooding in Houston using dynamic wave modeling and high-resolution terrain data. They highlight the highly accurate results they were able to yield and make recommendations for future research.
Authors: Seong Jin Noh, Jun-Hak Lee, Seungsoo Lee and Dong-Jun Seo.
As HPC systems become increasingly complex, they are exposed to a wide range of performance anomalies, but identifying those anomalies can be difficult or expensive. In this paper – written by a team from the University of Vigo and the Arab Academy for Science Technology and Maritime Transport – the authors compare the most commonly used machine learning techniques for anomaly detection, finding one algorithm to yield the most accurate results.
Authors: Mohamed Soliman Halawa, Rebeca P. Díaz Redondo and Ana Fernández Vilas
This extensive report discusses a project to construct “data building blocks to address major cyberinfrastructure challenges” in high-performance big data computing for use cases such as biomolecular simulation, polar science, pathology, and more. The authors – a team from seven different institutions – outline the building blocks they have built, their contributions so far, and their future trajectories.
Authors: Geoffrey Fox, Judy Qiu, David Crandall, Gregor von Laszewski, Oliver Beckstein, John Paden, Ioannis Paraskevakos, Shantenu Jha, Fusheng Wang, Madhav Marathe, Anil Vullikanti and Thomas Cheatham.
As earth system models near the exascale era, the incremental changes in code refactoring, software environments, and other factors make reproduction of solutions an extremely challenging task. In this paper, a team from Oak Ridge National Laboratory leverages the statistical properties of the global climate system to establish a greater level of model reproducibility. Using this statistical method, they are able to reproduce results from earlier versions of earth system models.
Authors: Salil Mahajan, Katherine J. Evans, Joseph H. Kennedy, Min Xu, Mathew R. Norman and Marcia L. Branstetter.
Thanks to improvements in sequencing technology, the data available to bioinformaticists is growing at a near-exponential rate. In order to leverage that data, researchers are increasingly turning to HPC architectures with high core count processors. In this paper, written by a team from the University of Tennessee and BioTeam Inc., the authors discuss their distributed implementation of the most popular sequence search and alignment tool, showing greater efficiency on modern HPC clusters.
Authors: Shane Sawyer, Mitchel Horton, Chad Burdyshaw, Glenn Brook and Bhanu Rekapalli.
More and more HPC systems are beginning to utilize mobile market processors – but this possible transition, of course, has trade-offs. In this paper, written by two researchers from Barcelona, the authors undertake a study to quantify trade-offs in energy performance when using mobile market processors using a set of detailed simulations. They conclude by outlining the characteristics of a mobile market processor that might minimize those trade-offs.
Authors: Adrià Armejach, Marc Casas and Miquel Moretó.
The EU has launched a number of major initiatives to support HPC development among its constituent countries – most recently, the major EuroHPC program launch. In this report, written by a team of European researchers, the authors discuss the mismatch between the hiring needs of the European HPC, AI and cybersecurity industries and the advanced educational programs offered in the EU. They outline specific geographic and technical areas where more specialized programs are necessary and recommend a path forward.
Authors: M. López Cobo, G. De Prato, G. Alaveras, R. Righi, S. Samoili, J. Hradec, L.W. Ziemba, K. Pogorzelska and M. Cardona.
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.