In November at SC19 in Denver, the HPCwire Readers’ and Editors’ Choice awards program celebrated its 16th year of honoring remarkable achievements in high-performance computing. With categories ranging from Best Use of HPC in Energy to Top HPC-Enabled Scientific Achievement, many of the winners contributed to groundbreaking developments in the sciences. This editorial highlights those awards.
Recognizing HPC’s contributions to our understanding of Earth
A number of awards went to researchers working to better model crucial Earth systems. Pawsey Supercomputing Center helped to build detailed 3D models of Perth’s groundwater aquifers using near-real-time data processing, earning the Editors’ Choice Award for Best Use of HPC in Physical Sciences. “The detailed knowledge of the groundwater system provided by this model assists the government in sustaining this valuable water resource,” said Mark Stickells, Pawsey’s executive director. “It is a great example of Pawsey-enabled collaborative research that has a direct impact on our community and quality of life.” The research is already being used to assist the Australian government in assessing groundwater sites.
The Ocean Conservancy also worked on modeling the world’s water, performing over 75 50-year ocean simulations on AWS to understand climate change’s effects on underwater ecosystems (and winning the Readers’ Choice Award for Best Use of HPC in the Cloud). “The computational density of this would never have been possible on premises, even using the supercomputers available at the world’s largest universities,” wrote Amazon CTO Werner Vogels on his blog.
The Leibniz Supercomputing Centre (LRZ), meanwhile, focused on even deeper systems, constructing 3D models of past shifts in tectonic plates and the Earth’s mantle, improving the safety of fossil fuel extraction processes and earning LRZ the Editors’ Choice Award for Best Use of HPC in Energy. The research, which used the SuperMUC-NG supercomputer, will help energy companies minimize their environmental impacts. “We are happy to be supporting cutting-edge research such as Prof. Bunge’s,” said Prof. Dieter Kranzlmüller, director of LRZ. “We are honored that this collaboration resulted in an HPCwire Editors’ Choice award, and are always happy when the HPC community hears about some of the innovations happening at LRZ specifically and the three GCS centres generally.”
Finally, the National Center for Supercomputing Applications (NCSA), the National Geospatial-Intelligence Agency, the University of Minnesota and Ohio State University won the Editors’ Choice Award for Best HPC Collaboration for their ongoing work to produce an unprecedented, high-resolution elevation model of the entire planet. “NCSA is honored to be awarded for our contributions in geospatial research,” said William “Bill” Gropp, director of NCSA. “Congratulations to our research partners, and thank you to HPCwire for this recognition.” The collaboration is planned as a “long-term strategic relationship” that will span “many years.”
Recognizing HPC’s contributions to understanding health and the human body
Moving from the macro to the micro, many of this year’s awards celebrated advances in human biology and disease treatments. Stony Brook University researchers used the Pittsburgh Supercomputing Computer’s Bridges and Bridges-AI systems to examine tumor tissue biology and build predictive, individualized treatment models, winning the Readers’ Choice Award for Best Use of HPC in Life Sciences – PSC’s tenth year with a win in the awards. “Having PSC’s contributions being recognized for ten years by the HPCwire awards, granted by the top trade publication in HPC, is a clear and strong sign that reaffirms PSC’s unwavering commitment to the advancement of science and engineering through HPC,” said Paola Buitrago, PSC’s director of artificial intelligence and big data.
CEA also tackled tumors, earning an Editors’ Choice Award for Best Use of HPC in Life Sciences by leveraging a PRACE allocation on GENCI’s Curie machine to implement a new algorithm that could lead to better breast cancer tumor detection. In essence, the researchers are using massively parallel codes to explore conversion of laser light into various forms, which they hope will lead to a breakthrough in high-resolution medical imaging and cancer detection. “PRACE is honored to receive the HPCwire Readers’ and Editors’ Choice Award for Best Use of HPC in Life Sciences,” said Serge Bogaerts, PRACE managing director. “This award acknowledges the endeavour of PRACE as a research infrastructure to support excellence in science.”
Research on other diseases was also recognized; another Editors’ Choice Award for Best Use of HPC in Life Sciences was awarded to a team from the University of Michigan, the San Diego Supercomputer Center (SDSC) and the Texas Advanced Computing Center (TACC) for their work using a model to examine differences in visualizations to better understand tuberculosis – specifically, how tuberculosis granulomas form and spread. “We’re delighted to be recognized once again,” said SDSC director Michael Norman, calling the life sciences “a key element of [SDSC’s] strategic thrust.”
Researchers at Mount Sinai Hospital, meanwhile, collaborated with Minerva HPC and won the Editors’ Choice Award for Best Use of High Performance Data Analytics & Artificial Intelligence for their development of an AI platform to detect neurodegenerative diseases using machine learning. The platform, called the Precise Informatics Platform, analyzed digitized microscopic slides of tissue samples to create a convolutional neural network capable of identifying “tangles” that serve as markers for neurodegenerative diseases. “Thank you @hpcwire for the #hpcwire Editors’ Choice Award,” Minerva HPC tweeted. “We couldn’t be more proud of our researchers and partners @Lenovo who have worked so hard.”
Finally, Tanmay Bakshi, a 15-year-old Canadian systems architect, used Nimbix Cloud to develop a neural network to create a new kind of biometric identification: Heart ID, which identifies an individual based on the unique electrical activity of their heart. Nimbix highlights how Heart ID requires no expensive hardware and has a very small software footprint, allowing it to run on many lightweight devices, including existing smartwatches with ECG sensors. Bakshi won the Readers’ Choice Award for Best Use of High Performance Data Analytics & Artificial Intelligence. “Thank you to the @hpcwire community,” Nimbix tweeted, “and congratulations to [Tanmay Bakshi] for this incredible accomplishment!”
Recognizing HPC’s contributions to our grasp of the universe
Yet other researchers set their sights on the heavens, leveraging immense supercomputing power to enhance our understanding of the cosmos. Scientists at Argonne National Laboratory used Oak Ridge National Laboratory’s Summit supercomputer to conduct three of the largest cosmological simulations in history to study the makeup of the universe. In the process, they broke records when the massive simulations necessitated a massive 2.9 PB data transfer, the largest single data transfer in history – winning them the Readers’ Choice Award for Best Use of HPC in Physical Sciences.
SDSC, TACC and the Pittsburgh Supercomputer Center (PSC) conducted supercomputer simulations under the National Science Foundation (NSF) XSEDE program to uncover new insights into the behavior of gravitational waves. They did this by analyzing the structures of neutron stars in “eccentric” configurations, such as binary neutron stars. The research, which was conducted in collaboration with the Perimeter Institute for Theoretical Physics and the Theoretical Physics Program of the University of Arizona, won the Readers’ Choice Award for Top HPC-Enabled Scientific Achievement. “We’re honored to receive these awards from the community,” said Dan Stanzione, TACC’s executive director.
Last but not least, the Editors’ Choice Award for Top HPC-Enabled Scientific Achievement went to the Event Horizon Telescope team, which crunched petabytes of data to produce groundbreaking cosmological images, including the first-ever image of a black hole, which went viral in April 2019. To find out more about the creation of the black hole image, which required crunching four petabytes of data, read HPCwire’s story here.
The HPCwire Readers’ and Editors’ Choice Awards are a way for our community to recognize the best and brightest innovators within the global HPC community. To see the full list of the 2019 winners, click here.