Preparing for science in the exascale era
As the future home to Aurora exascale system, Argonne National Laboratory has been ramping up efforts to ready the supercomputer and its future users for science in the exascale era.
It's national Exascale Day 2020! Visit the Exascale Computing Project's website to learn how the U.S Department of Energy's upcoming exascale supercomputers will change science as we know it.
Happy Exascale Awareness Week! Follow along this week (October 12-18, 2020) to learn how DOE's upcoming exascale supercomputers, including Argonne's Aurora system, will accelerate scientific discovery and innovation. Visit the Exascale Computing Project's event webpage for stories, videos, and more.
Congratulations to Hal Finkel on his new position with DOE! He has been named a computer science program manager for the DOE Office of Advanced Scientific Computing Research. From his work with compiler technologies and programming models to his efforts to help prepare the facility for exascale, Hal made a significant impact on scientific computing in his time with ALCF. He will certainly be missed but we look forward to continued collaborations in his new role with ASCR.
Over the past four decades, the U.S. Department of Energy's Office of Advanced Scientific Computing Research (ASCR) has played a pivotal role in advancing computational science through investments in basic research, leadership-class facilities and computers, workforce development programs, and much more. Check out the ASCR@40 report to learn more.
For 30 years, SCinet has provided SC conference attendees with the innovative network platform necessary to connect, transport, and display HPC research at SC from around the world. Check out this Q&A with Argonne's Linda Winker, a longtime SCinet volunteer, to learn more about the high-performance network and its impact on the HPC community.
Rick Stevens, associate laboratory director at Argonne National Laboratory, delivered a talk at ISC High Performance 2020 on how high-performance computing, machine learning, and AI are being used to advance the fight against COVID-19. Check out the new article from Science Node to learn more.
A new article from Proceedings of the National Academy of Sciences (PNAS) details some of the challenges associated with exascale computing, and how the upcoming supercomputers, like Argonne's Aurora system, stand to enable critical breakthroughs in medicine, climate research and more.
In a new interview with insideHPC, David Martin, manager of Industry Partnerships and Outreach at the ALCF, discusses upcoming exascale supercomputers and the importance of making them accessible to a wide variety of applications, environments, and people.
Researchers from the ALCF, NERSC and Cray-now-HPE teamed up to develop the Global Performance and Congestion Network Test (GPCNeT), a new benchmark suite for measuring contention and congestion in HPC networks. Check out this new article from the Next Platform to learn more.
Applications are now open for the 2021 BSSw Fellowship Program, which gives recognition and funding to leaders and advocates of high-quality scientific software. The deadline to apply is September 30, 202
Congratulations to ALCF computer scientist Murali Emani for receiving a new DOE award aimed at making artificial intelligence (AI) models and data more accessible and reusable to accelerate AI research and development. Emani will co-lead “HPC-FAIR: A Framework Managing Data and AI Models for Analyzing and Optimizing Scientific Applications,” one of five new projects that will apply Findable, Accessible, Interoperable, and Reusable (FAIR) Data Principles so that science data can drive innovations in AI. Emani’s three-year project received $2.2 million in funding.
In the Virtual Institute for I/O's new IO-500 list on filesystem performance, three of the top four systems are DAOS platforms, including Argonne National Laboratory's Presque testbed. It’s great to see the success of the DAOS work supported by Argonne's efforts to prepare for Aurora.
“The recent IO-500 results for DAOS demonstrate the continuing maturity of the software’s functionality enabled by a well-managed code development and testing process. The collaborative development program will continue to deliver additional capabilities for DAOS in support of Argonne’s upcoming exascale system, Aurora,” said Gordon McPheeters, HPC systems administration specialist at the ALCF.
The first all-virtual Argonne Training Program on Extreme-Scale Computing (ATPESC) is wrapping up week 1 with a series of talks on data-intensive computing and I/O. For anyone interested in following along from home, slide decks are being posted on the ATPESC website.
On Thursday, July 23 at 11 am (CT), Salman Habib, director of Argonne's Computational Science division, will present "Next-Generation High-Performance Computing" as part of the AI + Science seminar series hosted by the AI Joint Task Force of the University of Chicago, Argonne National Laboratory, Fermi National Accelerator Laboratory and Toyota Technological Institute at Chicago.
This year's P3HPC Forum will be held Sept. 1-2, 2020, as a virtual event. Register by July 31 to take part in the annual meeting dedicated to sharing best practices and ideas for achieving performance portability across current and future supercomputers.
Using the ALCF's Theta supercomputer, researchers from Princeton University are performing 3D supernova simulations to shed light on the physics behind the collapse of massive stars. In collaboration with the ALCF visualization team, the researchers have produced a new visualization of the early phase of a supernova explosion.
Valerie Taylor, director of Argonne National Laboratory's Mathematics and Computer Science Division, recently joined the This Week in HPC podcast for an important panel discussion on racism and high-performance computing.
The newly released SYCL 2020 Provisional Specification includes new features that will help researchers prepare applications to run on our upcoming exascale system, Aurora.
“Our users will benefit from features in the provisional SYCL 2020 specification,” said Hal Finkel, lead for compiler technology and programming languages at the ALCF. “New features, such as support for unified memory and reductions, are important capabilities for programming high-performance-computing hardware. In addition, support for C++17 will allow our users to write better C++ code, with both language features (such as deduction guides) and library features (such as std::optional).”
The First Workshop on LLVM Compiler and Tools for HPC is now underway at the ISC High Performance Digital 2020 conference. Johannes Doerfert (ALCF) is a co-organizer of the workshop and Hal Finkel (ALCF) is delivering the keynote talk, "LLVM and the Future of Computing: Adaptive, Reliable, and Heterogeneous."
Watch ALCF postdoctoral researcher Zheming Jin deliver a technical talk on using SYCL to improve the performance of medical imaging applications as part of the Intel HPC + AI pavilion at ISC 2020.
As part of the Intel HPC + AI pavilion at ISC 2020, the ALCF's Hal Finkel joins Codeplay's Andrew Richards and Intel's Sanjiv Shah to discuss the oneAPI initiative.
ISC High Performance Digital 2020 is officially underway. Watch Argonne National Laboratory's Rick Stevens deliver an overview of how supercomputing and AI are being used to advance COVID-19 research.
If you're planning to tune in for ISC High Performance Digital 2020 (June 22-25), be sure to check out Argonne National Laboratory's contributions to the digital conference, including talks on oneAPI, SYCL, LLVM, and using HPC and AI for COVID-19 research. Registration is free and all talks are available to registered participants for 14 days.
In a new report from WTTW - Chicago PBS, Argonne National Laboratory's Charles Macal discusses how scientists are using ALCF supercomputing resources to model the spread of COVID-19.
"In layman’s terms, Macal says researchers have essentially created a SimCity on steroids using Argonne’s Theta supercomputer to create forecasts of the spread of the virus for up to a simulated year."
Check out this new article from NBC Chicago to learn how Argonne National Laboratory scientists and tools, including the ALCF's Theta supercomputer, are advancing COVID-19 research.
Intel is hosting a few upcoming oneAPI webinars that may be of interest to researchers preparing for the ALCF's upcoming exascale system, Aurora.
- June 3: Offload Your Code from CPU to GPU … and Optimize It
- June 10: Profile DPC++ and GPU Workload Performance
The 236th Meeting of the American Astronomical Society kicks off this week. ALCF Deputy Director Jini Ramprakash will take part in the June 3 town hall session on National High Performance Computing User Facilities, informing attendees about the supercomputing resources that will be available in the next decade to support astronomy, cosmology, and astrophysics research.
Check out this new article from ZDNet to learn how researchers from Argonne National Laboratory are using supercomputers and AI to accelerate the search for potential drugs to combat COVID-19.
"We are doing in a few months what would normally take a drug development process years to do," said Argonne's Rick Stevens.
As home to one of NVIDIA's first DGX A100 systems, Argonne National Laboratory will use the cluster’s AI and computing power to better understand and fight COVID-19.
“The compute power of the new DGX A100 systems coming to Argonne will help researchers explore treatments and vaccines and study the spread of the virus, enabling scientists to do years’ worth of AI-accelerated work in months or days," said Rick Stevens, Argonne's associate laboratory director for Computing, Environment and Life Sciences.
With help from DOE supercomputers, including the ALCF's Theta system, a team led by Christine Goulet of the Southern California Earthquake Center is modeling the San Andreas Fault system to better understand the origin and impact of earthquakes. The team is using Theta “mostly for dynamic earthquake ruptures,” Goulet says. “That is using physics-based models to simulate and understand details of the earthquake as it ruptures along a fault, including how the rupture speed and the stress along the fault plane changes.” To learn more, check out the new story from ASCR Discovery.
ALCF researchers delivered a few noteworthy talks at last week’s International Workshop on OpenCL/SYCLcon 2020. Hal Finkel presented the keynote talk "Preparing to program Aurora at Exascale: Early experiences and future directions.” Brian Homerding and John Tramm presented "Evaluating the Performance of the hipSYCL toolchain for HPC kernels on NVIDIA V100 GPUs.” Slides and videos are available on the conference website.
Day 1 of our first virtual ALCF Computational Performance Workshop is in the books! We had nearly 100 participants join us for talks and breakouts designed to help boost code performance on ALCF supercomputing resources. Looking forward to two more days of collaborative online sessions.
In this new episode of the Code Together podcast, the ALCF’s Hal Finkel and Intel’s Alice Chan discuss the shift to modern C++ programming models, and how the cross-industry oneAPI initiative is addressing the need for programming portability and performance across diverse architectures.
The COVID-19 High Performance Computing (HPC) Consortium is helping to fast-track COVID-19 vaccine, drug, and epidemiology calculations by awarding researchers with free computing time on machines ranging from small clusters to some of the world’s largest supercomputers, including the ALCF's Theta system. “It’s sort of like a super user facility,” says Katherine Riley, ALCF Director of Science.
Congrats to ALCF Director Michael E. Papka for being named a 2020 Presidential Research, Scholarship and Artistry Professor by Northern Illinois University. In addition to his leadership roles at Argonne National Laboratory, Papka serves as a professor of computer science at NIU and co-director of its Data, Devices and Interaction Laboratory (ddiLab).
Check out this new article from the Texas Advanced Computing Center (TACC) to learn how researchers from Argonne and Brookhaven National Laboratories are using supercomputing power and AI to speed the identification of potential drugs to fight COVID-19.
In a new profile from the U.S. Department of Energy, ALCF Director Michael E. Papka discusses his role at Argonne National Laboratory, the impact of supercomputers on science, and the future of leadership computing.
“DOE Leadership Computing Facilities exist to serve the research community. As stewards of these resources, we must provide an environment that not only meets the needs of the user but the future they have not yet imagined.”
This new article from the University of Chicago details how researchers from the university and Argonne are studying antibodies to advance the development of COVID-19 treatments, vaccines, and testing regimens.
“Computational experiments using massively parallel supercomputers, such as Argonne’s Theta supercomputer and others, can suggest the best candidates for drugs, and then we’ll test to see whether any of these drugs inhibit central viral enzymes."
The first four chapters of Intel's forthcoming book, "Data Parallel C++: Mastering DPC++ for Programming of Heterogeneous Systems Using C++ and SYCL,” have been released in advance for free download. DPC++, an extension of C++ that incorporates SYCL and other new features, will be available on our Aurora exascale system.
A new article from Brookhaven National Laboratory highlights a collaborative effort with Stony Brook University and Argonne National Laboratory to develop computational models that will advance our understanding of the novel coronavirus responsible for COVID-19.
Check out this new article from Chicago Magazine to learn more about Argonne National Laboratory's efforts to combat the coronavirus pandemic.
"Using the laboratory’s supercomputers, researchers simulate the virus’s progression over time, dial up potential interventions, and predict the impact of those interventions on the infection rate."
Argonne National Laboratory's Stephen Streiffer joined CBS Chicago to discuss the lab's COVID-19-related research, including the use of supercomputers to study the virus and its properties. The goal, Streiffer said, is to “actually model parts of the virus so that we can better determine drugs that might be able to interfere with the reproduction of the virus, or ultimately give us clues as to how we might develop a vaccine for the virus.”
The ALCF is proud to be a part of the COVID-19 High Performance Computing Consortium, a private-public effort that brings together the federal government, industry, and academia to provide access to the world’s most powerful supercomputing resources in support of COVID-19 research. For more information, or to submit a proposal for COVID-19-related research, visit the consortium webpage.
The U.S. Department of Energy announced a plan to provide $60 million to establish multidisciplinary teams to develop new tools and techniques to harness supercomputers for scientific discovery. Letters of intent are required and are due on April 14, 2020 by 5 p.m. (ET).
ALCF researchers Pankaj Rajak and Ye Luo were part of a multi-institutional team that was recently recognized with the Best Paper Award at the HPC Asia 2020 conference. The team’s paper, "Quantum Dynamics at Scale: Ultrafast Control of Emergent Functional Materials,” describes their efforts to prepare a quantum molecular dynamics engine to run on Argonne National Laboratory's upcoming exascale system, Aurora. (Pictured here: Subodh Tiwari, a co-author from the University of Southern California, accepts the award on the team’s behalf.)
A new book, “Introduction to the Fast Multipole Method," co-written by ALCF computational scientist Victor Anisimov provides a user-friendly exploration of the theory and computer implementation of FMM, an algorithm that enables simulations in computational biophysics, chemistry, and materials science to deal with larger and more realistic models. Developed in 1987, FMM was recognized by Computing in Science and Engineering as one of the Top 10 Algorithms of the 20th century.
Anisimov is currently part of an effort to implement FMM into the NWChemEx molecular modeling package to improve its performance on Aurora and DOE's other upcoming exascale systems.
In the latest episode of the Let's Talk Exascale Podcast, the ALCF's Hal Finkel discusses the collaborative effort to develop the open-source LLVM compiler infrastructure for DOE's upcoming exascale systems.
“LLVM has really become a critical tool in our toolbox of technologies that form the basis of the programming environments on high-performance computing systems.”
Researchers from Argonne National Laboratory and our fellow DOE labs are contributing to the MLPerf-HPC working group, a broad collaborative effort working to develop a machine learning benchmark suite for supercomputers. Check out this InsideHPC feature from ALCF computer scientist Murali Emani to learn more about the group and how you can contribute.
Researchers from the ALCF's Aurora Early Science Program and DOE's Exascale Computing Project are visiting Argonne National Laboratory this week for Aurora ESP Workshop: TWO. The three-day, hands-on training event is designed to help research teams advance their applications and software development efforts for the ALCF's upcoming exascale system, Aurora.
Argonne's annual Introduce a Girl to Engineering Day (IGED) is one of our favorite days of the year! Serving as mentors and volunteers, several ALCF staff members are helping to inspire the next generation of engineers and scientists.
With help from ALCF supercomputing resources, researchers from University of Washington are breaking through computational barriers to advance the design of synthetic proteins for potential applications in medicine and technology.
Attend next week's IXPUG webinar (February 20) for an opportunity to learn how the Intel oneAPI programming model can help advance code development efforts for Aurora and other upcoming exascale systems. Led by Intel's Kevin O'Leary, the webinar will cover oneAPI’s performance optimization features and demonstrate how oneAPI can be used to port an application from CPU to GPU.
With an ALCF Data Science Program (ADSP) award, researchers from University of Cambridge and Argonne National Laboratory used data mining tools and ALCF supercomputing resources to create an auto-generated database of materials’ optical properties. Their novel approach, detailed in a recent paper published in Scientific Data, aims to accelerate the design and discovery of materials for solar cells and other optoelectronic applications.
Join us on Wednesday, February 19, for the next installment of the ALCF Developer Sessions webinar series. Rick Wagner, Globus professional services manager, will demonstrate how the Globus platform can be used to automate research data workflows.
Video: ALCF computer scientist Huihuo Zheng presents "Data Parallel Deep Learning" at the 2019 Argonne Training Program on Extreme-Scale Computing (ATPESC).
On Feb. 19, Christian Trott of Sandia National Laboratories will provide an introduction to the Kokkos C++ Performance Portability Ecosystem as part of the "Best Practices for HPC Software Developers" webinar series presented by IDEAS, the Exascale Computing Project, ALCF, OLCF, and NERSC.
Argonne National Laboratory is celebrating Thank Your Mentor Day today. ALCF staff members Christopher Knight and Yasaman Ghadar are a shining example of the positive impact that mentoring relationships bring across the lab.
"What I value most in our mentor/mentee relationship is the fact that Christopher has always found the time to listen and provide candid advice helping me to be confident in the decisions I make to help advance my career." - Yasaman Ghadar
A new article from ASCR Discovery highlights how a research team led by Parviz Moin of Stanford University is using ALCF computing resources to determine whether large-eddy simulation of commercial aircraft is both cost effective and sufficiently accurate to help engineers design more efficient aircraft. Their early results are promising.
"The future of aviation is bright and needs more development,” Moin says. “I think with time – and hopefully it won’t take too long – aerospace engineers will start to see the advantage of these high-fidelity computations in engineering analysis and design.”
With an Aurora Early Science Program project, a research team led by Noa Marom of Carnegie Mellon University is preparing to use the ALCF's forthcoming exascale system to identify new materials that can increase the efficiency of solar cells.
"The materials we are researching have unique properties that make them suitable for use in solar cells, and these properties are very rare and difficult to find out of the wide array of possible materials," Marom said. "We are trying to accelerate the process of material discovery through computer simulation on high-performance computers using sophisticated quantum-mechanical simulation software and machine learning tools.”
As part of the Fermilab/Argonne/UChicago Computing Seminars Series, Rebecca Willett, Professor of Statistics and Computer Science at the University of Chicago, visited Argonne National Laboratory today to deliver a talk on leveraging physical models in machine learning.
Happy Computer Science Education Week! ALCF Deputy Director Jini Ramprakash is one of many Argonne National Laboratory computing experts visiting area schools this week to lead "Hour of Code" activities designed to inspire students to explore the world of computer science. In a visit to Butler Junior High in Oak Brook, Illinois, Ramprakash introduced the students to the lab’s supercomputers and led the class through an activity that involved using data science to solve a mystery.
At the SC19 Conference last month, Argonne National Laboratory's Misha Salim was recognized with the Best Presentation Award at the XLOOP workshop for "Balsam: Near Real-Time Experimental Data Analysis on Supercomputers."
Ti Leggett, ALCF Deputy Director of Operations, recently visited Southbury Elementary in Oswego, Illinois, for American Education Week to highlight how the lab is using supercomputing and data science to solve complex problems in science and engineering. As part of his visit, Leggett led the class through an interactive whodunit activity that involved using data to identify the source of a disease outbreak in 1845 London.
As part of the ALCF's Aurora Early Science Program, Nicola Ferrier of Argonne National Laboratory is preparing to leverage exascale computing power to advance neuroscience research.
With a new project funded by the U.S. Department of Energy’s High Performance Computing for Manufacturing (HPC4Mfg) Program, researchers from Argonne National Laboratory, Purdue Northwest, and ArcelorMittal will leverage Argonne supercomputers to investigate ways that steel manufacturers can improve slab quality, increase productivity, and reduce energy consumption.
Congrats to the Argonne National Laboratory team for their award-winning entry in the SC Conference's first annual SCinet Technology Challenge! In collaboration with Northwestern University, StarLight, Northern Illinois University, the University of Chicago, and the Metropolitan Research and Education Network (MREN), the team's demo, "Real-Time Analysis of Streaming Synchrotron Data," took home the top recognition for an exemplary blend of networking, computing, and storage.
The researchers streamed experimental data from Argonne's Advanced Photon Source to the ALCF for real-time image reconstruction. They then used machine learning to enhance and de-noise the image data and streamed it back to the StarLight booth at SC19 for volume visualization.
Intel unveiled its oneAPI initiative – a unified and scalable programming model for application development across diverse computing architectures in the era of HPC/AI convergence – at the Intel DevCon yesterday at SC19. This tool will help researchers accelerate their efforts to prepare for Argonne National Laboratory’s upcoming Aurora exascale system.
SC19 is right around the corner! At the SC Conference next week in Denver, more than 90 researchers from Argonne National Laboratory will share insights and advances on topics ranging from exascale computing and big data analysis to artificial intelligence and quantum computing.
ALCF Director of Science Katherine Riley appeared on the latest episode of Big Picture Science to talk exascale and the far-reaching impacts of high-performance computing.
Solving the enigma of dark matter could shed light on the origins and structure of our universe. Learn how supercomputers at Argonne National Laboratory are helping researchers advance our understanding of the mysterious nature of dark matter and dark energy.
The ASCR Leadership Computing Challenge (ALCC) is now seeking proposals for supercomputing time at the ALCF, OLCF, and NERSC. Open to scientists from industry, academia, and national laboratories, the ALCC program aims to further DOE mission science and to broaden community access to leadership computing facilities.
Letters of intent are due Dec. 16, 2019. Full proposals are due Feb. 3, 2020.
Congrats to Argonne National Laboratory's Ian Foster for being named a 2019 SC Distinguished Scientist Fellow, a newly established honor from DOE’s Office of Science. Foster was recognized for his “pioneering work in distributed and high-performance computing with fundamental and long-lasting impacts on both computer science as a discipline and the practice of computing across the Office of Science.”
“This recognition of Argonne’s exceptional computer science program and culture, and the work of my many collaborators over many years, is extremely gratifying,” Foster said. “I am excited to be able to use this support to pursue new research directions at the intersection of artificial intelligence and science.”
The ALCF's 2019 Simulation, Data, and Learning Workshop welcomed several current and prospective facility users to Argonne National Laboratory for guidance on using ALCF supercomputing resources to advance their research.
The annual hands-on workshop is designed to help attendees leverage machine learning and deep learning frameworks, workflow services, containers, and other HPC and data science tools that can help improve application performance and productivity on ALCF systems.
Join us at 10 am (CT) on Friday (10/18) for a special webcast to celebrate Exascale Day!
Experts from the U.S. Department of Energy's Exascale Computing Project, Argonne National Laboratory, Oak Ridge National Laboratory, Lawrence Livermore National Laboratory, Cray Inc., and Hyperion Research will take part in a webinar to discuss how exascale computing will push the boundaries of what’s possible in computational science.
Our Mira supercomputer is being decommissioned at the end of the year, but not before enabling some amazing new science. Researchers from Argonne National Laboratory are currently using the machine to simulate how a large portion of our universe evolved over billions of years.
Elise Jennings and Timothy Williams of Argonne National Laboratory were part of a multi-institutional team that authored Nature Review Physics' first Expert Recommendation this month: "Enabling real-time multi-messenger astrophysics discoveries with deep learning."
The collaborative effort provided recommendations to accelerate the adoption of innovative signal-processing algorithms and computing approaches that can help enhance the potential for discovery in multi-messenger astrophysics.
Argonne National Laboratory is proud to be part of the Grace Hopper Celebration, the world's largest gathering of women technologists! The ALCF's Jini Ramprakash, Haritha Siddabathuni Som, and Janet Jaseckas are among the staff members representing the lab at this year's event.
GHC19 attendees - stop by the Argonne booth (216) to learn about exciting career opportunities at the forefront of science and technology.
Argonne National Laboratory's AI for Science initiative leverages the lab's broad capabilities and world-class facilities, including the ALCF, to explore and expand new artificial intelligence techniques aimed at accelerating scientific discovery and innovation.
Better Scientific Software (BSSw) is seeking applications for its 2020 Fellowship Program. Receive up to $25K for an activity that promotes practices, processes, or tools to improve developer productivity and software sustainability of scientific codes. The deadline to apply is October 15, 2019.
Earlier this month, Argonne National Laboratory hosted an HPC User Forum meeting that brought together users, vendors, and others interested in high-performance computing. The two-day event featured talks on the latest in HPC research, technologies, and initiatives. Argonne's Rick Stevens, Valerie Taylor, and David Martin were among the featured speakers.