Exascale Day
10.18.2022

“Creating an exascale supercomputer is an incredible achievement not just for HPE & AMD and those that worked on Frontier, but for the entire scientific community. Breaking the exascale barrier has been in the making for years and long been a goalpost for the industry. The Frontier supercomputer, powered by AMD EPYC processors and Instinct accelerators, opens the door for a new era of computational advancement and scientific discovery.”

"The ExaFLOP LINPACK result box is checked. This is great news, but our work is not done. ECP’s focus has always been on the longer-term impact and legacy of the project: A collection of applications, libraries, and tools that provide portable, GPU-accelerated computing capabilities for science and engineering, from desktop systems to supercomputers. ECP software is showing the value of DOE’s investment in GPU platforms and providing libraries and tools for other application teams to leverage, as well as examples to study via our application code development and experience. ECP algorithms and software investments will inform the efforts of the hundreds of other applications, libraries, and tools teams who will also need to portably utilize GPUs to achieve the performance levels needed for their own scientific computations. Regardless of where on the HPC spectrum these teams focus, portable GPU execution—whether on a single GPU or thousands of them—must be part of their strategy to remain relevant to HPC customers. ECP investments are meaningful in all these situations, laying groundwork that will pay off for many years to come."

"One of the best things about being on ECP projects is the collaborations we have formed with other labs and universities, and the strong connections with user facilities. I have been privileged to work with many excellent computer scientists, physical scientists, and engineers on both the ExaLearn and ExaFEL ECP projects and am excited about the science we have been a part of enabling."

"It is so exciting to be celebrating Exascale Day now that we have our first Exascale machine, Frontier. For so many years, reaching Exascale has been an aspiration, but now all of our hard work has made it a reality."

"My research involves studying proton-proton collisions, and without computing it’s impossible to analyze all this data. Exascale has the potential to benefit particle physics by introducing significantly more computing power to the field. We could potentially simulate more data, and process data so quickly that if someone had an idea we could quickly produce results versus waiting weeks or more to do the study."

"A decade ago, experts estimated that an exascale computer built with then-current technology would consume 120 MW of energy per year at a cost of $120 million. Reducing that figure to about 20 MW today is a great achievement that makes it practical to continue funding computing progress."

“At the start of the exascale effort, three major challenges were identified: power requirements, resilience and massive parallelism. We have clearly overcome the first, as Frontier consumes just over 19 megawatts per exaFLOP on the High Performance Linpack, and obviously even lower per peak exaFLOP—and future systems will reduce this even further. While the second issue remains a challenge as we bring up new systems, component-level reliability increases, along with system architectures that feature extremely capable individual nodes, has led to resilience being only about as challenging for exascale systems as for earlier large-scale systems. However, while we have identified new ways to structure applications that reveal greater parallelism to exploit, the third challenge is really about making systems easier to use, which remains an area where major advances are needed and will be the primary challenge as we look into the future beyond exascale systems.”

“I’m using AI on Frontier to understand the relation among 300,000 medical concepts from 55 million biological documents so that we will never have a pandemic like COVID in the future.”

“To see the promise of exascale computing being realized after a decade or more of planning and execution is very exciting and rewarding. The extraordinary amount of work by our staff that has gone into preparing our facilities, our software stack and our applications is all coming together at just the right time. Stockpile modernization is putting tremendous demands on our computing needs, and El Capitan will be a huge benefit for the NNSA’s nuclear deterrence and national security missions. It is not an exaggeration to claim that exascale will truly be game-changing, and I can’t wait for our system to arrive next year so we can finally unleash our people on it.”

“The future looks so bright. Everybody in the world will benefit from the research done on this machine.”

"Exascale computing gives us tremendous opportunities to make scientific advancements in plasma physics that were utterly beyond our reach before. Also, it allows us to do more simulations. So instead of doing a single simulation once, we can play with the problem and find out which physics conditions work best. For tumor therapy, this means we can discover new, innovative ways to improve the energy, quality, and precision of high-energy laser-driven electron beams."

“The problems we have to solve keep getting harder and harder. We need bigger and bigger computers to help develop the solutions. At this point, we’re not just predicting the future – we’re inventing it. “

"The whole reason for exascale computing is the science. To advance science, you need exascale computing. I like to think of the exascale computer as a very sophisticated scientific tool. These exascale computing tools will be used to understand something that we couldn’t understand before, so it allows us to push back those frontiers of science."

"To reach exascale computing has been an ambitious goal, which has now been achieved by the Exascale Computing Project, although more work is needed. The ECP also forged a lot of highly valuable collaborations between experts of many diverse areas, such as applications, software and hardware technologies and facilities. It has shown that much can be accomplished if we all work together towards a common goal."

“We are still in early days on the road to exascale, and bringing our efforts to fruition requires multi-disciplinary collaboration throughout the ecosystem to enable a qualitatively different computing experience. We are making it happen. Through partnerships at all levels, ECP scientists, engineers, and administrators work together to make this future a reality. Therefore, let’s take some time today to reflect on how exascale computing can benefit humanity and our planet, every step of the way.”

"The world is facing its biggest challenge in reducing its carbon emission footprint. Moving away from fossil fuels requires the development of new batteries able to store energy for long times without a loss. With the advent of exascale computing, we will be able to run very accurate simulations guided by artificial intelligence that will allow us to design materials that have the exact properties we want in an efficient super battery."

“With the formal acceptance of Frontier at ORNL, the easy part of achieving exascale is complete. Now comes the fun part: seeing what the scientists, engineers, and researchers can accomplish with the leadership infrastructure. To that end, streamlining access to and utilization of Frontier (and eventually to other exascale machines that will come on-line elsewhere in the U.S, Europe, and APAC) becomes paramount. Key to this will be minimizing the time users spend on exascale matters and optimizing the time they spend delivering far-reaching results in their respective domain areas of expertise. It is incumbent upon the leadership sites who are the stewards of the exascale-class machines, along with the vendors who are architecting and delivering the machines to them, to maintain the focus and investment required to unleash the achievements the users of these machines have to offer.”

"Exascale computing lets scientists do more types of computation at the same time than we can on smaller systems. Multi-tasking is especially important because future applications will not be an ‘AI’ or ‘Science’ application. They will be both with AI and simulation closely intertwined together. For instance, my colleagues and I are building exascale applications that continuously improve as simulation data feeds into the AI components that speed parts of the simulation."

“It’s a moonshot. The science that will be done on this machine will be truly game-changing and will have long-term, global impacts on society.”

“Exascale computing will allow us to numerically simulate Multiphysics problems at fine resolution – problems that have been out of our reach.”

"Combining the power of exascale computers with the advent of new observational data now being collected by large telescopes, we hope to obtain new insight into the nature of dark energy. We will carry out simulations of our universe at unprecedented detail and volume and study its formation from the earliest time until today."

"Many of the world’s greatest innovations have been made possible by high performance computing and the great minds behind the development of supercomputing today’s supercomputing technologies. The introduction of the first exascale system, officially ranked as both the fastest and most energy efficient in the world, is truly groundbreaking and proves that excellence in both areas is critical aspect of supercomputing and advancing scientific discoveries.”

"As developers of mathematical software, we are thrilled to provide newest capabilities on our first exascale machine, Frontier. These capabilities enable domain scientists to vastly improve their understandings of myriad physical phenomena, with unprecedented high-fidelity simulations and thorough analysis of a tremendous amount of data."

“My group uses Frontier and develops algorithms that allow us to understand and identify risk factors in real time for cancer prevention and treatment.”

"Almost every aspect of modern life is tied to computing. If we want to make better decisions, if we want to have a more detailed understanding of the world, we need faster computers. Exascale is just the beginning. We need a lot of compute power to solve simulation problems – whether it’s to predict the future climate, to design new batteries, to work on new cancer treatments and new drugs, or even to improve manufacturing or prototype the metaverse."

“We use Frontier to predict and understand new materials for everything from batteries to new types of computing. We have to use full compute power and all the memory on the system, and without them, these simulations would not be possible.”

"Fully exploiting the unprecedented capabilities of exascale computing for transformational advances requires the combined contributions of diverse teams with skills across mathematics, computer science, diverse scientific and engineering disciplines, data science, AI, and more. The ECP community is helping to foster a more vibrant, diverse, and inclusive workforce in high-performance computing through a multipronged Broadening Participation Initiative that is reaching out to talented people from underrepresented groups, including people who are Black or African American, Hispanic/Latinx, American Indian, Alaska Native, Native Hawaiian, Pacific Islanders, women, persons with disabilities, first-generation scholars, and people from smaller colleges and universities. During summer 2022, 16 faculty and 45 students representing 32 institutions (with 82% representing at least one element of diversity) participated in the newly established Sustainable Research Pathways for High-Performance Computing (SRP-HPC) Program, designed to connect students from (and faculty working with) underrepresented groups with DOE lab scientists to encourage lasting collaborations and jump start careers. SRP-HPC participants worked with ECP teams across 9 DOE laboratories to build software technologies that power HPC discoveries and to develop advanced simulation capabilities for chemistry, materials, energy, Earth, and space science, data analytics, optimization, and AI. Next year’s program is under way, spanning ECP and other computational and data science projects."

"Wouldn’t it be great to reduce the carbon footprint of your next overseas vacation? Using exascale computing and exascale-ready codes like Nek5000 and NekRS we are able to simulate engine fuel and hardware interactions at unprecedented insights. We are then able to use this data, knowledge and tools to develop sub-models that may be physics based, or data driven, that can then be transferred to industry so that they can design the next generation aircraft engines with sustainable biofuels."