Tagged: Supercomputing Toggle Comment Threads | Keyboard Shortcuts

  • richardmitnick 6:52 am on July 22, 2019 Permalink | Reply
    Tags: , , , , , , , QMCPACK, , Supercomputing, The quantum Monte Carlo (QMC) family of these approaches is capable of delivering the most highly accurate calculations of complex materials without biasing the results of a property of interest.   

    From insideHPC: “Supercomputing Complex Materials with QMCPACK” 

    From insideHPC

    July 21, 2019

    In this special guest feature, Scott Gibson from the Exascale Computing Project writes that computer simulations based on quantum mechanics are getting a boost through QMCPACK.

    2

    The theory of quantum mechanics underlies explorations of the behavior of matter and energy in the atomic and subatomic realms. Computer simulations based on quantum mechanics are consequently essential in designing, optimizing, and understanding the properties of materials that have, for example, unusual magnetic or electrical properties. Such materials would have potential for use in highly energy-efficient electrical systems and faster, more capable electronic devices that could vastly improve our quality of life.

    Quantum mechanics-based simulation methods render robust data by describing materials in a truly first-principles manner. This means they calculate electronic structure in the most basic terms and thus can allow speculative study of systems of materials without reference to experiment, unless researchers choose to add parameters. The quantum Monte Carlo (QMC) family of these approaches is capable of delivering the most highly accurate calculations of complex materials without biasing the results of a property of interest.

    An effort within the US Department of Energy’s Exascale Computing Project (ECP) is developing a QMC methods software named QMCPACK to find, predict, and control materials and properties at the quantum level. The ultimate aim is to achieve an unprecedented and systematically improvable accuracy by leveraging the memory and power capabilities of the forthcoming exascale computing systems.

    Greater Accuracy, Versatility, and Performance

    One of the primary objectives of the QMCPACK project is to reduce errors in calculations so that predictions concerning complex materials can be made with greater assurance.

    “We would like to be able to tell our colleagues in experimentation that we have confidence that a certain short list of materials is going to have all the properties that we think they will,” said Paul Kent of Oak Ridge National Laboratory and principal investigator of QMCPACK. “Many ways of cross-checking calculations with experimental data exist today, but we’d like to go further and make predictions where there aren’t experiments yet, such as a new material or where taking a measurement is difficult—for example, in conditions of high pressure or under an intense magnetic field.”

    The methods the QMCPACK team is developing are fully atomistic and material specific. This refers to having the capability to address all of the atoms in the material—whether it be silver, carbon, cerium, or oxygen, for example—compared with more simplified lattice model calculations where the full details of the atoms are not included.

    The team’s current activities are restricted to simpler, bulk-like materials; but exascale computing is expected to greatly widen the range of possibilities.

    “At exascale not only the increase in compute power but also important changes in the memory on the machines will enable us to explore material defects and interfaces, more-complex materials, and many different elements,” Kent said.

    With the software engineering, design, and computational aspects of delivering the science as the main focus, the project plans to improve QMCPACK’s performance by at least 50x. Based on experimentation using a mini-app version of the software, and incorporating new algorithms, the team achieved a 37x improvement on the pre-exascale Summit supercomputer versus the Titan system.

    ORNL IBM AC922 SUMMIT supercomputer, No.1 on the TOP500. Credit: Carlos Jones, Oak Ridge National Laboratory/U.S. Dept. of Energy

    ORNL Cray XK7 Titan Supercomputer, once the fastest in the world, to be decommissioned

    One Robust Code

    “We’re taking the lessons we’ve learned from developing the mini app and this proof of concept, the 37x, to update the design of the main application to support this high efficiency, high performance for a range of problem sizes,” Kent said. “What’s crucial for us is that we can move to a single version of the code with no internal forks, to have one source supporting all architectures. We will use all the lessons we’ve learned with experimentation to create one version where everything will work everywhere—then it’s just a matter of how fast. Moreover, in the future we will be able to optimize. But at least we won’t have a gap in the feature matrix, and the student who is running QMCPACK will always have all features work.”

    As an open-source and openly developed product, QMCPACK is improving via the help of many contributors. The QMCPACK team recently published the master citation paper for the software’s code; the publication has 48 authors with a variety of affiliations.

    “Developing these large science codes is an enormous effort,” Kent said. “QMCPACK has contributors from ECP researchers, but it also has many past developers. For example, a great deal of development was done for the Knights Landing processor on the Theta supercomputer with Intel. This doubled the performance on all CPU-like architectures.”

    ANL ALCF Theta Cray XC40 supercomputer

    A Synergistic Team

    The QMCPACK project’s collaborative team draws talent from Argonne, Lawrence Livermore, Oak Ridge, and Sandia National Laboratories.




    It also benefits from collaborations with Intel and NVIDIA.

    3

    The composition of the staff is nearly equally divided between scientific domain specialists and people centered on the software engineering and computer science aspects.

    “Bringing all of this expertise together through ECP is what has allowed us to perform the design study, reach the 37x, and improve the architecture,” Kent said. “All the materials we work with have to be doped, which means incorporating additional elements in them. We can’t run those simulations on Titan but are beginning to do so on Summit with improvements we have made as part of our ECP project. We are really looking forward to the opportunities that will open up when the exascale systems are available.”

    See the full article here .

    five-ways-keep-your-child-safe-school-shootings

    Please help promote STEM in your local schools.

    Stem Education Coalition

    Founded on December 28, 2006, insideHPC is a blog that distills news and events in the world of HPC and presents them in bite-sized nuggets of helpfulness as a resource for supercomputing professionals. As one reader said, we’re sifting through all the news so you don’t have to!

    If you would like to contact me with suggestions, comments, corrections, errors or new company announcements, please send me an email at rich@insidehpc.com. Or you can send me mail at:

    insideHPC
    2825 NW Upshur
    Suite G
    Portland, OR 97239

    Phone: (503) 877-5048

     
  • richardmitnick 8:08 am on July 18, 2019 Permalink | Reply
    Tags: , , , Proposed Expanse supercomputer, , Supercomputing   

    From insideHPC: “NSF Funds $10 Million for ‘Expanse’ Supercomputer at SDSC” 

    From insideHPC

    July 17, 2019

    SDSC Triton HP supercomputer

    SDSC Gordon-Simons supercomputer

    SDSC Dell Comet supercomputer

    The San Diego Supercomputer Center (SDSC) at the University of California San Diego, has been awarded a five-year grant from the National Science Foundation (NSF) valued at $10 million to deploy Expanse, a new supercomputer designed to advance research that is increasingly dependent upon heterogeneous and distributed resources.

    1
    Credit: Ben Tolo, SDSC

    “The name of our new system says it all,” said SDSC Director Michael Norman, the Principal Investigator (PI) for Expanse, and a computational astrophysicist. “As a standalone system, Expanse represents a substantial increase in the performance and throughput compared to our highly successful, NSF-funded Comet supercomputer. But with innovations in cloud integration and composable systems, as well as continued support for science gateways and distributed computing via the Open Science Grid, Expanse will allow researchers to push the boundaries of computing and answer questions previously not possible.”

    The NSF award, which runs from October 1, 2020 to September 30, 2025, is valued at $10 million for acquisition and deployment of Expanse. An additional award will be made in the coming months to support Expanse operations and user support.

    Like SDSC’s Comet [above]supercomputer, which is slated to remain in operation through March 2021, Expanse will continue to serve what is referred to as the ‘long tail’ of science. Virtually every discipline, from multi-messenger astronomy, genomics, and the social sciences, as well as more traditional ones such as earth sciences and biology, depend upon these medium-scale, innovative systems for much of their productive computing.

    “Comet’s focus on reliability, throughput, and usability has made it one of the most successful resources for the national community, supporting tens-of-thousands of users across all domains,” said SDSC Deputy Director Shawn Strande, a co-PI and project manager for the new program. “So we took an evolutionary approach with Expanse, assessing community needs, then working with our vendor partners including Dell, Intel, NVIDIA, Mellanox, and Aeon, to design an even better system.”

    Projected to have a peak speed of 5 Petaflop/s, Expanse will about double the performance of Comet with Intel’s next-generation processors and NVIDIA’s GPUs. Expanse will increase throughput of real-world workloads by a factor of at least 1.3 for both CPU and GPU applications relative to Comet, while supporting an even larger and more diverse research community. Expanse’s accelerated compute nodes will provide a much-needed GPU capability to the user community, serving both well-established applications in areas such as molecular dynamics as well as rapidly growing demand for resources to support machine learning and artificial intelligence. A low-latency interconnect based on Mellanox High Data Rate (HDR) InfiniBand will support a fabric topology optimized for jobs of one to a few thousand cores that require medium-scale parallelism.

    Expanse will support the growing diversity in computational and data-intensive workloads with a rich storage environment that includes 12PB of high-performance Lustre, 7PB of object storage, and more than 800TB of NVMe solid state storage.

    “While Expanse will easily support traditional batch-scheduled HPC applications, breakthrough research is increasingly dependent upon carrying out complex workflows that may include near real-time remote sensor data ingestion and big data analysis, interactive data exploration and visualization as well as large-scale computation,” said SDSC Chief Data Science Officer Ilkay Altintas, an Expanse co-PI and the director of SDSC’s Workflows for Data Science (WorDS) Center of Excellence. “One of the key innovations in Expanse is its ability to support so-called composable systems at the continuum of computing with dynamic capabilities. Using tools such as Kubernetes, and workflow software we have developed over the years for projects including the NSF-funded WIFIRE and CHASE-CI programs, Expanse will extend the boundaries of what is possible by integration with the broader computational and data ecosystem.”

    Increasingly, this ecosystem includes public cloud resources. Expanse will feature direct scheduler-integration with the major cloud providers, leveraging high-speed networks to ease data movement to/from the cloud, and opening up new modes of computing made possible by the combination of Expanse’s powerful HPC capabilities and ubiquity of cloud resources and software.

    Like Comet, Expanse will be a key resource within the NSF’s Extreme Science and Engineering Discovery Environment (XSEDE), which comprises the most advanced collection of integrated digital resources and services in the world.

    More details about the program will be available at the SDSC display at the SC19 in Denver.

    “The capabilities and services these awards will enable the research community to explore new computing models and paradigms,” said Manish Parashar, Office Director of NSF’s Office of Advanced Cyberinfrastructure, which funded this award. “These awards complement NSF’s long-standing investment in advanced computational infrastructure, providing much-needed support for the full range of innovative computational- and data-intensive research being conducted across all of science and engineering.”

    See the full article here .

    five-ways-keep-your-child-safe-school-shootings

    Please help promote STEM in your local schools.

    Stem Education Coalition

    Founded on December 28, 2006, insideHPC is a blog that distills news and events in the world of HPC and presents them in bite-sized nuggets of helpfulness as a resource for supercomputing professionals. As one reader said, we’re sifting through all the news so you don’t have to!

    If you would like to contact me with suggestions, comments, corrections, errors or new company announcements, please send me an email at rich@insidehpc.com. Or you can send me mail at:

    insideHPC
    2825 NW Upshur
    Suite G
    Portland, OR 97239

    Phone: (503) 877-5048

     
  • richardmitnick 9:12 am on July 10, 2019 Permalink | Reply
    Tags: , , Globus Data Transfer, , , , Supercomputing   

    From insideHPC: “Argonne Team Breaks Record with 2.9 Petabytes Globus Data Transfer” 

    From insideHPC

    Today the Globus research data management service announced the largest single file transfer in its history: a team led by Argonne National Laboratory scientists moved 2.9 petabytes of data as part of a research project involving three of the largest cosmological simulations to date.

    1

    “Storage is in general a very large problem in our community — the Universe is just very big, so our work can often generate a lot of data,” explained Katrin Heitmann, Argonne physicist and computational scientist and an Oak Ridge National Laboratory Leadership Computing Facility (OLCF) Early Science user.

    “Using Globus to easily move the data around between different storage solutions and institutions for analysis is essential.”

    The data in question was stored on the Summit supercomputer at OLCF, currently the world’s fastest supercomputer according to the Top500 list published June 18, 2019. Globus was used to move the files from disk to tape, a key use case for researchers.

    ORNL IBM AC922 SUMMIT supercomputer, No.1 on the TOP500. Credit: Carlos Jones, Oak Ridge National Laboratory/U.S. Dept. of Energy

    “Due to its uniqueness, the data is very precious and the analysis will take time,” said Dr. Heitmann. “The first step after the simulations were finished was to make a backup copy of the data to HPSS, so we can move the data back and forth between disk and tape and thus carry out the analysis in steps. We use Globus for this work due to its speed, reliability, and ease of use.”

    “With exascale imminent, AI on the rise, HPC systems proliferating, and research teams more distributed than ever, fast, secure, reliable data movement and management are now more important than ever,” said Ian Foster, Globus co-founder and director of Argonne’s Data Science and Learning Division. “We tend to take these functions for granted, and yet modern collaborative research would not be possible without them.”

    “Globus has underpinned groundbreaking research for decades. We could not be prouder of our role in helping scientists do their world-changing work, and we’re happy to see projects like this one continue to push the boundaries of what Globus can achieve. Congratulations to Dr. Heitmann and team!”

    “When it comes to data transfer performance, “the most important part is reliability,” says Dr. Heitmann. “It is basically impossible for me as a user to check the very large amounts of data upon arrival after a transfer has finished. The analysis of the data often uses a subset of the data, so it would take quite a while until bad data would be discovered and at that point we might not have the data anymore at the source. So the reliability aspects of Globus are key.”

    “Of course, speed is also important. If the transfers were very slow, given the amount of data we transfer, we would have had a problem. So it’s good to be able to rely on Globus for fast data movement as well. We are also grateful to Oak Ridge for access to Summit and for their excellent setup of data transfer nodes enabling the use of Globus for HPSS transfers. This work would not have been possible otherwise.”

    See the full article here .

    five-ways-keep-your-child-safe-school-shootings

    Please help promote STEM in your local schools.

    Stem Education Coalition

    Founded on December 28, 2006, insideHPC is a blog that distills news and events in the world of HPC and presents them in bite-sized nuggets of helpfulness as a resource for supercomputing professionals. As one reader said, we’re sifting through all the news so you don’t have to!

    If you would like to contact me with suggestions, comments, corrections, errors or new company announcements, please send me an email at rich@insidehpc.com. Or you can send me mail at:

    insideHPC
    2825 NW Upshur
    Suite G
    Portland, OR 97239

    Phone: (503) 877-5048

     
  • richardmitnick 2:38 pm on July 8, 2019 Permalink | Reply
    Tags: "Jülich Supercomputing Centre Announces Quantum Computing Research Partnership with Google", , , , , Supercomputing   

    From insideHPC: “Jülich Supercomputing Centre Announces Quantum Computing Research Partnership with Google” 

    From insideHPC

    July 8, 2019

    Today the Jülich Supercomputing Centre announced it is partnering with Google in the field of quantum computing research. The partnership will include joint research and expert trainings in the fields of quantum technologies and quantum algorithms and the mutual use of quantum hardware.

    JUWELS-Jülich Wizard for European Leadership Science

    JURECA-Jülich Research on Exascale Cluster Architectures
    2

    DEEP-EST – Dynamical Exascale Entry Platform – Prototype System DELL R640
    3

    Pilot systems of the Human Brain Project – JULIA and JURON
    JULIA from Cray and JURON from a consortium of IBM and NVIDIA
    4

    “Quantum computers have the potential to solve certain types of calculations much more efficiently than today’s technologies can,” said Peter Altmaier, Federal Minister for Economic Affairs and Energy. “Quantum computers and quantum algorithms are therefore very important technologies which will shape the future and are being followed closely around the world. At present, quantum computers are still very much at in their infancy, and it is difficult to predict what will become possible – and what perhaps will not. Researchers still have a lot of basic research to do in this area. It was the same situation when we were developing today’s computers. I am therefore delighted that Google and Forschungszentrum Jülich have decided to cooperate in the important forward-looking field of quantum computers.”

    Google has been working on the development of quantum processors and quantum algorithms for years. Exploring new technologies for quantum computers is also a key research focus at Forschungszentrum Jülich. The German research center will operate and make publicly accessible a European quantum computer with 50 to 100 superconducting qubits, to be developed within the EU’s Quantum Flagship Program, a large-scale initiative in the field of quantum technologies funded at the 1 billion € level on a 10 years timescale.

    Google and Forschungszentrum Jülich will support each other especially in training junior researchers and experts. “A shortage of specialists, like in the field of artificial intelligence, is also foreseeable in the field of quantum computing. For this reason, we invest in training and promoting top academic talent” says Dr. Markus Hoffmann, Head of Quantum Partnerships at Google.

    The partnership includes regular research exchange. “Hands-on workshop and spring schools will be organised at Forschungszentrum Jülich. The Jülich UNified Infrastructure for Quantum computing (JUNIQ), a European quantum computer user facility planned for the Jülich Supercomputing Centre (JSC), will be available for training industry professionals, and will be accessible in the cloud to European users,” says Prof. Kristel Michielsen from the JSC, head of the research group Quantum Information Processing.

    Forschungszentrum Jülich and Google have commenced the team work. Prof. Kristel Michielsen and Prof. Tommaso Calarco from Forschungszentrum Jülich have received Google Faculty Research Awards in 2018. Prof. Frank Wilhelm-Mauch, recipient of a Google Faculty Research Award in 2015, is a collaborator of Forschungszentrum Jülich within the subproject OpenSuperQ of the European “Quantum Flagship” project.

    See the full article here .

    five-ways-keep-your-child-safe-school-shootings

    Please help promote STEM in your local schools.

    Stem Education Coalition

    Founded on December 28, 2006, insideHPC is a blog that distills news and events in the world of HPC and presents them in bite-sized nuggets of helpfulness as a resource for supercomputing professionals. As one reader said, we’re sifting through all the news so you don’t have to!

    If you would like to contact me with suggestions, comments, corrections, errors or new company announcements, please send me an email at rich@insidehpc.com. Or you can send me mail at:

    insideHPC
    2825 NW Upshur
    Suite G
    Portland, OR 97239

    Phone: (503) 877-5048

     
  • richardmitnick 10:13 am on June 26, 2019 Permalink | Reply
    Tags: , IBM POWER9 NVIDIA Pangea III supercomputer, , Supercomputing   

    From insideHPC: “IBM Deploys World’s Most Powerful Commercial Supercomputer for Total” 

    From insideHPC

    IBM has deployed the world’s most powerful commercial supercomputer at Total, a global energy company. Coming in at number 11 on the TOP500, the new IBM POWER9-based supercomputer will help Total more accurately locate new resources and better assess the potential associated revenue opportunities.

    IBM POWER9 NVIDIA Pangea III supercomputer

    “Pangea III’s additional computing power enhances Total’s operational excellence. It enables Total to reduce geological risks in exploration and development, accelerate project maturation and delivery, and increases the value of our assets through optimized field operations, with all this at lower cost,” stated Arnaud Breuillac, President of Total Exploration & Production.”

    Exploring new oil and gas prospects, Total must first create an accurate picture of what lies underground, through the use of seismic data acquisition during exploration campaigns. Geoscientists then use these images to identify where oil and gas resources lie. This process creates massive amounts of data that IBM POWER9 systems can easily handle with industry-exclusive technology.

    In a competitive bidding process, Total selected IBM due to an industry-leading approach to GPU-accelerated computing. IBM worked with NVIDIA to jointly develop the industry’s only CPU-to-GPU NVIDIA NVLink connection, which allows for 5.6x faster memory bandwidth between the IBM POWER9 CPU and NVIDIA Tesla V100 Tensor Core GPUs than the compared x86-based systems. This will help Total to process the vast amount of data required in seismic modeling to get more accurate insights faster than they previously could.

    “Pangea III is designed to help Total not only improve performance but also improve energy efficiency in their HPC workloads. According to Total, Pangea III requires just 1.5 Megawatts, compared to 4.5 MW for its predecessor system.”

    New capabilities

    Pangea III delivered 17.86 Petaflops on the LINPACK benchmark and has a storage capacity of 50 petabytes. Some of the fields where Total states that they plan to apply Pangea III’s compute power include:

    Higher Resolution Seismic Imaging in exploration and development phase – new algorithms will process large data sets to produce higher resolution images to help Total more reliably locate hydrocarbons below ground. This is especially useful in complex geological environments where layers of salt in the Earth can make getting accurate readings challenging, such as in Brazil, the Gulf of Mexico, Angola and the Eastern Mediterranean.
    Reliable Development and Production Models – through increased computing power, Total will use innovative reservoir simulation methods that, for example, integrate the field’s production history to generate more reliable predictive production models faster
    Asset Valuation and Selectivity – through early assessment of the value of exploration acreage and asset opportunities, Total will enhance selectivity in their new ventures.

    “Based on the same IBM technology found in Summit and Sierra, the world’s smartest supercomputers, Pangea III demonstrates that IBM Power Systems are not just for the large government or research organizations,” said David Turek, Vice President of Exascale Systems for IBM Systems. “The world’s largest businesses, like Total, are now tapping that same technology to profoundly change how they operate. It also gives them room to explore the role IBM Power Systems can play against their most data-intensive workloads like hybrid could and AI.”

    See the full article here .

    five-ways-keep-your-child-safe-school-shootings

    Please help promote STEM in your local schools.

    Stem Education Coalition

    Founded on December 28, 2006, insideHPC is a blog that distills news and events in the world of HPC and presents them in bite-sized nuggets of helpfulness as a resource for supercomputing professionals. As one reader said, we’re sifting through all the news so you don’t have to!

    If you would like to contact me with suggestions, comments, corrections, errors or new company announcements, please send me an email at rich@insidehpc.com. Or you can send me mail at:

    insideHPC
    2825 NW Upshur
    Suite G
    Portland, OR 97239

    Phone: (503) 877-5048

     
  • richardmitnick 9:51 am on June 26, 2019 Permalink | Reply
    Tags: , Frontier Shasta based Exascale supercomputer, , Supercomputing   

    From insideHPC: “Cray to Deliver First Exabyte HPC Storage System for Frontier Supercomputer” 

    From insideHPC

    June 25, 2019

    At ISC 2019, Cray announced plans to deliver the worlds first Exabyte HPC storage system to Oak Ridge National Lab. As part of the Frontier CORAL-2 contract DOE and ORNL, the next generation Cray ClusterStor storage file system will be integrated as part of ORNL’s Frontier exascale supercomputer, built on Cray’s Shasta architecture.

    ORNL Cray Frontier Shasta based Exascale supercomputer with Slingshot interconnect featuring high-performance AMD EPYC CPU and AMD Radeon Instinct GPU technology

    “We are excited to continue our partnership with ORNL to collaborate in developing a next generation storage solution that will deliver the capacity and throughput needed to support the dynamic new research that will be done on the Frontier exascale system for years to come,” said John Dinning, chief product officer at Cray. “By delivering a new hybrid storage solution that is directly connected to the Slingshot network, users will be able to drive data of any size, access pattern or scale to feed their converged modeling, simulation and AI workflows.”

    The storage solution is a new design for the data-intensive workloads of the exascale era and will be based on next generation Cray ClusterStor storage and the Cray Slingshot high-speed interconnect. The storage system portion of the previously-announced Frontier contract is valued at more than $50 million, which is the largest single Cray ClusterStor win to date. The Frontier system is expected to be delivered in 2021.

    The new storage solution will be based on the next generation of Cray’s ClusterStor storage line and will be comprised of over one exabyte (EB) of hybrid flash and high capacity storage running the Lustre® parallel file system. One exabyte of storage is 1,000 petabytes (or one quintillion bytes), which is enough capacity to store more than 200 million high definition movies. The storage solution will be directly connected to ORNL’s Frontier system via the Slingshot system interconnect to enable seamless scaling of diverse modeling, simulation, analytics and AI workloads running simultaneously on the system. The Frontier system is anticipated to debut in 2021 as the world’s most powerful computer with a performance of greater than 1.5 exaflops.

    Compared to the storage for ORNL’s current Summit supercomputer, this next generation solution is more than four times the capacity (more than 1 EB (or 1,000 PB) versus 250 PB), and more than four times the throughput (up to 10 TB/s versus 2.5 TB/s) of their existing Spectrum Scale-based storage system.

    ORNL IBM AC922 SUMMIT supercomputer, No.1 on the TOP500. Credit: Carlos Jones, Oak Ridge National Laboratory/U.S. Dept. of Energy

    The new Cray ClusterStor storage solution for ORNL will be comprised of over 40 cabinets of storage and provide more than 1 EB of total capacity across two tiers of storage to support random and streaming access of data. The primary tier is a flash tier for high-performance scratch storage and the secondary tier is a hard disk tier for high capacity storage. The new storage system will be a center-wide system at ORNL in support of the Frontier exascale system and will be accessed by the Lustre global parallel file system with ZFS local volumes all in a single global POSIX namespace, which will make it the largest single high-performance file system in the world.

    “HPC storage systems have traditionally utilized large arrays of hard disks accessed via large and predictable reads and writes of data. This is in stark contrast to AI and machine learning workloads, which typically have a mix of random and sequential access of small and large data sizes. As a result, traditional storage systems are not well suited for the combined usage of these workloads given the mix of data access and the need for an intelligent high-speed system interconnect to quickly move massive amounts of data on and off the supercomputer to enable these diverse workloads to run simultaneously on exascale systems like Frontier.”

    The next generation ClusterStor-based storage solution addresses these challenges head on by providing a blend of flash and capacity storage to support complex access patterns, a powerful new software stack for improved manageability and tiering of data, and seamless scaling across both compute and storage through direct connection to the Slingshot high-speed network. In addition to scaling, the direct connection of storage to the Slingshot network eliminates the need for storage routers that are required in most traditional HPC networks. This results in lower cost, lower complexity and lower latency in the system overall, thus delivering higher unprecedented application performance and ROI. Additionally, since Slingshot is ethernet compatible, it can also enable seamless interoperability with existing third party network storage as well as with other data and compute sources.

    Cray’s Shasta supercomputers, ClusterStor storage and the Slingshot interconnect are quickly becoming the leading technology choices for the exascale era by combining the performance and scale of supercomputing with the productivity of cloud computing and full datacenter interoperability. The new compute, software, storage and interconnect capabilities being pioneered for leading research labs like ORNL are being productized as standard offerings from Cray for research and enterprise customers alike, with expected availability starting at the end of 2019.

    See the full article here .

    five-ways-keep-your-child-safe-school-shootings

    Please help promote STEM in your local schools.

    Stem Education Coalition

    Founded on December 28, 2006, insideHPC is a blog that distills news and events in the world of HPC and presents them in bite-sized nuggets of helpfulness as a resource for supercomputing professionals. As one reader said, we’re sifting through all the news so you don’t have to!

    If you would like to contact me with suggestions, comments, corrections, errors or new company announcements, please send me an email at rich@insidehpc.com. Or you can send me mail at:

    insideHPC
    2825 NW Upshur
    Suite G
    Portland, OR 97239

    Phone: (503) 877-5048

     
  • richardmitnick 10:23 am on June 23, 2019 Permalink | Reply
    Tags: , , , Mellanox HDR 200G InfiniBand is powering next-gen supercomputers, Supercomputing   

    From insideHPC: “Mellanox HDR 200G InfiniBand is powering next-gen supercomputers” 

    From insideHPC

    June 23, 2019

    Today Mellanox announced that HDR 200G InfiniBand is powering the next generation of supercomputers world-wide, enabling higher levels of research and scientific discovery. HDR 200G InfiniBand solutions include the ConnectX-6 adapters, Mellanox Quantum switches, LinkX cables and transceivers and software packages. With its highest data throughput, extremely low latency, and smart In-Network Computing acceleration engines, HDR InfiniBand provides world leading performance and scalability for the most demanding compute and data applications.

    HDR 200G InfiniBand introduces new offload and acceleration engines, for delivering leading performance and scalability for high-performance computing, artificial intelligence, cloud, storage, and other applications. InfiniBand, a standards-based interconnect technology, enjoys the continuous development of new capabilities, while maintaining backward and forward software compatibility. InfiniBand is the preferred choice for world leading supercomputers, replacing lower performance or proprietary interconnect options.

    “We are proud to have our HDR InfiniBand solutions accelerate supercomputers around the world, enhance research and discoveries, and advancing Exascale programs,” said Gilad Shainer, senior vice president of marketing at Mellanox Technologies. “InfiniBand continues to gain market share, and be selected by many research, educational and government institutes, weather and climate facilities, and commercial organizations. The technology advantages of InfiniBand make it the interconnect of choice for compute and storage infrastructures.”

    The Texas Advanced Computing Center’s (TACC) Frontera supercomputer, funded by the National Science Foundation, is the fastest supercomputer at any U.S. university and one of the most powerful systems in the world.

    TACC Frontera Dell EMC supercomputer fastest at any university

    Ranked #5 on the June 2019 TOP500 Supercomputers list, Frontera utilizes HDR InfiniBand, and in particular multiple 800-port HDR InfiniBand switches, to deliver unprecedented computing power for science and engineering.

    “HDR InfiniBand enabled us to build a world-leading, 8,000+ node, top 5 supercomputer that will serve our users’ needs for the next several years,” said Dan Stanzione, TACC Executive Director. “We appreciate the deep collaboration with Mellanox and are proud to host one of the fastest supercomputers in the world. We look forward to utilizing the advanced routing capabilities and the In-Network Computing acceleration engines to enhance our users’ research activities and scientific discoveries.”

    Located at the Mississippi State University High Performance Computing Collaboratory, the new HDR InfiniBand-based Orion supercomputer will accelerate the university research, educational and service activities.

    Dell EMC Orion supercomputer at Mississippi State University

    Ranked #62 on the June 2019 TOP500 list, the 1800-node supercomputer leverages the performance advantages of HDR InfiniBand and its application acceleration engines to provide new levels of application performance and scalability.

    “HDR InfiniBand brings us leading performance and the ability to build very scalable and cost efficient supercomputers utilizing its high switch port density and configurable network topology,” said Trey Breckenridge, Director for High Performance Computing at Mississippi State University. “Over 16 years ago MSU became one of the first adopters of the InfiniBand technology in HPC. We are excited to continue that legacy by leveraging the latest InfiniBand technology to enhance the capabilities of our newest HPC system.”

    CSC, the Finnish IT Center for Science, and the Finnish Meteorological Institute Selected HDR 200G InfiniBand to accelerate a multi-phase supercomputer program. The program will serve researchers in Finnish universities and research institutes, enhancing their research into climate science, renewable energy, astrophysics, nanomaterials, and bioscience, among a wide range of exploration activities. The first supercomputer is ranked #166 on the TOP500 list.

    “The new supercomputer will enable our researchers and scientists to leverage the most efficient HPC and AI platform to enhance their competitiveness for years to come,” said Pekka Lehtovuori, Director of services for research at CSC. “The HDR InfiniBand technology, and the Dragonfly+ network topology will provide our users with leading performance and scalability while optimizing our total cost of ownership.”

    Cygnus is the first HDR InfiniBand supercomputer in Japan, located in the Center for Computational Sciences at the University of Tsukuba.

    Cygnus FPGA GPU supercomputer at University of Tsukuba Japan

    Ranked #264 on the TOP500 list, Cygnus leverages HDR InfiniBand to connect CPUs, GPUs and FPGAs together, enabling accelerated research in the areas of astrophysics, particle physics, material science, life, meteorology and artificial intelligence.

    The Center for Development of Advanced Computing (C-DAC) has selected HDR InfiniBand for India’s national supercomputing mission. The C-DAC HDR InfiniBand supercomputer advances India’s research, technology, and product development capabilities.

    “The Center for Development of Advanced Computing (C-DAC), an autonomous R&D institution under the Ministry of Electronics and IT, Government of India with its focus in Advanced Computing is uniquely positioned to establish dependable and secure Exascale Ecosystem offering services in various domains. As our nation embarks upon its most revolutionary phase of Digital Transformation, C-DAC has committed itself to explore and engage in the avant-garde visionary areas excelling beyond in the present areas of research transforming human lives through technological advancement,” said Dr Hemant Darbari, Director General, C-DAC.”

    See the full article here .

    five-ways-keep-your-child-safe-school-shootings

    Please help promote STEM in your local schools.

    Stem Education Coalition

    Founded on December 28, 2006, insideHPC is a blog that distills news and events in the world of HPC and presents them in bite-sized nuggets of helpfulness as a resource for supercomputing professionals. As one reader said, we’re sifting through all the news so you don’t have to!

    If you would like to contact me with suggestions, comments, corrections, errors or new company announcements, please send me an email at rich@insidehpc.com. Or you can send me mail at:

    insideHPC
    2825 NW Upshur
    Suite G
    Portland, OR 97239

    Phone: (503) 877-5048

     
  • richardmitnick 11:17 am on June 19, 2019 Permalink | Reply
    Tags: , , , LLNL’s Lassen IBM NVIDIA supercomputer leaps to No. 10 on TOP500 list, Supercomputing   

    From Lawrence Livermore National Laboratory: “LLNL’s Lassen supercomputer leaps to No. 10 on TOP500 list, Sierra remains No. 2” 

    From Lawrence Livermore National Laboratory

    June 18, 2019
    Jeremy Thomas
    thomas244@llnl.gov
    925-422-5539

    1
    Lawrence Livermore National Laboratory’s Lassen IBM NVIDIA supercomputer

    Lawrence Livermore National Laboratory’s Lassen joined its companion system Sierra in the top 10 of the TOP500 list of the world’s most powerful supercomputers, announced Monday at the 2019 International Supercomputing Conference (ISC19) in Frankfurt, Germany.

    Lassen, an unclassified, heterogenous IBM/NVIDIA system with the same architecture as Sierra but smaller, placed No. 10 on the list with a High Performance Linpack (HPL) benchmark score of 18.2 petaFLOPS (18.2 quadrillion point operations per second) boosting its original 15.4 petaFLOP performance from last November. Sierra, LLNL’s classified system that went into production earlier this year, remained unchanged in the second spot at 94.6 petaflops.

    “We are pleased with the results of the June 2019 TOP500 list, in which not only does Sierra continue to occupy the second position but also Lassen has risen to tenth,” said Bronis de Supinski, chief technical officer for Livermore Computing. “These successes demonstrate that LLNL’s strategy of both programmatic and institutional investments supports the complete range of applications required to meet our mission.”

    The improved HPL score for Lassen was attributed to an upgrade on the system, according to a TOP500 press release. LLNL’s IBM/Blue Gene system Sequoia, which had been the 10th most powerful computer in the world in the previous list and is expected to be retired later this year, dropped to 13th.

    LLNL Sequoia IBM Blue Gene Q petascale supercomputer

    Oak Ridge National Laboratory’s Summit, also an IBM/NVIDIA supercomputer, maintained its top spot on the list and slightly improved its result from six months ago, delivering a record 148.6 petaFLOPS. Los Alamos National Laboratory’s Trinity, another Department of Energy/National Nuclear Security Administration supercomputer, placed seventh at 20.2 petaFLOPS.

    ORNL IBM AC922 SUMMIT supercomputer, No.1 on the TOP500. Credit: Carlos Jones, Oak Ridge National Laboratory/U.S. Dept. of Energy

    The 53rd edition of the TOP500 marks a milestone. For the first time in the 26-year history of the list, all 500 systems on the list registered HCL benchmark scores of a petaFLOP or more. The benchmark reflects the performance of a dedicated system for solving a dense system of linear equations.

    See the full article here .


    five-ways-keep-your-child-safe-school-shootings

    Please help promote STEM in your local schools.

    Stem Education Coalition

    LLNL Campus

    Operated by Lawrence Livermore National Security, LLC, for the Department of Energy’s National Nuclear Security Administration
    Lawrence Livermore National Laboratory (LLNL) is an American federal research facility in Livermore, California, United States, founded by the University of California, Berkeley in 1952. A Federally Funded Research and Development Center (FFRDC), it is primarily funded by the U.S. Department of Energy (DOE) and managed and operated by Lawrence Livermore National Security, LLC (LLNS), a partnership of the University of California, Bechtel, BWX Technologies, AECOM, and Battelle Memorial Institute in affiliation with the Texas A&M University System. In 2012, the laboratory had the synthetic chemical element livermorium named after it.
    LLNL is self-described as “a premier research and development institution for science and technology applied to national security.” Its principal responsibility is ensuring the safety, security and reliability of the nation’s nuclear weapons through the application of advanced science, engineering and technology. The Laboratory also applies its special expertise and multidisciplinary capabilities to preventing the proliferation and use of weapons of mass destruction, bolstering homeland security and solving other nationally important problems, including energy and environmental security, basic science and economic competitiveness.

    The Laboratory is located on a one-square-mile (2.6 km2) site at the eastern edge of Livermore. It also operates a 7,000 acres (28 km2) remote experimental test site, called Site 300, situated about 15 miles (24 km) southeast of the main lab site. LLNL has an annual budget of about $1.5 billion and a staff of roughly 5,800 employees.

    LLNL was established in 1952 as the University of California Radiation Laboratory at Livermore, an offshoot of the existing UC Radiation Laboratory at Berkeley. It was intended to spur innovation and provide competition to the nuclear weapon design laboratory at Los Alamos in New Mexico, home of the Manhattan Project that developed the first atomic weapons. Edward Teller and Ernest Lawrence,[2] director of the Radiation Laboratory at Berkeley, are regarded as the co-founders of the Livermore facility.

    The new laboratory was sited at a former naval air station of World War II. It was already home to several UC Radiation Laboratory projects that were too large for its location in the Berkeley Hills above the UC campus, including one of the first experiments in the magnetic approach to confined thermonuclear reactions (i.e. fusion). About half an hour southeast of Berkeley, the Livermore site provided much greater security for classified projects than an urban university campus.

    Lawrence tapped 32-year-old Herbert York, a former graduate student of his, to run Livermore. Under York, the Lab had four main programs: Project Sherwood (the magnetic-fusion program), Project Whitney (the weapons-design program), diagnostic weapon experiments (both for the Los Alamos and Livermore laboratories), and a basic physics program. York and the new lab embraced the Lawrence “big science” approach, tackling challenging projects with physicists, chemists, engineers, and computational scientists working together in multidisciplinary teams. Lawrence died in August 1958 and shortly after, the university’s board of regents named both laboratories for him, as the Lawrence Radiation Laboratory.

    Historically, the Berkeley and Livermore laboratories have had very close relationships on research projects, business operations, and staff. The Livermore Lab was established initially as a branch of the Berkeley laboratory. The Livermore lab was not officially severed administratively from the Berkeley lab until 1971. To this day, in official planning documents and records, Lawrence Berkeley National Laboratory is designated as Site 100, Lawrence Livermore National Lab as Site 200, and LLNL’s remote test location as Site 300.[3]

    The laboratory was renamed Lawrence Livermore Laboratory (LLL) in 1971. On October 1, 2007 LLNS assumed management of LLNL from the University of California, which had exclusively managed and operated the Laboratory since its inception 55 years before. The laboratory was honored in 2012 by having the synthetic chemical element livermorium named after it. The LLNS takeover of the laboratory has been controversial. In May 2013, an Alameda County jury awarded over $2.7 million to five former laboratory employees who were among 430 employees LLNS laid off during 2008.[4] The jury found that LLNS breached a contractual obligation to terminate the employees only for “reasonable cause.”[5] The five plaintiffs also have pending age discrimination claims against LLNS, which will be heard by a different jury in a separate trial.[6] There are 125 co-plaintiffs awaiting trial on similar claims against LLNS.[7] The May 2008 layoff was the first layoff at the laboratory in nearly 40 years.[6]

    On March 14, 2011, the City of Livermore officially expanded the city’s boundaries to annex LLNL and move it within the city limits. The unanimous vote by the Livermore city council expanded Livermore’s southeastern boundaries to cover 15 land parcels covering 1,057 acres (4.28 km2) that comprise the LLNL site. The site was formerly an unincorporated area of Alameda County. The LLNL campus continues to be owned by the federal government.

    LLNL/NIF


    DOE Seal
    NNSA

     
  • richardmitnick 11:27 am on June 15, 2019 Permalink | Reply
    Tags: , , , Japan's new supercomputer Fugaku, Kyodo News, Supercomputing   

    From insideHPC via Kyodo News: “Japan’s new supercomputer Fugaku to begin operations around 2021” 

    From insideHPC

    via

    Kyodo News

    1

    Fugaku supercomputer Japan

    Japan’s new supercomputer “Fugaku” is set to begin operations around 2021 with the aim of regaining the title of the world’s fastest computer, replacing the current supercomputer “K,” government-backed research institute Riken said Thursday.

    The Fugaku, a nickname for Mt. Fuji, aims to be about 40 to 120 times faster than the K, the first supercomputer in the world to achieve a speed of over 10 quadrillion computations per second.

    “A supercomputer is essential to solving social challenges such as drug development and disaster prevention,” Riken President Hiroshi Matsumoto said. “We will dedicate our best effort to its success and operation.”

    The new supercomputer, developed at a cost of about 110 billion yen ($999 million), will be utilized in a wide range of research by various companies and universities including forecasting heavy rains.

    The institute received nearly 5,100 entries for potential names between February and April from the public, with only two entries for Fugaku.

    The new computer will be placed in the institute’s Center for Computational Science in Kobe, replacing the K when it retires in August.

    “We’re aiming for the world’s fastest computing speed,” Satoshi Matsuoka, head of the center, said.

    The United States and China have both revealed plans to release supercomputers with equal computing ability to the Fugaku in 2020 or 2021.

    In June 2011, the K, which refers to the way 10 quadrillion is written in Japanese, ranked first in the world in computing speed, going into full-scale operation in September 2012. It fell to 18th place in the most recent rankings reported in November last year.

    According to the “Top 500” list of the world’s fastest supercomputers made biannually by researchers, the United States ranked top with its supercomputer “Summit” in November 2018. China had held first place from 2013 until June 2018, when the United States regained top spot.

    See the full article here .

    five-ways-keep-your-child-safe-school-shootings

    Please help promote STEM in your local schools.

    Stem Education Coalition

    Founded on December 28, 2006, insideHPC is a blog that distills news and events in the world of HPC and presents them in bite-sized nuggets of helpfulness as a resource for supercomputing professionals. As one reader said, we’re sifting through all the news so you don’t have to!

    If you would like to contact me with suggestions, comments, corrections, errors or new company announcements, please send me an email at rich@insidehpc.com. Or you can send me mail at:

    insideHPC
    2825 NW Upshur
    Suite G
    Portland, OR 97239

    Phone: (503) 877-5048

     
  • richardmitnick 11:13 am on June 14, 2019 Permalink | Reply
    Tags: "Preparing Scientific Applications for Exascale Computing", , , , Brookhaven Lab's Computational Science Initiative hosted a four-day coding workshop focusing on the latest version of OpenMP, Supercomputing   

    From Brookhaven National Lab: “Preparing Scientific Applications for Exascale Computing” 

    From Brookhaven National Lab

    June 11, 2019
    Ariana Tantillo
    atantillo@bnl.gov

    Brookhaven Lab’s Computational Science Initiative hosted a four-day coding workshop focusing on the latest version of OpenMP, a widely used programming standard that is being upgraded with new features to support next-generation supercomputing.

    1
    The 2019 OpenMP hackathon at Brookhaven Lab—hosted by the Computational Science Initiative from April 29 to May 2—brought together participants from Brookhaven, Argonne, Lawrence Berkeley, Lawrence Livermore, and Oak Ridge national labs; IBM; NASA; Georgia Tech; Indiana University; Rice University; and University of Illinois at Urbana-Champaign.

    Exascale computers are soon expected to debut, including Frontier at the U.S. Department of Energy’s (DOE) Oak Ridge Leadership Computing Facility (OLCF) and Aurora at the Argonne Leadership Computing Facility (ALCF), both DOE Office of Science User Facilities, in 2021.

    ORNL Cray Frontier Shasta based Exascale supercomputer with Slingshot interconnect featuring high-performance AMD EPYC CPU and AMD Radeon Instinct GPU technology

    Depiction of ANL ALCF Cray Intel SC18 Shasta Aurora exascale supercomputer

    These next-generation computing systems are projected to surpass the speed of today’s most powerful supercomputers by five to 10 times. This performance boost will enable scientists to tackle problems that are otherwise unsolvable in terms of their complexity and computation time.

    But reaching such a high level of performance will require software adaptations. For example, OpenMP—the standard application programming interfaces for shared-memory parallel computing, or the use of multiple processors to complete a task—will have to evolve to support the layering of different memories, hardware accelerators such as graphics processing units (GPUs), various exascale computing architectures, and the latest standards for C++ and other programming languages.

    3
    Exascale computers will be used to solve problems in a wide range of scientific applications, including to simulate the lifetime operations of small modular nuclear reactors (left) and to understand the complex relationship between 3-D printing processes and material properties (right). Credit: Oak Ridge National Lab.

    Evolving OpenMP toward exascale with the SOLLVE project

    In September 2016, the DOE Exascale Computing Project (ECP) funded a software development project called SOLLVE (for Scaling OpenMP via Low-Level Virtual Machine for Exascale Performance and Portability) to help with this transition.

    The SOLLVE project team—led by DOE’s Brookhaven National Laboratory and consisting of collaborators from DOE’s Argonne, Lawrence Livermore, and Oak Ridge National Labs, and Georgia Tech—has been designing, implementing, and standardizing key OpenMP functionalities that ECP application developers have identified as important.

    Driven by SOLLVE and sponsored by ECP, Brookhaven Lab’s Computational Science Initiative (CSI) hosted a four-day OpenMP hackathon from April 29 to May 2, jointly organized with Oak Ridge and IBM. The OpenMP hackathon is the latest in a series of hackathons offered by CSI, including those focusing on NVIDIA GPUs and Intel Xeon Phi many-core processors.

    “OpenMP is undergoing substantial changes to address the requirements of upcoming exascale computing systems,” said local event coordinator Martin Kong, a computational scientist in CSI’s Computer Science and Mathematics Group and the Brookhaven Lab representative on the OpenMP Architecture Review Board, which oversees the OpenMP standard specification. “Porting scientific codes to the new exascale hardware and architectures will be a grand challenge. The main motivation of this hackathon is application engagement—to interact more deeply with different users, especially those from DOE labs, and make them aware of the changes they should expect in OpenMP and how these changes can benefit their scientific applications.”

    Laying the foundation for application performance portability

    ORNL IBM AC922 SUMMIT supercomputer, No.1 on the TOP500. Credit: Carlos Jones, Oak Ridge National Laboratory/U.S. Dept. of Energy

    Computational and domain scientists, code developers, and computing hardware experts from Brookhaven, Argonne, Lawrence Berkeley, Lawrence Livermore, Oak Ridge, Georgia Tech, Indiana University, Rice University, University of Illinois at Urbana-Champaign, IBM, and the National Aeronautics and Space Administration (NASA) participated in the hackathon. The eight teams were guided by national lab, university, and industry mentors who were selected based on their extensive experience in programming GPUs, participating in the OpenMP Language Committee, and conducting research and development in tools that support the latest OpenMP specifications.

    Throughout the week, the teams worked on porting their scientific applications from central processing units (CPU) to GPUs and optimizing them using the latest OpenMP version (4.5+). In between hacking sessions, the teams had tutorials on various advanced OpenMP features, including accelerator programming, profiling tools to assess performance, and application optimization strategies.

    Some teams also used the latest OpenMP functionalities to program IBM Power9 CPUs accelerated with NDIVIA GPUs. The world’s fastest supercomputer—the Summit supercomputer at OLCF—is based on this new architecture, with more than 9000 IBM Power9 CPUs and more than 27,000 NVIDIA GPUs.

    Taking steps toward exascale

    The teams’ applications spanned many areas, including nuclear and high-energy physics, lasers and optics, materials science, autonomous systems, and fluid mechanics.

    Participant David Wagner of the NASA Langley Research Center High Performance Computing Incubator and colleagues Gabriele Jost and Daniel Kokron of the NASA Ames Research Center came with a code for simulating elasticity. Their goal at the hackathon was to increase single-instruction, multiple-data (SIMD) parallelism—a type of computing in which multiple processors perform the same operation on many data points simultaneously—and optimize the speed at which data can be read from and stored into memory.

    “Scientists at NASA are trying to understand how and why aircraft and spacecraft materials fail,” said Wagner. “We need to make sure that these materials are durable enough to withstand all of the forces that are present in normal use during service. At the hackathon, we’re working on a mini app that is representative of the most computationally intensive parts of the larger program to model what happens physically when the materials are loaded, bent, and stretched. Our code has lots of little formulas that need to run billions of times over. The challenge is performing all of the calculations really fast.”

    According to Wagner, one of the reasons NASA is pushing for this computational capability now is to understand the processes used to generate additively manufactured (3-D printed) parts and the different material properties of these parts, which are increasingly being used in aircraft. Knowing this information is important to ensuring the safety, reliability, and durability of the materials over their operational lifetimes.

    “The hackathon was a success for us,” said Wagner. “We got our code set up for massively parallel execution and running correctly on GPU hardware. We’ll continue with debugging and parallel performance tuning, as we expect to have suitable NASA hardware and software available soon.”

    Another team took a similar approach in trying to get OpenMP to work for a small portion of their code, a lattice quantum chromodynamics (QCD) code that is at the center of an ECP project called Lattice QCD: Lattice Quantum Chromodynamics for Exascale. Lattice QCD is a numerical framework for simulating the strong interactions between elementary particles called quarks and gluons. Such simulations are important to many high-energy and nuclear physics problems. Typical simulations require months of running on supercomputers.

    4
    A schematic of the lattice for quantum chromodynamics calculations. The intersection points on the grid represent quark values, while the lines between them represent gluon values.

    “We would like our code to run on different exascale architectures,” said team member and computational scientist Meifeng Lin, deputy group lead of CSI’s new Quantum Computing Group and local coordinator of previous hackathons. “Right now, the code runs on NVIDIA GPUs but upcoming exascale computers are expected to have at least two different architectures. We hope that by using OpenMP, which is supported by major hardware vendors, we will be able to more easily port our code to these emerging platforms. We spent the first two days of the hackathon trying to get OpenMP to offload code from CPU to GPU across the entire library, without much success.”

    Mentor Lingda Li, a CSI research associate and a member of the SOLLVE project, helped Lin and fellow team member Chulwoo Jung, a physicist in Brookhaven’s High-Energy Theory Group, with the OpenMP offloading.

    Though the team was able to get OpenMP to work with a few hundred lines of code, its initial performance was poor. They used various performance profiling tools to determine what was causing the slowdown. With this information, they were able to make foundational progress in their overall optimization strategy, including solving problems related to initial GPU offloading and simplifying data mapping.

    Among the profiling tools available to teams at the hackathon was one developed by Rice University and University of Wisconsin.

    5
    John Mellor-Crummey gives a presentation about the HPCToolkit, an integrated suite of tools for measuring and analyzing program performance on systems ranging from desktops to supercomputers.

    “Our tool measures the performance of GPU-accelerated codes both on the host and the GPU,” said John Mellor-Crummey, professor of computer science and electrical and computer engineering at Rice University and the principal investigator on the corresponding ECP project Extending HPCToolkit to Measure and Analyze Code Performance on Exascale Platforms. “We’ve been using it on several simulation codes this week to look at the relative performance of computation and data movement in and out of GPUs. We can tell not only how long a code is running but also how many instructions were executed and whether the execution was at full speed or stalled, and if stalled, why. We also identified mapping problems with the compiler information that associates machine code and source code.”

    Other mentors from IBM were on hand to show the teams how to use IBM XL compilers—which are designed to exploit the full power of IBM Power processors—and help them through any issues they encountered.

    “Compilers are tools that scientists use to translate their scientific software into code that can be read by hardware, by the largest supercomputers in the world—Summit and Sierra [at Lawrence Livermore],” said Doru Bercea, a research staff member in the Advanced Compiler Technologies Group at the IBM TJ Watson Research Center. “The hackathon provides us with an opportunity to discuss compiler design decisions to get OpenMP to work better for scientists.”

    According to mentor Johannes Doerfert, a postdoctoral scholar at ALCF, the applications the teams brought to the hackathon were at various stages in terms of their readiness for upcoming computing systems.

    6
    QMCPack can be used to calculate the ground and excited state energies of localized defects in insulators and semiconductors—for example, in manganese (Mn)4+-doped phosphors, which are promising materials for improving the color quality and luminosity of white-light-emitting diodes. Source: Journal of Physical Chemistry Review Letters.

    “Some teams are facing porting problems, some are struggling with the compilers, and some have application performance issues,” explained Doerfert. “As mentors, we receive questions coming from anywhere in this large spectrum.”

    Some of the other scientific applications that teams brought include a code (pf3d) for simulating the interactions between high-intensity lasers and plasma (ionized gas) in experiments at Lawrence Livermore’s National Ignition Facility, and a code for calculating the electronic structure of atoms, molecules, and solids (QMCPack, also an ECP project). Another ECP team brought a portable programming environment (RAJA) for the C++ programming language.

    “We’re developing a high-level abstraction called RAJA so people can use whatever hardware or software frameworks are available on the backend of their computer systems,” said mentor Tom Scogland, a postdoctoral scholar in the Center for Applied Scientific Computing at Lawrence Livermore. “RAJA mainly targets OpenMP on the host and CUDA [another parallel computing programming model] on the backend. But we want RAJA to work with other programming models on the backend, including OpenMP.”

    “The theme of the hackathon was OpenMP 4.5+, an evolving and not fully mature version,” explained Kong. “The teams left with a better understanding of the new OpenMP features, knowledge about the new tools that are becoming available on Summit, and a roadmap to follow in the long term.”

    “I learned a number of things about OpenMP 4.5,” said pf3d team member Steve Langer, a computational physicist at Lawrence Livermore. “The biggest benefit was the discussions with mentors and IBM employees. I now know how to package my OpenMP offload directives to use NVIDIA GPUs without running into memory limitations.”

    A second OpenMP hackathon will be held in July at Oak Ridge and a third in August at the National Energy Research Scientific Computing Center, a division of Lawrence Berkeley, a DOE Office of Science User Facility, and the primary computing facility for DOE Office of Science–supported researchers.

    See the full article here .


    five-ways-keep-your-child-safe-school-shootings

    Please help promote STEM in your local schools.

    Stem Education Coalition

    BNL Campus


    BNL Center for Functional Nanomaterials

    BNL NSLS-II


    BNL NSLS II

    BNL RHIC Campus

    BNL/RHIC Star Detector

    BNL RHIC PHENIX

    One of ten national laboratories overseen and primarily funded by the Office of Science of the U.S. Department of Energy (DOE), Brookhaven National Laboratory conducts research in the physical, biomedical, and environmental sciences, as well as in energy technologies and national security. Brookhaven Lab also builds and operates major scientific facilities available to university, industry and government researchers. The Laboratory’s almost 3,000 scientists, engineers, and support staff are joined each year by more than 5,000 visiting researchers from around the world. Brookhaven is operated and managed for DOE’s Office of Science by Brookhaven Science Associates, a limited-liability company founded by Stony Brook University, the largest academic user of Laboratory facilities, and Battelle, a nonprofit, applied science and technology organization.
    i1

     
c
Compose new post
j
Next post/Next comment
k
Previous post/Previous comment
r
Reply
e
Edit
o
Show/Hide comments
t
Go to top
l
Go to login
h
Show/Hide help
shift + esc
Cancel
%d bloggers like this: