Intel® OPA is interconnect of choice for six fastest academic clusters in world.
The evidence is in. Powering some of the fastest supercomputers in the world, Intel® Omni-Path Architecture (Intel® OPA) strongly increased our contribution interconnecting TOP500 systems, with a record number of total PFLOPS increased to 146.2.1 This represents a 26% increase since June 2018.2
Continuing Intel® OPA’s deployment momentum at leading research institutions, Intel® OPA is the interconnect of choice for the six fastest academic clusters in the world, as noted on TOP500, with Leibniz Supercomputing Centre (also known as Leibniz Rechenzentrum or LRZ) now joining the TOP500 list at number 8.
To see Intel® OPA in action among the TOP500, check out highlights from the Intel® HPC Forum, held in Dallas, Texas during Supercomputing 2018 (SC18). Guest presenter, Professor Dr. Dieter KranzlmÜller, from LRZ in Bavaria, Germany, shared LRZ research plans using their newly installed SuperMUC-Next Generation (NG) with peak performance rated at 26.7 PFLOPs using 6,480 Intel® Xeon® Scalable processors and the Intel® OPA interconnect.
In his presentation, “The quest for high performance—supporting science with SuperMUC-Next Generation at LRZ,” Dr. KranzlmÜller described how his team helps drive scientific innovation, big data analytics, and cloud file system sharing among local academic institutions, as well as biotech, IT, and environmental firms.
Also at the Intel HPC Forum, Dr. Sofia Vallecorsa shared her research at CERN OpenLab, exploring the possibility of using deep-learning methods for fast simulation, in contrast to classical Monte Carlo simulations. In her presentation, Dr. Vallecorsa shared results comparing theories, accuracy and performance, conducted on University of Texas Advanced Computing Center (TACC) Stampede2, which is connected with Intel OPA and is number 17 on the TOP500. Dr. Vallecorsa described the application of Generative Adversarial Networks, a form of deep neural network (DNN), showing potential to dramatically increase the speed of high energy physics (HEP) simulations in preparation for a future upgrade to CERN’s Large Hadron Collider (LHC), the world’s largest particle accelerator.
Finally, at number 19 on the TOP500, the Marconi supercomputer was recently described by Dr. Marco Rorro, HPC CTO at Cineca, and Dr. Valeriu Codreanu, Senior HPC Consultant at SURFsara. In their webinar, “TensorFlow*: Performance and Optimization on the Marconi Supercomputer,” they shared step by step tutorial insights on how to setup, optimize, and improve performance of TensorFlow workloads at HPC scale, as well as insights on how Intel® OPA provides the fabric performance and efficiency required to scale connectivity to thousands of Intel® Xeon® Scalable processors.
Beyond the TOP500, Intel® OPA has been deployed across hundreds of accounts, including customers in government, academic research, and commercial enterprise, with some customers deploying their 2nd and 3rd round of cluster expansion focused on HPC and artificial intelligence (AI) workloads. Additionally, Intel® OPA has a large and growing ecosystem consisting of hardware vendors, storage vendors along with support on both open source and commercially available applications.
We are excited about Intel® OPA’s continued growth in 2019 as we help enable optimal HPC, AI, and deep learning workload performance on our customers’ systems. For more information on how Intel® OPA can help your institution or company deliver its next scientific breakthrough, please visit intel.com/omnipath.
Software and workloads used in performance tests may have been optimized for performance only on Intel® microprocessors. Performance tests, such as SYSmark* and MobileMark*, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to www.intel.com/benchmarks.
Performance results are based on testing as of 11/12/2018 and may not reflect all publicly available security updates. See configuration disclosure for details. No product or component can be absolutely secure. Intel® technologies’ features and benefits depend on system configuration and may require enabled hardware, software or service activation. Performance varies depending on system configuration. Check with your system manufacturer or retailer or learn more at www.intel.com.
2 Based on FLOPS for Intel OPA systems ranked in the top 500 of the November 2018 TOP500 List https://www.top500.org/lists/2018/11/ and the June 2018 TOP500 List https://www.top500.org/lists/2018/06/.