Last week at SC15, Numascale announced the successful installation of a large shared memory Numascale/Supermicro/AMD system at a customer datacenter facility in North America. The system is the first part of a large cloud computing facility for analytics and simulation of sensor data combined with historical data. “The Numascale system, installed over the last two weeks, consists of 108 Supermicro 1U servers connected in a 3D torus with NumaConnect, using three cabinets with 36 servers apiece in a 6x6x3 topology. Each server has 48 cores in three AMD Opteron 6386 CPUs and 192 GBytes memory, providing a single system image and 20.7 TBytes to all 5184 cores. The system was designed to meet user demand for “very large memory” hardware solutions running a standard single image Linux OS on commodity x86 based servers.”
In this podcast, Jorge Salazar from TACC interviews two winners of the 2015 ACM Gordon Bell Prize, Omar Ghattas and Johann Rudi of the Institute for Computational Engineering and Sciences, UT Austin. As part of the discussion, Ghattas describes how parallelism and exascale computing will propel science forward.
In this video from SC15, Rich Brueckner from insideHPC moderates a panel discussion with Hewlett Packard Enterprise HPC customers. “Government labs, as well as public and private universities worldwide, are using HPE Compute solutions to conduct research across scientific disciplines, develop new drugs, discover renewable energy sources and bring supercomputing to nontraditional users and research communities.”
Today Russia’s RSC Group announced that Team TUMuch Phun from the Technical University of Munich (TUM) won the Highest Linpack Award in the SC15 Student Cluster Competition. The enthusiastic students achieved 7.1 Teraflops on the Linpack benchmark using an RSC PetaStream cluster with computing nodes based on Intel Xeon Phi. TUM student team took 3rd place in overall competition within 9 teams participated in SCC at SC15, so as only one European representative in this challenge.
Basic optimization techniques that include an understanding of math functions and how to simplify can go a long way towards better performance. “When optimizing for a parallel SIMD system such as the Intel Xeon Phi coprocessor, it is also important to make sure that the results match the scalar system. Using vector data may cause parts of the computer program to be re-written, so that the compiler can generate vector code.”
Software for data analysis, system management, and for debugging other software were be among the innovations on display at SC15 last week. In addition to the software, novel and improved hardware will also be on display, together with an impressive array of initiatives from Europe in research and development leading up to Exascale computing.
In this video from SC15, Sam Mahalingam from Altair discusses the HyperWorks Unlimited Virtual Appliance and the new open source version of PBS Pro. “Our goal is for the open source community to actively participate in shaping the future of PBS Professional driving both innovation and agility. The community’s contributions combined with Altair’s continued research and development, and collaboration with Intel and our HPC technology partners will accelerate the advancement of PBS Pro to aggressively pursue exascale computing initiatives in broad classes and domains.”
Last week at SC15, Fujifilm announced that its next-generation LTO Ultrium 7 data cartridge has been qualified by the LTO technology provider companies for commercial production and is available immediately. FUJIFILM LTO Ultrium 7 has a compressed storage capacity of 15.0TB with a transfer rate of 750MB/sec assuming 2.5:1 compression ratio. This capacity achievement represents a 2.4X increase over the current LTO-6 generation.