ISC 2017
BittWare
BittWare’s (A-1431) FPGA boards combine the largest and most powerful FPGAs from Intel and Xilinx with high-end networking interfaces and memory options. The company will demonstrate an artificial intelligence/machine learning application on the A10PL4 PCIe board—programmed using OpenCL for rapid development and easy upgrades as machine learning libraries improve. The board features an Intel Arria 10 FPGA, as does the upcoming A10SA4—the first board to use BittWare’s exclusive Spider platform for top performance.
BittWare’s Xilinx boards include the high-powered XUPP3R, featuring an UltraScale+ 16nm FPGA. This board will be used to demonstrate our latest IP core: Arkville DPDK from Atomic Rules. This IP core brings the industry-standard DPDK to allow high-performance FPGA-CPU data moving with kernel bypass. BittWare’s Xilinx UltraScale+ boards also feature 100G networking and up to 256 Gbytes of DDR4 memory.
DDN
At ISC17, DDN (E-1010) will present live demonstrations of the latest in HPC data management solutions including Flash-Native, Scale-Out caching with Infinite Memory Engine accelerating I/O for Lustre parallel file systems. DDN will also be showcasing GRIDScaler parallel file system appliance use cases for industrial analytics and machine learning and WOS object storage for cloud and active archive applications.
At the DDN User Group Tuesday, June 20, leading HPC sites will present their experience supporting some of the most data-intensive workflows in the world. Talks will include the results from test-bed and production sites working with the latest DDN products and industry technologies including Lustre 2.7, OPA and NVMe-based burst buffers.
Meet with DDN at booth E-1010 or join our user group Tuesday June 20. User Group attendees also receive tickets to the DDN Customer Appreciation Party.
DKRZ
At ISC’17 DKRZ (L-440) presents its mission as a national HPC center - DKRZ provides high performance computers, high capacity data storage and management for German climate research. In Addition DKRZ’s specialists offer a range of services covering modeling and programming, from support in carrying out climate simulations, through parallelisation and optimisation of the simulation programs, to analysis and visualization of the simulation results.
DKRZ is involved in numerous projects. As an example it informs about the new Centre of Excellence in Simulation of Weather and Climate in Europe (ESiWACE), which is coordinated by DKRZ. DKRZ computers are regularly used for the extensive simulations that contribute to the IPCC reports. On a climate globe DKRZ presents climate simulations.
During the conference DKRZ employees contribute to the program as organisers, chairs and/or speakers to keynotes, workshops and BOFs.
The DKRZ research group "Scientific Computing" participates in the student cluster competition.
EPCC
EPCC, (B-1223) the supercomputing centre at the University of Edinburgh, is home to many of the UK’s national HPC and research data services, including ARCHER and the Research Data Facility. We’re also home to Cirrus, our new Tier 2 HPC service for science and industry – one of our key focuses. Together, our facilities and expertise make us the premiere industry HPC provider in Europe.
Come and see us at ISC to learn more about:
The Fortissimo [1] project, helping the European manufacturing industry access on-demand HPC resources;
EPCC involvement in the Centres of Excellence initiative via BioExcel [2] and CompBioMed [3], enabling use of HPC and HTC in life sciences;
The push towards the Exascale goal through FETHPC projects, including leadership of NEXTGenIO [4] and INTERTWinE [5].
EPCC’s commitment to increasing diversity through founding the Women in HPC [6] network.
[2] www.bioexcel.eu
[4] www.nextgenio.eu
Nvidia
AI is the future and NVIDIA (H-730) is uniquely suited to drive it forward. NVIDIA has been at the forefront of deep learning, a key area of AI, for years because it believes it will revolutionise computing and transform HPC. At ISC, NVIDIA( booth #730) will demonstrate how the revolution has already begun, how the intersection of AI and HPC is helping accelerate the world’s most challenging scientific research. NVIDIA will showcase its new Volta-architecture and its newly expanded portfolio of purpose-built AI supercomputers, including the Volta-based NVIDIA DGX-1 and the all-new personal AI workstation, the NVIDIA DGX Station.
On Tuesday, June 20, Steve Oberlin will present the “Dawn of Smart Supercomputing: How HPC & AI Are Engaging in a Virtuous Cycle That Will Leave Both Forever Transformed & Immeasurably Improved” in room Panorama 1.
PRACE
PRACE (B-1201) will highlight results from its HPC systems, and promote the workshops organised by PRACE and partners. Visitors are invited to listen to a speech by the PRACE Managing Director, Serge Bogaerts, and receive a special gift. The PRACE Fact Sheets will be available, as well as the PRACE Annual Report 2016. The PRACE booth will feature again the popular short but in-depth presentations on the PRACE Peer Review process, PRACE Training, and PRACE for Industry. The winner of the PRACE-ISC Research Poster Award will be announced during the Research Poster Session and Reception.
PRACE, the Partnership for Advanced Computing in Europe is a persistent pan-European Research Infrastructure of world-class supercomputers in all major architectural classes. PRACE powers excellent science and engineering in academia and industry, addressing society’s grand challenges and serving as a vital catalyst in fostering European competitiveness and economic growth.
Univa
Univa (C-1214) Grid Engine is the leading workload management system. The solution maximises the use of shared resources in a data centre and applies advanced policy management tools to deliver products and results faster, more efficiently, and with lower overall costs. The product can be deployed in any technology environment, including containers: on-premise, hybrid or in the cloud. A variety of add-ons can be utilised to extend workload management capabilities and create a customised solution for any enterprise infrastructure.
