BrandPosts are written and edited by members of our sponsor community. BrandPosts create an opportunity for an individual sponsor to provide insight and commentary from their point-of-view directly to our audience. The editorial team does not participate in the writing or editing of BrandPosts.
The 5th meeting of the Dell EMC HPC Community (DellHPC.org) was held on June 21-22, 2017, in Frankfurt, Germany, during the week of the ISC (International Supercomputing) High Performance 2017 conference and exhibition (isc-hpc.com). The Dell EMC HPC Community is a Dell EMC – sponsored forum designed to foster the exchange of HPC strategies, technologies and best practices among Dell EMC executives and staff, our technology partners, and most importantly, our HPC customers.
For this Community meeting, we continued the discussion started at the last Community meeting in April ’17 in Austin on the evolution of HPC from its traditional roots in modeling and simulation, to the much broader areas of data analytics and artificial intelligence, which was also one of the themes of the ISC event. HPC provides the infrastructure for artificial intelligence, so the explosion of interest in it in HPC circles should not be a surprise. And no less than Michael Dell has commented on the impart AI will have in the coming years:
“When you get all this data coming in from hundreds of billions of connected devices and apply to that artificial intelligence, it’s almost like a fourth industrial revolution and an incredible opportunity for companies to re-imagine themselves in this digital age….The last 30 years have been incredible in IT, but the next 30 years will make it look like child’s play.”
Michael DellSapphire SAP Conference May 16th 2017
In his keynote at the Community meeting, Armughan Ahmad, SVP & GM, Ready Solutions and Alliances, Dell EMC, spoke about the company’s history as a leader in HPC and how our customers workloads are evolving to be increasingly data centric, with examples in life sciences (genomics) and financial services (risk analysis), using machine/deep learning technologies to discover patterns in the data, and make decisions based on that. This is leading to new generation of HPC systems designed for huge volumes of data, processed via machine and deep learning technologies.
Armughan presenting to the Community meeting.
With the increased focus on data centric HPC, systems need to be architected differently. In the first customer keynote, Paul Calleja, Director of Research Computing at the University of Cambridge in the UK, described how the Cambridge Research Computing Service has implemented a new system aimed at addressing large scale High I/O data intensive science use cases, such as their research programs in life sciences, high energy physics and astronomy, including support of the Square Kilometer Array project. The system includes two new TOP500 Dell EMC HPC Systems, Wilkes-2 at #100 based on PowerEdge C4130 servers with NVIDIA Tesla P100 GPUs and Peta4-KNL at #404 based on PowerEdge C6320p servers with Intel Xeon Phi (KNL) processors. (Per the June 2017 TOP500 list, www.top500.org)
Then, in a session focused on Machine/Deep Learning and Artificial Intelligence, experts from Dell EMC, Intel, NVIDIA, the University of Pisa in Italy, and The Commonwealth Scientific and Industrial Research Organization (CSIRO) of Australia discussed Machine/Deep Learning technologies and use cases. As a very new technology, machine/deep learning has many tool sets for developing applications, many system technologies to build them on, and only an emerging understanding of which combinations of these are optimal for any particular workload.
Jay Boisseau, HPC Technology Strategist at Dell EMC, presented on Dell EMC’s Machine Learning Strategy. The talk focused on Dell EMC’s views on machine learning’s potential for impact in areas such as healthcare, research, financial services, and more, and outlined our plans for providing effective solutions for various workloads and verticals.
Joe Curley, Sr. Director HPC Platforms and Ecosystems at Intel, talked about Intel’s flexible portfolio of technologies offered by Dell EMC, to enable technologists to build more advanced AI solutions and stimulate new ideas.
Jonas Lööf, Deep Learning Solution Architect at NVIDIA, presented an introduction to what deep learning is, why it is becoming more and more important, and how it is in use in different fields. The importance of GPU computation was discussed, including an overview of different GPU solutions for deep learning.
Antonio Cisternino, CIO of the University of Pisa, shared present and future Deep and Machine Learning Activities at the University of Pisa. Researchers at the university have been evaluating different algorithms and technologies for various ML/DL/AI workloads, and have concluded that there is much work yet to be done to determine the optimal tools and system architectures that are optimal for each.
Angus Macoustra, Deputy Director, Information Management & Technology at CSIRO discussed how Australia’s National science agency, the Commonwealth Scientific and Industrial Research Organization (CSIRO) is applying machine learning with scientific computing to tackle the big challenges facing the nation.
One feature of the HPC Community meetings is the ample time built in for networking among the participants. Thus, the Community meeting started on the evening of the 21st with a networking reception and dinner “on the beach” at the Frankfurter Botschaft in Frankfurt. Yes, a beach in Frankfurt. With over 100 attendees, and perfect weather, we had a great time networking, enjoying German BBQ, and getting sand in our dress shoes.
On the beach, with Community members Happy Sithole from CHPC in South Africa, Maurizo Davini and Antonio Cisternino of the University of Pisa in Italy, Tom Tabor from Tabor Communications, Ed Turkel of Dell EMC, and members of the Dell EMC-sponsored CHPC Student Cluster Challenge team, who placed 2nd in this year’s competition.
The presentations for all Community talks are available to Community members at DellHPC.org. Membership is open to all Dell EMC customers, technology partners and staff, and requires that the member has an active non-disclosure agreement with Dell EMC.
The next meeting of the Community will be at the Accelerating Understanding forum at the University of Pisa in Italy, September 26 and 27.
See the website for more information on how to join and register for future meetings.