At ISC, Intel® Technologies Unleash the Intertwined Future of AI and HPC

Intel is excited to be a part of ISC High Performance 2018 in Frankfurt, Germany, June 25–28. ISC brings together academic and commercial disciplines to share knowledge in the field of high performance computing. Our presence at the event will include keynotes, sessions, and booth demos that will be focused on the future of HPC technology, including Artificial Intelligence (AI) and visualization. Also featured in the Intel booth: The Intel Collaboration Hub hosting eight in-booth panel discussions and 25 Intel and partner presentations.

Throughout Intel’s presence at ISC, we will showcase how the industry and the world are entering an advanced era of HPC that now includes AI technologies. The data and algorithmic innovations powered with deep learning have led to a sharp increase in the need for compute. As deep learning and other AI techniques evolve with massive data sets and image sizes, they will continue to require high compute power that scales to the size of HPC infrastructures. Intel will showcase how it’s helping AI developers, data scientists and HPC programmers transform industries such as healthcare, retail and finance by tapping into HPC to power the AI solutions of today and the future.

Let’s take a look at a few key sessions and demos presented by Intel. We’ll also discuss where and how to engage with Intel at the show.

Key Intel Sessions

Keynote: Supercomputing Circa 2025
Dr. Raj Hazra, Corporate Vice President at Intel, will discuss emerging technologies that will accelerate discovery and innovation. The convergence of AI, analytics and simulation and modeling provide new opportunities and challenges for every organization and will require new system architecture for HPC. Dr. Hazra will explore new system level architecture approaches that will help global HPC leaders reach new breakthroughs in system performance and capabilities.

Speaker: Raj Hazra, Corporate Vice President at Intel
Date: Monday, June 25, 2018
Time: 6:00 to 6:45 p.m.
Location: Panorama 2, Forum

Keynote at the HP-CAST Event: New Era in HPC
Trish Damkroger, Vice President, Data Center Group, General Manager, Technical Computing Initiative, Enterprise and Government at Intel will keynote at the HP-CAST, HPE’s user group meeting for high performance computing on the New Era in HPC. She will discuss the changing landscape of technical computing, key trends in HPC, and the convergence of HPC-AI-HPDA that is transforming the industry and ensuring technical computing continues to fulfill its potential as a scientific tool for insight and innovation. Trish will also provide an overview of Intel’s technical computing strategy, products, and ecosystem collaborations that are driving this transformation.

Speaker: Trish Damkroger, Vice President, Data Center Group, General Manager, Technical Computing Initiative, Enterprise and Government at Intel
This keynote will take place at the HP-CAST event.

Machine Learning Day at ISC
Within the last few years, machine learning has become a vital topic within the HPC community. Vendors see the development as an opportunity to expand their high-performance solutions into a new and fast-growing market, and traditional HPC users are incorporating these techniques into their workflows to speed processing and extract greater insights. ISC High Performance will provide a special focus on this subject matter with a one-day track on Wednesday, June 27, where Pradeep Dubey, Intel Fellow at Intel Labs and the Director of the Parallel Computing Lab, will speak in the following session:

Session: On Scale-out Deep Learning Training for Cloud and HPC
Pradeep Dubey, Intel Fellow at Intel Labs and the Director of the Parallel Computing Lab will be one of the featured speakers for Machine Learning Day at ISC. He will discuss how the exponential growth of Artificial Intelligence (AI) and Deep Learning (DL) has accelerated the need for training deep neural networks in hours or even minutes. This can only be achieved through scalable and efficient distributed training, since a single node/card cannot satisfy the compute, memory, and I/O requirements of today's state-of-the-art neural networks. However, scaling Stochastic Gradient Descent (SGD) is still a challenging problem and requires continued research/development. This entails innovations spanning algorithms, frameworks, communication libraries, and system design. In this talk, Pradeep will describe the philosophy, design, and implementation of Intel Machine Learning Scalability Library (MLSL), support in popular DL frameworks, and present proof-points demonstrating scaling DL training on 100s to 1000s of nodes across Cloud and HPC systems.

Speaker: Pradeep Dubey, Intel Fellow at Intel Labs
Session: Scalable Machine Learning Systems
Time: Wednesday, June 27, 1:45 to 2:15 p.m.
Location: Panorama 2

Intel booth activities located in Messe Frankfurt, Hall 3 (upper level) F-930

The Intel Collaboration Hub, located inside the Intel booth, will be the place to get updates from Intel, meet experts, experience panel discussions, and see the latest from Intel. Below are two panel discussion highlights. Click the link below to see the full schedule:

Panel: Intel® Select Solutions – Intro to Professional Visualization
Hosts: Jeff Jeffers, Senior Director, Senior Principle Engineer, Visualization Engineering at Intel; Ullrich Becker-Lemgau, Technical Marketing Engineer, Intel; and Werner Krotz-Vogel, Technical Marketing Manager, Intel
Time: 11 a.m.
Location: Intel Collaboration Hub inside the Intel booth

Panel: Convergence of AI and HPC
Host: Binay Ackalloor, Director, Business Development, Artificial Intelligence & Machine Learning at Intel
Time: 12 p.m.
Location: Intel Collaboration Hub inside the Intel booth

Please see the full schedule of the Intel Collaboration Hub speakers and panels below.

Intel ISC Booth Demos

Demo: AI on IA—Brain Tumor Segmentation
See an interactive demo of a solution built by Intel to help identify brain tumors. This solution uses deep CNNs to segment brain tumors in 2D MRI images, using 2016 Brain and Tumor Segmentation (BraTS) dataset and containing 31,000 2D MRI scans at 240x240 pixel resolution. Each brain has 255 slices, and the dataset represents ~120 individuals. Intel will deliver a flexible deployment architecture that provides great performance and T.C.O for GE Imaging C.T. group to deliver deep learning applications for their imaging modalities. The result is a trained model that is able to segment brain tumors just as well as a neuroradiologist. Intel then passes along the final MRI images and segmentation maps to EchoPixel* for visualization.

Demo: Defined Visualization
See the latest Intel® Select Solution that uses software-defined visualization and state-of-the-art Intel ingredients for modeling with visualization of even the most complex and large data sizes. The demo will showcase automotive CFD Analysis with OpenFOAM*, ParaView v5.5* using OSPRay* to demonstrate high visual fidelity of software-defined visualization using ray tracing to display an animation of airflow over a moving vehicle.

Demo: Application Acceleration with FPGA
This demo is built around the Intel® Programmable Acceleration Card with Intel® Arria® 10 GX FPGA (Intel® PAC with Intel® Arria® 10 GX FPGA) and demonstrates how FPG-based processing provides high performance and low power. See how to dynamically configure an FPGA to run multiple functions or workloads using partial reconfiguration. The workloads run on the FPGA will be file compression, genome sequencing, image recognition, and video segmentation. The demo will use the Intel® PAC with Intel® Arria 10 GX FPGA, a robust collection of software, firmware, and tools to make it easier to use Intel® FPGAs for workload optimization. The Acceleration Stack abstracts the FPGA design, enabling software application developers to leverage the benefits of FPGAs.

Demo: Distributed Deep Reinforcement Learning on CPU
An Adam optimization algorithm with a batch size of up to 2048 is a viable choice for carrying out large scale machine learning computations. See how this optimization is used to train an Atari agent to play games, which resulted in substantially decreasing the training time. The results were collected using a multi-node Intel® Xeon® processor-based cluster (formerly codenamed Haswell) and the optimized TensorFlow* for Intel® architecture. We will demonstrate the progress of the training model, including detailed hardware configuration and software stack, as well as visualization of all critical parameters of the training algorithm and their progress over time.

At ISC, see how Intel is leading the way in the convergence of AI and HPC, and partnering with organizations to create solutions that are solving the problems of today and tomorrow. You can engage with us in various sessions (see the ISC schedule) and in the Intel booth (Messe Frankfurt, Hall 3 (upper level) F-930). You can also follow and engage with us on social media @IntelHPC and follow the #ISC18 hashtag.

The full Intel Collaboration Hub schedule (taking place in Messe Frankfurt, Hall 3 (upper level) F-930): 

MONDAY, June 25th
TIME Topic Host/Speakers Guest
13:00 VENDOR SHOWDOWN (in Panorama 3) Nash Palaniswamy
15:00 Intel Optane and NVM solutions for HPC Andrey Kudryavtsev, Intel
15:20 Towards a modular supercomputer architecture for HPC and Data Analytics Norbert Eicker, Morris Riedel, and Hans-Christian Hoppe, Juelich Supercomputing Center and Intel
15:40 BigDL: Distributed Deep Learning on Apache Spark* Michael Steyer, Intel



OpenMP Panel: API Version 5.0 is Knocking at your Door – Grill the ARB Mike Klemm, Intel Host  Christian Terboven, RWTH Aachen
Martin Schulz, Technical University of Munich
Barbara Chapman, Stony Brook University
Xavier Teruel, Barcelona Supercomputing Center
17:00 Accelerating modern physics simulations with FPGAs Tobias Kenter and Christian Plessl, Paderborn University
17:20 OpenHPC Deep Dive: What’s New in PBS Pro v18 Piush Patel and Bill Nitzberg, Altair
17:40 Exploiting HPC Technologies to Accelerate Big Data Processing and Associated Deep Learning Dhabaleswar Panda and Xiaoyi Lu, Ohio State University



KEYNOTE: Supercomputing: Redefined  (in Panorama 2)  Dr. Raj Hazra, Corporate Vice President at Intel
19:00 Building a Balanced Artificial Intelligence System Pamela Sufi,
Penguin Computing
19:15 Hybrid HPC Jean-Luc Assor, Lori Ingle HPE
19:30 MEGWARE HPC Clusters, an Energy Efficient Intel® Select Solution for Simulation and Modeling Axel Auweter,
MEGWARE Computer Vertrieb und Service GmbH
19:45 The Power of Neptune: How to Increase Performance & Reduce Energy Consumption Vinod Kamath, Patrick Moakley and Lidia Wojdat,
TUESDAY, June 26th - Industrial Day
TIME Topic Host/Speakers Guest
09:40 EXHIBIT HALL OPENS 10:00 a.m.
10:00 Fast and Accurate Training of an AI Radiologist on Intel Xeon-based Dell EMC Supercomputers Lucas Wilson, Alex Filby, Nishanth Dandapanthula, Onur Celebioglu, Valeriu Codreanu and Damian Podareanu,
Dell/EMC, SURFsara
10:20 Moving to Skylake and Beyond with the latest Intel® Compilers v19.0 Keith Perrin, MSC Software
10:40 Recent advances in OSPRay rendering in ParaView David Demarle,
Michael Migliore
and Joachim Pouderoux,
Kitware SAS



PANEL: Intel Select Solutions - Intro to Professional Visualization Jim Jeffers,
Dan Stanzione , TACC, TX Luis Paulo Santos, Univ of Porto, Portugal Francesco Ruffino, Amazon  Michael Migliore, Kitware  Ullrich Becker-Lemgau and Werner Krotz-Vogel, Intel



PANEL: Convergence of AI and HPC Binay Ackalloor,
Prof. David Keyes, KAUST
Vikram Saletore, Intel
Valeriu Codreanu, SURFsara B.V.
Gaurav Kaul, Amazon
Sofia Vallecorsa, CERN
13:00 3D Convolutional GANs for fast detector simulation Sofia Vallecorsa, Federico Carminati, and Gulrukh Khattak, CERN
13:20 Advances Made in ANSYS Fluent® on Intel® Omni-Path Architecture and Latest Intel® Processors Wim Slagter,  ANSYS®
13:40 The Exascale Deep Learning and Simulation Enabled Precision Medicine for Cancer Project Thomas Brettin, Rick Stevens, and Fangfang Xia, Argonne National Lab



PANEL: HPC in the Cloud - Hybrid HPC, What Works What Not, Cloud HPC Providers Sharing their Views Bill Magro,
Jon Thor Kristinsson, Advania
Dave Hanlon,
Gabriel Broner, Rescale
Geoff Newell, Appsbroker  Hugo Meiland, Microsoft



PANEL: Women in HPC: Diversity Challenges Susan Bobholz,
Toni Collis, Women in HPC Carolyn Coke Reed Devany, Data Vortex
Marie-Christine Sawley, Intel
16:00 Dell EMC Solutions, in partnership with Intel, are providing the infrastructure to enable a broader definition of HPC Adnan Khaleel and Monika Gupta, Dell/EMC
16:20 Harness the Power of Supercomputing in the Cloud Girish Bhatia and Stacey Mccallum, Cray
16:40 Hyper-converged and ultra-compact RSC Tornado solution for HPC, datacenter and cloud Alexey Shmelev, RSC Group
17:00 What's new with Intel® Omni-Path Architecture Susan Bobholz, Intel
17:20 Application performance and scaling on OPA Jeffrey Salmond, University of Cambridge
17:40 Designing efficient big memory systems in HPC Andrey Kudryavtsev, Intel
18:00 EXHIBIT CLOSES 18:00
WEDNESDAY, June 27th - Machine Learning Day
TIME Topic Host/Speakers Guest
10:00 EXHIBIT HALL OPENS 10:00 a.m.



PANEL: HPC Workloads/Apps  - Skylake & OPA Aswathy Nair, Intel Wim Slagter, ANSYS®
Keith Foston, Siemens
Piush Patel, Altair
Sameer Shende, University of Oregon
11:20 Moving to Skylake and Beyond with the Latest Intel Compilers v19.0 Fabio Baruffa,
11:40 Intuitive Visual Performance Modeling & Tuning - Using Roofline Analysis for Performance Tuning Zakhar Matveev and Mike Lee,



Innovative NVM Solutions for HPC Storage and Memory
Andrey Kudryavtsev,
Sven Oehme, Spectrum Scale Research
Zivan Ori, Dualport NVMe Storage, E8 Storage
Dr. Alexander Moskovsky, General Manager, RSC Technologies
Vladimir Mironov, Senior Researcher, Lomonosov Moscow State University
13:00 TAU Performance System and TAU Commander* Sameer Shende,
University of Oregon
13:20 Scaling up Deep Learning on Supercomputers Damian Podareanu, Valeriu Codreanu, SURFsara
Vikram Saletore,
13:40 Application Optimization
Case analysis of large-scaled on Xeon + OmniPath HPC
Jerry Liu, Inspur
14:00 Next Generation MPI - Intel® MPI Library 2019 Technical Preview Dmitry Durnov and Mike Lee, Intel
14:20 Swarm64 S64DA – Accelerate high-performance data processing and analytics Juri Schmidt,
Swarm64 AS
14:40 HPC for industrial applications on Intel® platforms R. Fischer, M. Ast, H. Manz,
INTES - Ingenieurgesellschaft für technische Software GmbH
and M. Klemm, Intel



PANEL: HPC in Manufacturing
IT Transformation for Industrial Competitiveness
Marie-Christine Sawley,
Gerd Buttner, Airbus
Wim Slagter, ANSYS®
TBD, Gridcore
Rolf Fischer, Intes,
16:00 Intel® Xeon® Scalable Processor Family Vladimir Kostarev, Intel
16:20 Interactive Supercomputing to Enable Machine Learning at Lincoln Laboratory Albert Reuther,
Massachusetts Institute of Technology
16:40 True HPC Cloud (IaaS): The Holy Grail of HPC Gisli Kr. and Jon Thor Kristinsson,



HPC in the Cloud
User Panel: 
How do Users Consume Cloud Services? Customer Experiences in Consuming HPC Solutions from the Cloud. What Sets the Players Apart?  Does it Support Hybrid?  What about GDPR?
Ullrich Becker-Lemgau,
Andrew Cusick, DXC
Tony Spagnuolo, Rescale
Gisli Kr. , Advania
Mark Parsons, EPCC, University of Edinburgh
Karlheinz Pischke, Microsoft Azure
18:00 EXHIBIT CLOSES 18:00