High Performance Computing

 View Only

Attend IBM User Group Meetings and Seminars at SC19 in Denver

By Andy Morris posted Tue October 22, 2019 11:35 AM

  

High performance computing touches lives everywhere — every day.

Join IBM at SC19 in Denver, Booth #1525

The lines between traditional high performance computing (HPC) and new AI workflows are quickly eroding. Today, our work in HPC and AI is more essential than ever, driven by an urgent need to provide computational solutions for the world’s greatest challenges. The impact of these changes goes beyond traditional research boundaries to enhance our daily lives.

In addition, listen to IBM subject matter experts as they cover a variety of hot topics during the week, ranging from AI to supercomputers, and from containers to Kubernetes for HPC environments.

The following are abstracts for select IBM Technical Seminars and User Group Meetings at SC19. You can register here.

SC19 User Group Meetings | November 17-21, 2019
Denver, Colorado
Hyatt Regency Hotel, Third Floor Briefing Center

IBM LSF & HPC User Group Meeting including MC04 Briefing: Containers & Kubernetes for HPC Environments

Tuesday, November 19, 2019 Centennial Ballroom G
2:30 pm – 5:30 pm

Abstract: Come learn about the latest developments In IBM HPC and AI technology. We have a full agenda of customer and expert presentations on topics including AI-fueled intelligent simulation, and Spectrum LSF enhancements for accelerated and containerized workloads, kubernetes, and hybrid cloud bursting. The meeting will be followed by a cocktail reception.

2.30 Welcome and introductory remarks

Keri Olson, Director, HPC & Analytics Offerings


2.35 Containers & Kubernetes for HPC Environments (MCB04)

Khalid Ahmed, Distinguished Engineer, Spectrum Computing and IBM Cloud Private


3.10 Applied Bayesian Optimization in HPC

Chris Porter, HPC Cloud Offering Manager, IBM Spectrum Computing


3.35 Using Watson Machine Learning Community Edition with IBM Spectrum LSF

John Welch, Technical Sales Specialist, Washington Systems Center


4.00 Trends in High Performance Computing: Exascale Systems and Facilities

Eric Zhang, Sr. Client Technical Advisor, IBM Federal CTO Office


4.25 What's new in IBM Spectrum LSF

Bill McMillan, Global Offering Leader - LSF, IBM Spectrum Computing


4.50 Open Floor Q&A with Refreshments

5.30 close

For more information email Bill McMillan at bill.mcmillan@uk.ibm.com

SC19 IBM Technical Seminars | November 19-20
Hyatt Regency Hotel, Third Floor Briefing Center, Centennial Ballroom F

All seminars are 60 minutes

MC04: Containers & Kubernetes for HPC Environments

Containers and Kubernetes are defining the next-generation IT infrastructure for cloud native applications enabling new applications in an agile manner with a micro-services architecture. This session will discuss how the cloud native stack and HPC software stack can be integrated to support new applications and workflows that blend agility and performance. For customers looking to have a common infrastructure the need for effective resource scheduling and workload management across Kubernetes/Openshift and HPC environments is important. This will deliver a consistent way to manage SLAs , provide performance, optimally utilize expensive resources like GPUs, reduce operational expenses and improve developer productivity.

Speaker:

Bill McMillan, Global Offering Lead, IBM Spectrum Scale and Khalid Ahmed, Distinguished Engineer, IBM Cloud Private and IBM Spectrum Computing

Date/Time:

Tuesday, November 19, 2:30pm to 3:30pm

 

MC08: Optimizing Simulation Time to Final Design

HPC environments have been used for decades to rapidly explore design spaces for a number of important product-oriented reasons: Quality, Longevity, Performance, Cost, Efficiency just name a few. Those design spaces, from drugs to wind turbines, have grown ever more complex with the advancement of scientific and engineering technologies from which the designs can draw. The challenge for HPC is, with more complex design spaces, how best to trade off fidelity of simulation (run time) with design space coverage (number of simulations) to identify a product to take to market.

Speakers:

Edward Pyzer-Knapp, Research Lead, Machine Learning and AI

Date/Time: Wednesday, November 20, 1:00pm to 2:00pm

 

Also attend IBM user group meetings at SC19 to collaborate with peers from around the globe:


Additional Briefings:

MC01: AI needs IA: The Critical Role of Information Architecture for AI Success

With the growing importance of organizational data assets, the management of data and its storage takes on critical importance. In this session, we describe the data management requirements of different elements of the AI data pipeline and how key storage management technologies work together to provide the infrastructure needed to initiate and grow AI projects. We will cover best-of-breed solutions based on NVIDIA DGX and IBM Power for unstructured file data (IBM Spectrum Scale), objects (IBM Cloud Object Storage), and meta- data (IBM Spectrum Discover) on high performance computing architectures represented in the IBM Spectrum Storage for AI reference architecture.

Speakers:

Doug O'Flaherty, Director IBM Storage Product Marketing
Tony Paikeday, Director of Product Marketing, NVIDIA AI and Deep Learning

Date/Time:

Tuesday, November 19, 10:00am to 11:00am

MC02: Designing a CORAL Summit System for YOUR Center

In this session, we will describe how clients are installing their own instances of the world's smartest and most powerful computer. Architecture includes consideration of networking, servers, storage, software and services. Specific installations at major research institutions such as MIT and Miami will be described and documented. Speakers will include actual engineers who designed the systems, as well as installed them at the client sites, and will include successes and lessons learned.

Speaker:

Fred Robinson, IBM Lab Services

Date/Time:

Tuesday, November 19, 11:30am to 12:30pm

 

MC03: Faster Time to Value and Flexibility for Your AI/ML/Analytics Workloads

This session will focus on the newest member of the Elastic Storage Server family, which brings ground-breaking capability for Artificial Intelligence (AI) and Deep Learning (DL), providing full GPU resource utilization in servers such as Power AC922 and the fastest benchmarks with NVIDIA DGX systems. New containerized deployment and upgrade capability provide faster time-to-value and easy maintainability for resource-limited IT organizations. This session will cover IBM’s new high-performance storage tier, perfect for supporting data acceleration required by data-hungry AI and ML workloads. Powered by NVMe drives with active-active internal dual CPU controllers in a compact 2U enclosure delivering flexible configurations starting as small as 22TB growing up to 369TB within the 2U enclosure, at nearly linear performance scalability. For larger configurations, additional 2U systems can be stacked, or even larger Elastic Storage Server capacity models can be leveraged.

Speakers:

Doug Petteway, Offering Manager, IBM Elastic Storage Server

Date/Time:

Tuesday, November 19, 1:00pm to 2:00pm


MC05: Meet the IBM Champions of Storage and Power

The IBM Champion recognition program recognizes the innovative thought leaders in the technical community and rewards their contributions by amplifying their voice and increasing their sphere of influence. IBM Champions are client and business partner enthusiasts and advocates: IT professionals, business leaders, developers, executives, educators, and influencers. Come to this session to hear from a panel of current IBM Champions representing Storage and Power about their Champion experiences and expertise in leading technology.

Speaker:

Libby Ingrassia, IBM Champions Global Program Manager
Robert Oesterlin, Principal Engineer, Nuance, Inc.
Simon Thompson, Infrastructure Architect, University of Birmingham
Niel Balsino, Sr. Power Engineer-Linux, Sirius Computer Solutions, Power Champion
Peter Basmajian, IBM Storage Global Marketing and Alliances Manager

Date/Time:

Tuesday, November 19, 4:00pm to 5:00pm

 

MC06: Machine Learning and Deep Learning at Scale: Field Stories

Join us on a journey through Machine Learning / Deep Learning and explore what it means for the enterprise, using real world customer case studies. We’ll explain how to leverage the breadth of the IBM cognitive portfolio and hear how multiple users, using multiple frameworks, systems, and architectures can utilize the latest and greatest deep learning technologies.

Speaker:

Eric Fiala, AI ML DL Solution Architect, IBM Systems

Date/Time:

Wednesday, November 20, 10:00am to11:00am

 

MC07: Object Storage for HPC Workloads - Can I Really?

An interactive discussion. Keeping massive amounts of data online for HPC, AI or
Analytics can quickly become cost prohibitive and expensive to manage. Object storage is the fastest growing option for large amounts of data with parallel access, custom metadata, ease of scalability, and access from any location. In this session we will quickly cover the basics and then take a detailed look at object storage and some of the expanding possibilities for HPC workloads with IBM Cloud Object Storage. We will also discuss some future directions and ways that object storage is expanding beyond backup and archive.

Speakers:

David Wohlford, WW Marketing IBM Cloud Object Storage

Date/Time:

Wednesday, November 20, 11:30am to 12:30pm


MC09: The Secret [Weapon] Behind the World's Smartest Supercomputers

IBM Spectrum Scale, the industry-leading parallel file system behind the most powerful supercomputers in both the private and public sectors, continues to advance. Latest enhancements make it ideal for HPC, Big Data, AI, and other demanding workloads where there is no room for compromise in either performance or scalability. In this session we will cover the advances in the past year that have made IBM Spectrum Scale easier to use and more flexible on a variety of workloads and platforms. Plus, with the introduction of Erasure Code Edition, Spectrum Scale brings those same world-leading enterprise capabilities to storage rich servers with commodity hardware. We will highlight a handful of installations to explore usage patterns and outline possible future product directions.

Speakers:

Christopher Maestas, Executive Architect, IBM Spectrum Scale

Date/Time:

Wednesday, November 20, 2:30pm to 3:30pm


Additional User Group Meetings:

IBM Spectrum Scale User Group Meeting Sunday, November 17, 2019
Centennial Ballroom D
9:30 am –12:00

The morning session is dedicated to new and potential users of IBM Spectrum Scale. Experienced users and technical leaders will host an introductory session featuring key use cases, basic concepts and an interactive Q&A. Lunch provided. 12:30 pm –5:30 pm
The global user-led meeting provides an opportunity for peer-to-peer learning and interaction with IBM's Spectrum Scale technical development leadership. This session will feature presentations on actual customer use cases. Topics include how to design an IBM Spectrum Scale environment, product strategy and AI update. Client presenters from companies such as Nuance, SciNet, Brown University, University of Queensland and others will present details. This year's agenda will also feature lightning talks on user-generated topics. A cocktail reception will follow the main agenda at 6 pm to continue the discussion. Lunch is also provided.

Open Compute Rack and Server Project Meeting Monday, November 18, 2019
Centennial Ballroom F
1:00 pm – 4:00 pm

Abstract: The goal of this meeting is to introduce the Open Compute Project and its HPC sub-project to the technical community at SC19. The Open Compute Projects (OCP) is reimaging hardware, making it more efficient, flexible, and scalable. The meeting will share specifications on the OpenCompute Modular Accelerator project and share some of the latest OCP designs that could benefit scale-out HPC organizations.

Spectrum Scale Design Thinking Tuesday, November 19, 2019 Centennial Ballroom G
10:00 am – 12:00 pm

Abstract: A collaborative co-creative session to explore how Spectrum Scale can deliver differentiating user experiences to address future needs and technologies. We desire to get first-hand information from our customers to get input for our roadmap. Both sessions will have a different focus. We will announce details at the Spectrum Scale User Group Meeting on Sunday Nov 17th.

IBM HPC & AI University User Meeting Thursday, November 21, 2019 Centennial Ballroom F
8:30 am – 12:30 pm

Abstract: The 4th annual IBM University HPC/AI User Group will highlight user experiences and computational sciences on IBM Power Systems from the University of Miami, Princeton University, The University of Illinois Urbana Champaign, and others. The meeting is open to all academia interested parties worldwide.


#SpectrumComputingGroup
0 comments
27 views

Permalink