Dell AI and HPC University Roadshow

Dell AI and HPC University Roadshow

• 1.
  1 Industry Trends inAnalytics Business Intelligence, Data Warehousing, Big Data, Artificial Intelligence AI and HPC University Roadshow Bill Wong – Artificial Intelligence and Data Analytics Practice Leader Bill Kiatipis - High Performance Computing Strategist Dell Technologies
• 2.
  2 Agenda  Key DataTrends and Challenges  Digital Transformation Through Analytics  HPC Platform Computing  Summary
• 3.
  3 Higher Education AdvancedAnalytics Drivers Enhance Student Experience to Attract and Retain Students, Improve Post- Education Outcomes Personalized Learning to Improve Student Outcome, Improve Student Performance and Graduation Rates Drive and support Research, Partnerships and Entrepreneurial Initiatives in Key Industries Smarter Campus to enhance the student experience and campus facilities by transforming the economic, social, and technology foundation
• 4.
  Q. What arethe technology areas where your organization will be spending the largest amount of new or additional funding in 2019? n = 3,086. Q. What are the technology areas where your organization will be reducing funding by the highest amount in 2019 compared to 2018? n = 2,819. Multiple responses allowed, excludes “don’t know.” Source: Gartner “The 2019 CIO Agenda: Securing a New Foundation for Digital Business,” (G00366991) Data and Analytics Investments Leads New Digital Transformation Initiatives (Again) for CIOs Analytics investments continue to increase Plans to Increase Investment for Digital Transformation
• 5.
  5 The AI TechnologyEnablers 1950 1960 1970 1980 1990 2000 2010 2020 Cost of Compute Amount of Data Artificial Intelligence Machine Learning Deep Learning - Artificial Intelligence (AI) is human intelligence mimicked by machine algorithms, examples: Chess, Go, Facial Recognition - Machine Learning (ML) is a subset of AI algorithms to parse data, learn from data, and then make a determination or prediction, example: Spam Detection, Preventative Maintenance - Deep Learning (DL) a subset of machine learning algorithms that leverage artificial neural networks to develop relationships among the data, examples: Driverless Cars, Cyber-Security Accelerators Algorithms Big Data Traditional Programming Machine Learning
• 6.
  6 Expectations Plateau of Productivity Peak ofSlope of EnlightenmentInnovationTrigger Trough of Disillusionment Inflated Expectations Hype Cycle for Artificial Intelligence “Narrow" AI is becoming better than humans at defined tasks. "General" AI is still a long way off.” Time Plateau will be reached less than 2 years 2 to 5 years 5 to 10 years more than 10 years Deep Learning Infrastructure Transformation Autonomous Vehicles “AI, one of the most disruptive classes of technologies, will become more widely available due to cloud computing, open source and the “maker” (developers, data scientists and AI architects) community. While early adopters will benefit from continued evolution of the technology, the notable change will be its availability to the masses. As of July 2019 AI PaaS Artificial General Intelligence Machine Learning NLP FPGA Accelerators GPU Accelerators DNN ASICs Quantum Computing Neuromorphic Hardware Computer Vision Speech Recognition
• 7.
  7 Top 10 Typesof Hardware for AI Delivery* 1. Processors (CPU, GPU, FPGA, ASIC) 2. HPC / Supercomputer Infrastructure 3. Communication Network 4. Personal Devices 5. Connected Home Devices 6. AR / VR Head-Mounted Displays (HMD) 7. Drones 8. Robotics 9. Automotive 10.Sensors and Application Components (audio, camera, LiDAR, etc.) *The Business Impact and Use Cases for Artificial Intelligence, Gartner, 2017 Accelerate computational performance AI-enabled endpoints AI-enabled autonomous endpoints
• 8.
  8 AI Accelerators Flexibility Efficiency “Startingtoday (November 13, 2019), Microsoft is providing Azure customers with access to chips made by the British startup Graphcore.” Microsoft Sends a New Kind of AI Processor Into the Cloud – https://www.wired.com/story/microsoft-sends-a-new-kind-of-ai-processor-into- the-cloud/
• 9.
  9 Dell EMC DSS8440 A Dynamic Machine Learning and HPC Platform Differentiated Features Benefits Up to 10 Accelerator cards including Industry leading NVIDIA Telsa V100 GPUs and Graphcore Colossus C2 IPU cards (Post RTS) Exceptional horsepower in a 4U chassis Supports up to 205W CPUs with accelerators in 35C environments Thermally unconstrained provides flexibility for a variety of configurations/ environments Up to 10 drives of local storage (NVMe and SAS/SATA) Accelerated access to training data 8 x PCIe Gen3 Extensive I/O options for network/IO traffic DSS 8440 with NVIDIA GPUs DSS 8440 with Graphcore IPUs
• 10.
  10 Worldwide Artificial Intelligence2018 Share Snapshot Dell Technologies' 2018 AI revenue grew 72.6% to $1.89 billion, from $1.10 billion in 2017. Dell revenue in the AI market is primarily from infrastructure (server and storage). *IBM's AI revenue grew 19.0% to $2.58 billion, from $2.17 billion in 2017. Revenue is divided across software, hardware, and services, with services and hardware significantly larger than software ($349.6M).
• 11.
  © Copyright 2019Dell Inc.11 Computer Vision Analytics Solutions - Examples • Agent Video Intelligence: automates video analysis to detect and alert for events of interest, expedite search in recorded video and extract statistical data from the footage captured by surveillance cameras. • BriefCam: provider of Video Synopsis® and Deep Learning solutions for rapid video review and search, real-time alerting and quantitative video insights. Transforms raw video into actionable intelligence. • Deep Vision AI: highly customizable “deep learning” platform that provides, facial and object recognition, as well as behavioral analysis. Used in more specialized situations where customer has unique models such as for uniform detection, sanitary processes, foreign objects in food processing as a few examples.
• 12.
  © Copyright 2019Dell Inc.12 • Eliminate inefficient islands of storage – Infrastructure consolidation for both clinical and non-clinical workloads • Scales as data growth and number of instruments, modalities, and digital clinical applications increases • Enable better information sharing • Accelerate data analytics to gain new insight • Extends into the cloud • Prepared for next generation analytics Dell EMC Data Lake Caffe2 Data Lake Storage Platform
• 13.
  13 Dell EMC HPC MarketLeadership generations of servers and storage in HPC clusters10 1st HPC cluster HPC solutions program officially launches Industry’s 1st HPC Solution Bundle #4 Tungsten Thunderbird #6 DCS formed C-series joins PowerEdge Stampede #7 Zenith System launched at Dell EMC HPC Innovation Lab Dell EMC merger — Isilon joins HPC portfolio 1999 2001 2004 2005 2008 2012 2015 2016 Dell EMC AI solutions announced 2017 2019 Frontera 2018 First systems with DCLC (U. Michigan) and HDR (OSC)
• 14.
  14 World-class infrastructure inthe Innovation Lab Zenith • TOP500-class system based on Intel Scalable Systems Framework (OPA, KNL, Xeon, OpenHPC) • 424 nodes dual Intel Xeon Gold processors, Omni-Path • +160 Intel Xeon Phi (KNL) servers. • Over 1 PF combined performance! • #396 on Top500, 1.86 PF theoretical peak • Lustre, Isilon H600, Isilon F800 and NSS storage • Liquid cooled and air cooled Rattler • Research/development system with Mellanox, NVIDIA and Bright Computing • 88 nodes with EDR InfiniBand and Intel Xeon Gold processors • 32x PowerEdge C4140 nodes with 4x NVIDIA GPUs Other systems • 32 node AMD cluster, storage solutions, etc. 13K ft.2 lab, 1,300+ servers, ~10PB storage dedicated to HPC in collaboration with the community
• 15.
  15 Compute/ Storage/ Networking Verticals/ Use Cases Processor/ Accelerator Software& Frameworks Management/ Orchestration Virtualization Xeon Phi R840 In-Memory Analytics C4140 R740C6320p Training Machine Learning & Deep Learning eco-system – solving real world problems Xeon FPGA Adapter Consumption Models Math Libraries Crest Family Big Accelerator system BigDL FPGA Adapter C6420 R740 Inference Xeon T640R940XA Enterprise ISV ML Software Open Source Frameworks hpcWeb Recommendation Image Smart Disease Predictive Fraud Smart Traffic Threat Inventory engines classification chatbots Identification marketing Detection Core to Edge IoT Predictions Management BrightML BigData & HPC orchestration Containers orchestration asaka Bitfusion Accelerator Virtualization & pooling Service Providers Systems IntegratorsServices Solution BuildBuy Hyperconverged appliance In-Memory Analytics Training Inference
• 16.
  16 Dell Value-Add forHPC Customers Dell EMC Ready Solution for HPC Life Sciences Dell EMC Ready Solution for HPC Research HPC Sales Specialists Dell EMC Ready Solution for NFS Storage Dell Financial Services HPC Specific Executive Briefings HPC Estimation Tool HPC/AI Innovation Lab Dell EMC Annual HPC Community Mtg & DellXL Reference Architectures Ready Solutions HPC Resources & Tools HPC Services Dedicated HPC organization HPC Solution Architects HPC Vertical SME’s HPC TSR’s & OSE’s Access to Future Roadmaps ProSupport Add-on for HPC Remote Cluster Mgmt And support packs ProDeploy for HPC Cloud Services Office of CTO HPC Visionaries Consulting Services for HPC Dell EMC Ready Solution for HPC Digital Mfg. HPC Power/Cooling and Cable Calculators Partner with leading suppliers and channel partners Dell EMC Ready Solution for HPC AI Dell EMC Ready Solution for Lustre®
• 17.
  17 Business needs Simon FraserUniversity needed increased scale and capacity to compete and excel globally using big data and big compute tools. Solutions at a glance • Heterogeneous cluster with Dell EMC PowerEdge C4130/C6320 servers, Intel® Xeon® E5-2650v4/E5- 2683-v4 processors, Intel Omni-Path, NVIDIA® Tesla® P100 & V100 GPU accelerators Business results • Expanded compute, storage and cloud resources for over 11,000 researchers across Canada • Ability to run multiple simultaneous jobs of up to 2,600 CPU cores each • Allows researchers to much more quickly analyze the DNA of microbes • Ability to identify infectious disease outbreaks faster • More rapid tracking and understanding of origins and spread of infectious disease outbreaks With greater computational power than all of Compute Canada’s legacy systems combined, Cedar is built for big data. The system can support researchers collecting, analyzing or sharing immense volumes of data https://www.youtube.com/watch?v=3RqF8m65r8g Simon Fraser University Reduce the impact of infectious disease outbreaks
• 18.
  18 Decision Criteria forAI Infrastructure/Solutions Data Scientist Perspective IDC 2018
• 19.
  The Digital FutureDemands a New Perspective Cloud First Data First Infrastructure-centric Business-centric Takes into consideration: • Data gravity • Data velocity • Data control • Data privacy and compliance Driven by: • Lower infrastructure CapEx • Offload infrastructure maintenance • Improve time to market (deployment time for infrastructure)
• 20.
  20 Analytics Development Challengesand Best Practices Data Environment Challenge Best Practices Data Quality Inconsistently formatted across the organization, often contains errors and could create biases. Minimize duplication. Implement Data Governance strategy and tools to cleanse, validate and enforce data quality. Data Silos Inability to leverage data across the enterprise for analytics Develop/engineer a common data repository to access for advanced analytics New Sources of Data Machine/Deep Learning often requires large amounts of data, from various sources, and formats (unstructured). Develop a data integration strategy and use tools to process unstructured data that requires additional processing, cleansing, and/or normalization. Supporting Near Real- time Requests Data repositories are challenged scaling transactional and query- based workloads together Consider translytical databases to address near real-time data response times for advanced analytical applications. Privacy, Security, Ethics, Auditability Often not considered at design time and leads to governance exposures. Adoption of design principles to provide data transparency for key stakeholders and accountability for governance/regulation organizations.
• 21.
  21 The Value ofDell for AI Infrastructure - Comprehensive and Scalable AI/Analytics Platform Portfolio - Workstations, Servers, Clusters, Storage, Networking - Infrastructure and Data Science and Analytics Expertise - HPC and AI Innovation Lab - IoT / Intelligent Video Analytics Lab - Solution-based Offerings - Pre-configured AI Ready Offerings - IoT / Safety and Security and Thermal Vision Solutions - GPU Virtualization - ML Platforms Infrastructure Scalability Reduce Complexity Address Demand Partner Ecosystem Cost Effective
• 22.
  22 - Appendix - DellTechnologies AI and Data Analytics Solutions
• 23.
  23 Dell Technologies AIand Data Analytics Solutions AI / Machine Learning / Deep Learning • Domino Data Science Platform Design Document • HPC for AI and Data Analytics Ready Architecture • Retail Loss Prevention Ready Solutions • DataRobot Reference Architecture • H2O AI Reference Architecture • Kubeflow Reference Architecture • OneConvergence Dkube Reference Architecture • Iguazio Reference Architecture • Deep Learning with NVIDIA Ready Solutions • Isilon with NVIDIA DGX-1 Reference Architecture • Isilon with NVIDIA DGX-2 Reference Architecture • Isilon with Dell Precision 7920 Data Science Workstation Reference Architecture • Isilon with Dell EMC DSS8440 Reference Architecture • Noodle.ai (OEM) Solution Bundle IoT / Streaming / Machine Data Analytics • IntelliSite (OEM) Thermal Detection Solution • Retail Loss Prevention Ready Solutions • Dell IoT Safety and Security Portfolio • Real-Time Data Streaming Ready Architecture • Splunk Enterprise on Dell EMC Infrastructure • Streaming Data Platform • ElasticSearch (OEM) Solution Bundle © Copyright 2020 Dell Inc. Augmented Analytics and Data Warehouse • Spark on Kubernetes • Kinetica (OEM) Solution Bundle • ThoughtSpot (OEM) Solution Bundle • Pivotal Greenplum • Dell Boomi Data Lake / Unstructured Data Infrastructure • Microsoft SQL Server 2019: Big Data Cluster Ready Solution • Cloudera Hadoop Ready Architecture • Hortonworks Hadoop Ready Architecture • Kubernetes Containers with Diamanti (OEM) Solution Bundle • Grid Dynamics Reference Architecture • Red Hat OpenShift Reference Architecture HPC Ready Solutions • HPC Digital Manufacturing • HPC Life Sciences • HPC Research • HPC BeeGFS Storage • HPC Lustre Storage • HPC NFS Storage • HPC PixStor Storage *Note, some products can deliver capabilities that address multiple use cases Product Offerings and Technical Collateral for Analytical Use Cases