Bill Magro, Ph.D. Intel Fellow, Software and Services Group Chief Technologist, High Performance Computing

Transcription

1 Bill Magro, Ph.D. Intel Fellow, Software and Services Group Chief Technologist, High Performance Computing

2 2 Legal Disclaimer & Optimization Notice INFORMATION IN THIS DOCUMENT IS PROVIDED AS IS. NO LICENSE, EXPRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE, TO ANY INTELLECTUAL PROPERTY RIGHTS IS GRANTED BY THIS DOCUMENT. INTEL ASSUMES NO LIABILITY WHATSOEVER AND INTEL DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY, RELATING TO THIS INFORMATION INCLUDING LIABILITY OR WARRANTIES RELATING TO FITNESS FOR A PARTICULAR PURPOSE, MERCHANTABILITY, OR INFRINGEMENT OF ANY PATENT, COPYRIGHT OR OTHER INTELLECTUAL PROPERTY RIGHT. Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. Copyright 2014, Intel Corporation. All rights reserved. Intel, Xeon, Xeon Phi, and the Intel logo are trademarks of Intel Corporation in the U.S. and other countries. Optimization Notice Intel s compilers may or may not optimize to the same degree for non-intel microprocessors for optimizations that are not unique to Intel microprocessors. These optimizations include SSE2, SSE3, and SSSE3 instruction sets and other optimizations. Intel does not guarantee the availability, functionality, or effectiveness of any optimization on microprocessors not manufactured by Intel. Microprocessor-dependent optimizations in this product are intended for use with Intel microprocessors. Certain optimizations not specific to Intel microarchitecture are reserved for Intel microprocessors. Please refer to the applicable product User and Reference Guides for more information regarding the specific instruction sets covered by this notice. Notice revision #

3 Legal Information This presentation contains the general insights and opinions of Intel Corporation ( Intel ). The information in this presentation is provided for information only and is not to be relied upon for any other purpose than educational. Statements in this document that refer to Intel s plans and expectations for the quarter, the year, and the future, are forward-looking statements that involve a number of risks and uncertainties. A detailed discussion of the factors that could affect Intel s results and plans is included in Intel s SEC filings, including the annual report on Form 10-K. Any forecasts of goods and services needed for Intel s operations are provided for discussion purposes only. Intel will have no liability to make any purchase in connection with forecasts published in this document. Intel accepts no duty to update this presentation based on more current information. Intel is not liable for any damages, direct or indirect, consequential or otherwise, that may arise, directly or indirectly, from the use or misuse of the information in this presentation. Intel technologies features and benefits depend on system configuration and may require enabled hardware, software or service activation. Learn more at intel.com, or from the OEM or retailer. Copyright 2017 Intel Corporation. Intel, the Intel logo, Xeon. Movidius and Stratix are trademarks of Intel Corporation in the U.S. and/or other countries. *Other names and brands may be claimed as the property of others 3

4 FASTER DISCOVER Y Better Products Deeper Insights

5 Astrophysics Life Sciences Weather Climate Manufacturing Energy Financial Security 5

6

7 What is HPC? "High-Performance Computing," or HPC, is the application of "supercomputers" to computational problems that are either too large for standard computers or would take too long. - NICS High-performance computing (HPC) is the use of super computers and parallel processing techniques for solving complex computational problems. - Techopedia The term high performance computing (HPC) refers to any computational activity requiring more than a single computer to execute a task. - HPC Wales HPC: An activity characterized by the workload s nature, intent, and response to scale

8 Scope Sc Ale Delivery Analytics and AI Exascale HPC in the Cloud

9

10 By 2020 Avg. internet user 1.5 GB OF TRAFFIC / DAY Autonomous vehicles4 TB OF DATA / DAY CONNECTED AIRPLANE5 TB OF DATA / DAY Smart Factory1 PB OF DATA / DAY

11 The next big wave of computing mainfra mes Standar dsbased servers Cloud computi ng Data deluge COMPUTE breakthrough Innovation Artificia surge l intellige nce AI Compute Cycles will grow 12X by 2020 Source: Intel

12 Smart Economics Autonomous driving Highly compute intensive: dense matrix m Highly fabric intensive: frequent global exc

13 EXPERIMEN TAL Observation THEORETIC AL Mathematical Model COMPUTATI ONAL Numerical Simulation & Modeling Data Analytics Visualization Computational image: CERN

14

15 Performance 10 EFlop/s 1 EFlop/s 100 PFlop/s 10 PFlop/s Sum #1 #500 1 PFlop/s 100 TFlop/s 10 TFlop/s 1 TFlop/s 100 GFlop/s 10 GFlop/s 1 GFlop/s 100 MFlop/s 10 MFlop/s 1 MFlop/s Data source: top500.org

16 2009 Outlook: Impractical Power Consumption Power (KW) 1000, , Power Consumption? An ExaFLOPS machine based on 2009 technology 500 MW 1 GW Other misc. power consumptions: Power supply losses Cooling etc PFLOP Disk 10MW 10EB TFLOP Comm 100MW ~100pJ per FLOP 100 GFLOP MFLOP Memory Compute 150MW 200MW ~1500pJ per Byte ~400W / Socket Source: Intel internal estimates; not product representative

17 2009 Outlook: Extreme System Scale Total Performance 1 EFLOPS (10**18) 10,000,0001,000, ,000 10,000 1 PFLOPS (10**15) ~ ,000 1, TFLOPS (10**12) ~ GFLOPS (10**9) ~1987 Note: Numbers are based on Linpack Benchmark. Dates are approximate. 100 GFLOPS 1 TFLOPS 10 TFLOPS Performance per CPU 100 TFLOPS

18 Intel Global Race to 18 Exascale Computing Initial systems

19 Compute MEMORY/Sto rage Fabric Software

20

21 Challenges for Existing users Image courtesy of Cycle Computing

22 Challenges for New Users Infographic source: NCMS.org Top barriers to adoption: Awareness Defensible ROI Business risk Access to technology and expertise

23 Data value/sensitivity Low High Example HPC workload fit for the cloud Private Cloud Dedicated HPC System Public Cloud Low High Application Coupling & Infrastructure sensitivity

24 HPC Cluster HPC in the Cloud Applications Platform Provisioning, Messaging, Runtimes, System Management, Resource Management Infrastructure High Performance Components Applications Platform Messaging, Runtimes, System Management, Resource Management HPC Provisioning Layer IAAS Standard or High Performance Components Server Storage SSDs Switch Fabric Infrastructure Server Storage SSDs Network Infrastructure

25 Data value/sensitivity Low High Cloud s Expanding HPC frontier Security Certifications batch interfaces HPC PAAS (Orchestration) Public Cloud High Performance Instance types Placement groups Fast Networking fabrics Bare metal instances Dedicated HPC System Low High Application Coupling & Infrastructure sensitivity

26 HPC by Deployment Type 30% HPC: On-Premise HPC: Cloud *Based on Intel internal analysis

27 Cloud: An HPC Center Threat OR Opportunity?? Tailored HPC Systems Public cloud Adding cloud technologies and capabilities Adding HPC technologies and performance

28 Simulation Visualization Artificial intelligence Data Analytics Compute MEMORY/Sto rage Fabric Intel Xeon Nervana Scalable Neural Network Processor Processor 3dXPoint Technology Omni-PatH Architecture

29 Simulation Visualization Artificial intelligence Data Analytics Software Compute MEMORY/Sto rage Fabric

30 Modeling & simulation visualizatio n ai High Performanc e Data Analytics Small Clusters Through Supercomputers Compute and Data-Centric Standards-Based Programmability On-Premise and Cloud-Based

31