Transcription
Nikolay Mester, HPC and CSP verticals, Eastern Europe, Intel
The HPC OpportunityModeling & Simulation 515averagereturn per 1of HPCinvestment1HPC Data AnalyticsArtificial IntelligenceVisualization18% revenue 55% revenue 30% revenueCAGR; 3 CAGR; 47 CAGR; 1.6billion inbillion inbillion in2020220203202041 Source: Source: IDC HPC and ROI Study Update, September 20152 Source IDC Worldside High-Performance Data Analytics Forecast 2016-2020, June 20163 Source: IDC Worldwide Semiannual Cognitive/Artificial Intelligence Systems Spending Guide, Oct 20164 Source: MarketsandMarkets Visualization and 3D Rendering Software Market by Application, March 2016Intel and the Intel logo are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries. Other names and brandsmay be claimed as the property of others. All products, dates, and figures are preliminary and are subject to change without any notice. Copyright 2013, Intel Corporation.4
A Holistic Architectural ApproachSystemInnovative TechnologiesApplicationTighter IntegrationPerformanceComputeModernized geSystemSoftwareFPGAGraphicsProprietaryI/OTime3
Key Elements ofIntel SSFINTEL Scalable System FrameworkIntel Market1LeadingHighlyParallelIntel ntelSupported* Other names and brands may be claimed as the property of others. 1Source: Intel estimates.*Flexibility& StabilityExtremeScalability11
Code Modernization for Higher PerformanceDPLot of performance isbeing left on thetable152xVP Vectorized & Parallelized (MT)SP Scalar & Parallelized (MT)VS Vectorized & Single-Threaded (ST)SS Scalar & Single-Threaded (ST)We believe mostcodes are here4C4C6C8C12C14C22CModernization (i.e. parallelization and vectorization) of your code is thesolutionSoftware 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 usingspecific 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 toassist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to www.intel.com/benchmarks.5
Tighter System-Level IntegrationInnovative Memory-Storage HierarchyHigher Bandwidth.Lower Latency and chesLocal memory is now faster& in processor packageOn-Package HighBandwidth Memory*Memory BusLocal MemoryComputeNodeLocal StorageMuch larger memory capacities keepdata in local memoryI/O Node storage movesto compute nodeIntel DIMMs based on3D XPoint TechnologyIntel Optane Technology SSDsI/O NodeSSD StorageSome remote data movesonto I/O nodeRemoteStorage*cache, memory or hybrid modeParallel File System(Hard Drive Storage)Burst Buffer Node withIntel Optane Technology SSDsParallel File System(Hard Drive Storage)Copyright 2016 Intel Corporation. All rights reserved.*Other names and brands may be claimed as the property of others.Andrey Semin 29 September 2017 Slide 6LocalMemory
What to use for your situation?Why Xeon Phi ?ImprovePerformance-AND/OR-ImproveROIWhich Apps?1Scalable to 60 coresOptimized -OR-UnlockPotentialIf yes Local memoryBW boundIf no Commonly-UsedParallel Processor*Intel Xeon Phi is optimal for applications that scale to 60cores and are highly threaded or memory bandwidth bound1Performanceresults on Intel Xeon Phi will vary depending on app characteristics. For more information,see: https://software.inte
The HPC Opportunity Modeling & Simulation HPC Data Analytics Artificial Intelligence Visualization 4 55% revenue CAGR; 47 billion in 20203 1 Source: Source: IDC HPC and ROI Study Update, September 2015 2 Source IDC Worldside High-Performance Data Analytics Forecast 2016-2020, June 2016
