1、 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.C M P 3 4 0HPC at Scale with AWS Parallel Computing Service(PCS)Michael GujralHPC Manager,ShellSatoshi IkemureAdministration Leader,Toyota Central R&D Labs.Ian Colle
2、Director,Adv Compute&SimAmazon Web Services 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.AgendaWhy HPC needs a managed serviceIntroduction to AWS PCS and overviewWho needs AWS PCSArchitecture,Design patterns,Cost and PricingToyota Central R&Ds journey with AWS PCSShells journey
3、 with AWS PCSOrchestrating HPC and AI with AWS Batch 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Why HPC needs to get simplerIndustry TrendsSimulation and AI workloads expanding in every sectorCluster sizes and GPU fleets growing year over yearBusiness outcomes increasingly de
4、pend on HPC resultsConstraintsDemand for HPC capacity grows faster Longer lead times to design,deploy,and validate clustersEngineers/Researchers pulled into cluster operations instead of product/research work 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.The Need for Simplified
5、HPC SolutionsCommon challengesComplex cluster setupHard to scale and operate reliably Need for HPC/Slurm expertiseSimple,repeatable setupAutomatic,policy-driven scalingManaged Expertise(and best practices!)What a managed service should provideSource:DataIntelo,Cloud HPC(High Performance Computing)Ma
6、rket,2025-2033 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Introduction to AWS Parallel Computing Service(PCS)A fully managed service that makes it easier for you to run and scale your high-performance computing(HPC)workloads on AWS 2025,Amazon Web Services,Inc.or its affiliat