借助英特尔® 至强® 加速人工智能创新.pdf

1、Accelerating AI Innovation with Intel XeonJames McFadden,Milan Mehta,David HanleyOct 14th20252Lenovos AI Product PortfolioJames McFaddenExecutive Director,Product Quality EngineeringLenovoSD650-NSR680aSR780aHPC&AI Workload PortfolioPCIe Card Acceleration HPC AISC777Enterprise AI&LLMsEntry level AI&H

2、PC acceleration Lightweight AI model inferenceScalable AI training&deep learningAdvanced AI&deep learning Scientific modelingLLM(move to first)HG660X V3SR650aPlatforms designed to accelerate AI model training,HPC simulations and LLM development.Equipped with high GPU density and innovative Neptune c

3、ooling for maximum performance and thermal efficiency.SR650SC7502025 Lenovo.All rights reserved4AgileStrategic Quality Management Outperforms Reactive Quality ManagementE2E QualityPartnershipMeasurementCommunicationsScalableManufacturingSupplier ManagementMeasuredRepeatableTransparentPCRs Change Con

4、trolT&C DocumentsScorecardsOn Site Verification(OSV)Yield ReviewsPCN ManagementManufacturing Readiness ReviewsIQCFAIsAuditsProcess FMEAProcess ControlQuality Excursion ManagementClosed Loop ReviewsCustomer FeedbackDaily Roll CallFPY,TPY,PRRSupplier ScorecardLenovo-ClientCommunicationPCN ManagementAc

5、countablePN and SN TraceabilityUnique NPI MethodologyPN/BOM ControlOffering Delivery TeamAccuracyNUDD ReviewCertificationsLenovo Quality ApproachLenovo SR680a v3 5thGeneration Intel CPUsHigh Compute+Memory bandwidth accelerates scientific and engineering modelsLow-latency networking andCPU-GPU syner

6、gyMulti-GPU NVLink architectureModeling&SimulationTraining LLMsFinTechGPU Power and NVMe storage deliver real-time renderingRenderingMIG and TruScale make it ideal for shared GPU resources in hybrid environmentsGPUaaSLenovo SR680a v3 Lenovo SR780a v3 Liquid cooling6Lenovo High Performance ComputingL