4479 - AI 数据引擎 – 来自 IBM 存储的最新进展、路线图和用例.pdf

1、Orlando,FLOctober 69IBM TechXchange 20254479 AI Data EngineFrank LeeIBM StorageAI Data EngineAgenda010102More flexible deploymentsFueling AI Factory with Global Data PlatformFueling AI Factory with Global Data PlatformEnabling AI Common with Fusion AIIBM TechXchange|2025 IBM CorporationCompute PodCo

2、mpute Pod4 Scalable Units128 Compute Nodes1024 H100 GPUs82 TB of GPU Ram12,288 Physical Cores256 TB of RAM3481 TB NVME Local StorageScalable UnitScalable Unit32 Compute Nodes256 H100 GPUsIBM Blue VelaModular Deployment UnitsIBMs Purpose-Built Storage Solution for AI Factory IBM Storage Scale System

3、6000up to 1PB Usable Flash per node4310GB/s Read310GB/s Read155GB/s Write155GB/s Write1+TB/s Read1+TB/s Read775GB/s Write775GB/s Write3+TB/s Read3+TB/s Read1.5TB/s Write1.5TB/s WritePer Rack!Per Rack!310 GB/s310 GB/s155 GB/s155 GB/sStorage Scale Storage Scale System 6000 System 6000 From AI Factory

4、to AI CommonFrom AI Factory to AI CommonAI FactoryAccelerateAccelerateOptimizeOptimizeConvergeConvergeHPC ClusterHPCDGX/HGX Node with GPUtrainingGlobal Data PlatformAgenda010102More flexible deploymentsFueling AI Factory with Global Data PlatformEnabling AI Common with IBM Fusion AIEnabling AI Commo

5、n with IBM Fusion AIIBM TechXchange|2025 IBM CorporationOver time,platforms Over time,platforms and methodsand methodsF FindableindableA AccessibleccessibleI InteroperablenteroperableR ReusableeusableFAIRFAIR-able RDM(Research Data Management)able RDM(Research Data Management)Across tools,domains,fi

6、elds,disciplines and institutionsFor large,distributed and unstructured dataBy Applications,workflows,AI modelsCATALOG1DISCOVER3INTEGRATE2ORCHESTRATE4metaEngine pilot MD AndersonCancer Center“finding needle in the haystack”202120172019metaOcean pilot Broad Institute“clean up data swamp”Launch of IBM