Learn about Samsung's innovative Petabyte SSD and MangoBoost's DPU technology in this technical presentation that addresses the challenges of data-centric AI workloads. Explore how the combination of Samsung's compact, high-capacity storage solution and MangoBoost's Data Processing Unit enhances data transfer efficiency and network performance for large-scale AI infrastructure. Discover performance improvements and TCO savings through real-world evaluations using state-of-the-art AI workloads and optimized AI/ML frameworks. Principal Engineer Yunho Josh Yang from Samsung and Storage Team Lead Wonsik Lee from Mangoboost demonstrate how their collaborative solution meets the exponential surge in AI data demands while optimizing space and power requirements.
DPU-Accelerated Petascale Storage Subsystem for Near Data Computing
Open Compute Project via YouTube
Overview
Syllabus
DPU Accelerated Peta Scale Storage Subsystem for Near Data Computing
Taught by
Open Compute Project
Related Courses
-
Accelerating GPU Server Access to Network-Attached Disaggregated Storage Using DPU
-
High Performance Data Center Storage Using DPUs - Optimizing Power and Performance
-
Green Computing with Computational Storage Devices - Energy Efficiency in Data Centers
-
Storage in the Era of Large-scale AI Computing
-
Accelerating Near Real-time Analytics with High Performance Object Storage
-
Optimizing Data Center TCO - An In-Depth Analysis and Sensitivity Study
