Learn about best practices for programming Computational Memory in this technical talk featuring experts from Elephance Memory, Dell, Metisx, and Mips. Explore optimization techniques for performance while maintaining portability and transparency in both programmable and fixed-function near-data processing offload. Discover insights from an OCP whitepaper developed through collaboration with 41 industry and academic participants, including major tech companies like Samsung, SK hynix, Intel, Marvell, Arm, MIPS, and Nvidia. Gain understanding of complementary standardization efforts in memory data movement and acceleration through SNIA's SDXI Technical Workgroup.
Programming Near Data Processing - Best Practices and Industry Standards
Open Compute Project via YouTube
Overview
Syllabus
Programming Near Data Processing
Taught by
Open Compute Project
Related Courses
-
Memory Semantics and Data Movement with CXL and SDXI - Standards and Applications
-
Smart Data Accelerator Interface - Use Cases, Futures and Proof Points
-
SDXI Internals and Standardizing Memory to Memory Data Movement
-
Smart Data Accelerator Interface (SDXI) - Use Cases, Proof Points and Future Development
-
Envisioning a Computational Storage Architecture with SDXI Data Mover - Early Efforts
-
SDXI Specification v1.0 and Beyond - Memory Data Movement and Acceleration Interface
