Smart Data Accelerator Interface: Use Cases, Proof Points, v1.1 and beyond

Mon Sep 16 | 2:30pm

Location:

Cypress

Abstract

The SNIA Smart Data Accelerator Interface (SDXI) Technical Work Group, provides an update on this SNIA standard for a memory-to-memory data movement and acceleration interface. Learn about applicable use cases for SDXI-based accelerators including those in emerging areas like artificial intelligence, and how you can participate in future work in this growing ecosystem. This talk will also discuss software enablement work, PoC implementations, and a preview of features being developed for the v1.1 specification.

Learning Objectives

Learn the use cases driving SDXI development, and adoption.
Learn about some of the PoC demos demonstrated publicly.
Learn about the basics of the standard. New features being developed by the working group.
Learn how to join/contribute to the work items being developed by this workgroup.

Download Presentation

Abstract

The SNIA Smart Data Accelerator Interface (SDXI) Technical Work Group, provides an update on this SNIA standard for a memory-to-memory data movement and acceleration interface. Learn about applicable use cases for SDXI-based accelerators including those in emerging areas like artificial intelligence, and how you can participate in future work in this growing ecosystem. This talk will also discuss software enablement work, PoC implementations, and a preview of features being developed for the v1.1 specification.

Learning Objectives

Learn the use cases driving SDXI development, and adoption.
Learn about some of the PoC demos demonstrated publicly.
Learn about the basics of the standard. New features being developed by the working group.
Learn how to join/contribute to the work items being developed by this workgroup.

---

Shyam Iyer

Dell

Related Sessions

Magic Memory Access

CXL for AI/ML: A Practical Guide to Unleashing the Potential of AI and Machine Learning

Steve Scargall

MemVerge

Favorites

Magic Memory Access

Optimized Resource Allocation for CXL Tiered-Memory Systems

Main memory dominates data center server cost, and hence, data center operators are exploring alternative technologies such as CXL-attached memory to improve cost without jeopardizing performance.

Heiner Litz

University of California, Santa Cruz (UCSC)