Uploaded May 2026 | Updated September 2026, 2 weeks ago

---

This presentation talks about how data flows through a typical LLM training sequence and its relationship to system/GPU memory. We will examine a simple system and how it scales to larger more complex systems. We will also talk about the need and differences between Scale-up and Scale-out networks.

Presented by Bill Lynn | AMD - Fellow, Systems Design Engineering

Learn More:
• SDC: StorageAI Website: snia.org/sniadeveloper/storageai
• SNIA Website: snia.org
• SNIA Educational Library: snia.org/library
• X: twitter.com/SNIA
• LinkedIn: linkedin.com/company/snia