Generative AI in the Enterprise with AMD Accelerators

• 1

  Overview
• 2

  Document purpose
• 3

  Audience
• 4

  Revision history

• 1

  Overview
• 2

  Large language models
• 3

  Inferencing
• 4

  Retrieval augmented generation
• 5

  Model customization

• 1

  Architecture overview
• 2

  Hardware components
• 3

  Networking design
• 4

  Storage design
• 5

  Software components
• 6

  Security considerations
• 7

  Rack design
• 8

  Sizing and scaling the infrastructure for LLM inferencing and fine-tuning
• 9

  Summary

• 1

  Overview
• 2

  System configurations for validation
• 3

  Building the base container
• 4

  Inference validation
• 5

  RAG validation
• 6

  Fine-tuning validation
• 7

  Summary

• 1

  Benchmarking overview
• 2

  Memory bandwidth
• 3

  High Performance LINPACK for Accelerator Introspection
• 4

  ROCm Communication Collectives Library Benchmarking
• 5

  High-Performance Linpack
• 6

  Benchmarking caveats

• 1

  Overview
• 2

  Professional Services
• 3

  Customer Solution Centers (CSC)

• 1

  Summary
• 2

  We value your feedback

• 1

  Overview
• 2

  Dell Technologies documentation
• 3

  AMD documentation
• 4

  Broadcom documentation
• 5

  Other documentation

• 1

  Example configuration file