Turbocharge your data center infrastructure with AMD
summary
AMD EPYC Rome and upcoming Milan CPUs use 7nm chiplet design to deliver 64 cores, 128 PCIe lanes, eight memory channels and up to 4 TB RAM, and AWS exposes them as M5a/R5a/T3a/C5a (including C5a bare-metal). AWS pricing typically shows about 10% cost savings vs. comparable instances (Mumbai announcement claims ~45% on some M5r types). AppStream 2.0 runs Radeon Pro GPUs with SR‑IOV. Densify’s analytics finds ~51% of workloads mis‑sized and automates safe migrations; Ample Organics reports equal or better performance and ~10% savings using Terraform (add “a”).
AI-written summary · about
chapters (32)
- 0:00Introduction
- 0:25Guest speakers
- 2:13Cloud market
- 2:59Market is growing
- 3:45People are spending more
- 4:32Poll Question
- 5:24Breakthroughs
- 8:01Where we are today
- 10:42Amazon with AWS
- 11:15C5A Instance Type
- 13:02Instance Types
- 16:57Availability
- 18:21Ecosystem Support
- 20:08Database Scalability
- 21:59Resource Optimization
- 23:06Ecosystem
- 26:02Hybrid Cloud Optimization
- 29:12Machine Learning Recommendations
- 32:08Behind the Scenes
- 36:58Hybrid Journey
- 37:50Customer Example
- 40:13Developers Perspective
- 41:45Continuous Optimization
- 43:35Declarative Optimization
- 44:33Full Upstream Self Service
- 45:38Savings Plans
- 47:06Machine Learning
- 49:40Ample Organics
- 54:38AMD AppStream
- 55:08Security
- 57:02AMD GPUs
- 57:44Summary
related talks
38mConquer cloud challenges with a competitive edge for less
47mHow AMD is taking the cloud to new heights, and what’s next
34mDeploy new workloads efficiently without additional investments
46mAMD + Azure: Outstanding performance and the latest security features
18mNext Gen Data Architecture for Multicloud
28mWhat's new in Amazon EC2
43mWhat’s new in Amazon EC2
16mChange Your Chip Architecture and Save the World