Class Central is learner-supported. When you buy through links on our site, we may earn an affiliate commission.

YouTube

Resilience for AI Workloads at Scale - The Fast and the Finicky!

USENIX via YouTube

Overview

Coursera Spring Sale
40% Off Coursera Plus Annual!
Grab it
Explore the critical networking challenges that impact AI workloads at scale in this 25-minute conference talk from SREcon25 Europe/Middle East/Africa. Learn how AI workloads depend on network fabrics like RoCE and Infiniband for both high throughput and reliable communication paths, drawing parallels to Formula 1 racing where high-speed performance requires maintaining grip on the track despite potential obstacles. Discover how network "debris" such as NIC and link flaps can cause performance degradation, job crashes, and costly rollbacks that erode return on investment. Watch practical demonstrations showing how these networking problems directly impact the performance and reliability of AI jobs. Gain insights into strategies for overcoming these networking challenges to ensure AI workloads maintain peak performance, just like a Formula 1 champion's car powering through to the finish line despite track obstacles.

Syllabus

SREcon25 Europe/Middle East/Africa - Resilience for AI Workloads at Scale: The Fast and the Finicky!

Taught by

USENIX

Reviews

Start your review of Resilience for AI Workloads at Scale - The Fast and the Finicky!

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.