Master AI and Machine Learning: From Neural Networks to Applications
UC San Diego Product Management Certificate — AI-Powered PM Training
Overview
Google, IBM & Meta Certificates – 40% Off
One plan covers every Professional Certificate on Coursera.
Unlock All Certificates
Explore a 37-minute conference talk from SNIA Storage Developer Conference 2024 examining how Infrastructure Processing Units (IPUs) can be deployed as standalone servers to create efficient, scalable clusters. Discover the evolution from traditional PCIe endpoint deployment to a new paradigm where IPUs function as complete servers with PCIe root-port interfaces, embedded Arm cores, and Linux OS capabilities. Learn how to configure IPUs for multiple storage use cases and build cost-effective, energy-efficient server clusters optimized for scale-out applications like Apache Cassandra NoSQL database. Gain practical insights into performance tuning techniques for minimizing database tail latency while maintaining high throughput, particularly relevant for AI/ML applications requiring increased storage capacity. Presented by Intel's Ian Wigmore, delve into the technical requirements for developing and managing multi-node IPU clusters, with specific focus on optimizing Cassandra p99 tail latency and throughput in IPU environments.
Syllabus
SNIA SDC 2024 - Scaling Out IPUs to Create a Server Cluster
Taught by
SNIAVideo