NVIDIA & Marlowe: Workload Performance and Health Monitoring in Data Centers

Abstract

This session explores how to efficiently run and monitor large workloads in hyper-scalars and the kind of telemetry data we can monitor from the Infra and workload side. We will review various tools offered by NVIDIA to monitor the performance of a workload and how they improve stability with increasing scale.

Special Speaker | Aparnaa Ramani

Aparnaa Ramani is an AI Solution Architect working closely with leading AI labs. She collaborates closely with them on optimizing workloads and data center bring-ups with the latest NVIDIA architecture. Before joining NVIDIA, she worked as a GPU kernel developer at Qualcomm. Aparnaa has a Master's degree from Georgia Tech with focus on statistical ML and Advanced DSP.