profiler

OLCF will host a two-hour interactive discussion on selecting profiling tools at different stages of performance analysis. The session includes a tool-agnostic overview of profiling strategies, brief introductions from participating tool vendors, and a moderated panel discussion. The event concludes with an open Q&A, giving participants the opportunity to discuss…

This session builds on Part One of the How to Profile series and focuses on analyzing and optimizing multi-process GPU applications using MPI. Participants will learn how to connect GPU and network performance insights to better understand application behavior and improve performance at scale.

This session introduces the fundamentals of profiling high-performance computing (HPC) applications using AMD profiling tools, with a focus on GPU-accelerated workloads. Participants will learn how to systematically analyze performance and identify opportunities for optimization. No prior profiling experience is required.