The performance engineering courses at NHR@FAU cover node-level hardware analysis and optimization in depth, using established tools and performance models. Topics range from microarchitecture fundamentals and cache behavior to solver-level performance analysis, and are frequently taught in collaboration with performance engineering experts from partnering institutions.
- 2026, Jul 20: Practical Roofline Analysis by Example
online, full-day
Registration Link - 2026, Aug 30: Practical Roofline Analysis by Example
in Poznan, Poland, full-day - 2026, Oct 5: Core-Level Performance Engineering
online, full-day
Registration Link - 2026, Dec 1-3: Node-Level Performance Engineering
online, three-day
- Core-Level Performance Engineering
Performance engineering at the CPU core level, covering x86 and ARM architecture, assembly code analysis, and the OSACA tool. - Introduction to the LIKWID Tool Suite
Introduction to the LIKWID command-line performance tool suite for x86, ARM, and GPU architectures on GNU/Linux. - Node-Level Performance Engineering
Performance engineering at the compute node level using the Roofline model and hardware bottleneck analysis for modern CPUs. - Performance Engineering for Linear Solvers
Performance modeling and optimization for iterative linear solvers on CPU and GPU using the Roofline model. - Practical Roofline Analysis by Example
Hands-on introduction to the Roofline model for identifying hardware bottlenecks and guiding code optimization on CPUs and GPUs.
For an overview of all NHR@FAU courses, visit the course overview page.