General

This course covers performance engineering approaches on the compute node level. Even application developers who are fluent in OpenMP and MPI often lack a good grasp of how much performance could at best be achieved by their code. This is because parallelism takes us only half the way to good performance. Even worse, slow serial code tends to scale very well, hiding the fact that resources are wasted. This course conveys the required knowledge to develop a thorough understanding of the interactions between software and hardware. This process must start at the core, socket, and node level, where the code gets executed that does the actual computational work. We introduce the basic architectural features and bottlenecks of modern processors and compute nodes. Pipelining, SIMD, superscalarity, caches, memory interfaces, ccNUMA, etc., are covered. A cornerstone of node-level performance analysis is the Roofline model, which is introduced in due detail and applied to various examples from computational science. We also show how simple software tools can be used to acquire knowledge about the system, run code in a reproducible way, and validate hypotheses about resource consumption. Finally, once the architectural requirements of a code are understood and correlated with performance measurements, the potential benefit of code changes can often be predicted, replacing hope-for-the-best optimizations by a scientific process.

Lecturers: Georg Hager and Jan Eitzinger, Erlangen National High Performance Computing Center, Bill Wiliams, Center for Information Services and High Performance Computing (ZIH)

Course date: June 3-5, 2025 (9:00 am - 4:00 pm) and June 6, 2025 (9:00 am - 12:15 pm)

This course will be conducted online as a Zoom event. Details will be sent vie e-mail to registered participants.

Course Outline:

Introduction

• Basic architecture of multicore systems: threads, cores, caches, sockets, memory
• The important role of system topology

Tools topology and affinity in multicore environments

• Overview
• likwid-topology and likwid-pin

Roofline model: basics

• Model assumptions and construction
• Simple examples
• Limitations of the Roofline model

Tools: hardware performance counters

• Why hardware performance counters?
• likwid-perfctr
• Applications

Roofline case studies

• Stencil algorithms
• Tall & Skinny dense matrix-matrix multiplication
• Sparse matrix-vector multiplication

Basic skills in performance engineering

Optimal use of parallel resources

• Single Instruction Multiple Data (SIMD)
• Cache-coherent Non-Uniform Memory Architecture (ccNUMA)

Extending Roofline: The ECM performance model

Performance Engineering using Score-P and Vampir

• Analyzing MiniMD
• Analyzing SpMV