Hochleistungsrechnen in den Ingenieurwissenschaften

What it is about

You will learn techniques of high-performance computing with a focus on numerical algorithms. You will be able to optimize sequential performance, scale programs for multi-core and parallel computers, and evaluate results using profiler tools and performance models.

What you will be able to do

• Characterize implementations of numerical algorithms
• Predict the performance of an algorithm based on hardware properties
• Verify achieved performance with profiler tools
• Classify differences between performance models and real execution
• Apply programming techniques for performant code

What the module consists of

• VorlesungDelivery of theoretical foundations and hardware properties; supported by script and presentations
• praktische Übungen/ProjektarbeitenInvestigation and optimization of algorithms on laptops and parallel computers; programming tasks with performance analysis

Teaching method

• Präsentation / VortragIntroduction of theoretical foundations and key concepts
• Skript als BegleitmaterialSupports the lecture and serves as a basis for own notes
• praktische Arbeiten an RechnernEnable examination of implementations and application of profiler tools
• Zusammenarbeit mit Leibniz-RechenzentrumPractical insights and use of parallel computing resources