-
Theory
- Example of where Parallel Programming is necessary
- Theory of PP performance Analysis : intro to Ahmdahl's Law (Strong Scaling) and Gustafson's Law (Weak Scaling)
-
Languages/Libraries/Directive that can be used OPENMP, MPI GPU(not covered)
- Shared(OPENMP) Memory Programming Model
- Introduce basic example of openmp ("Hello Word")
- Examples:
- serial to parallel matrix multiplication
- reduction example: Numerical Integration
- (Scheduling example: SpMV, show performance depends on matrix structure -> scheduling)
- (ccNUMA example: 2D heat with Jacobi, Show performance depends on the data layout when using multi-socket system)
- Final Example: Mandelbrot Set
- Examples:
- Introduce basic example of openmp ("Hello Word")
- Shared(OPENMP) Memory Programming Model
-
Distributed (MPI) Memory Programming Model (Maybe only for advanced group ? and basic example for beginner group ? )
- Message Passing Interface (MPI - Importance of communication
- Introduce basic example of OpenMPI ("Hello Word")
-
Examples:
- serial to parallel matrix multiplication
- reduction example: Numerical Integration
-
Final Example: MandelBrot Set
-
- "Hof. Moritz Travis" moritz.hof@dlr.de
- "Overmars. Nigel" nigel.overmars@dlr.de
created Edited: Moritz Hof, 07.10.2020 last Edited: Moritz Hof, 16.10.2020