Parallel programming with OpenMP (7 hours)

OpenMP provides a very convenient programming model for scientific applications that run multiple threads. It is supported by C, C++ and Fortran compilers. Its main advantage is that it is mostly used through annotation of code, which makes it easy to pick low hanging fruit when starting to parallelize code. The training covers OpenMP 4.5, and emphasizes pitfalls and best practices.

Subjects:

• programming and execution model
• worksharing directives
• data environment
• tasking
• SIMD constructs
• thread placement
• pitfalls and best practices

Prerequisites:

• experience in C, C++, or Fortran.