Parallel Programming

A laboratory course providing practical training in high-performance programming and parallel programming techniques:

• Basic concepts of parallel computing

• The impact of memory hierarchy on performance

• SSE instructions

• Programming with POSIX Threads

• Thread programming with OpenMP

• Introduction to CUDA programming

Operating system: Linux. Programming languages: C and C++. Special-purpose languages for GPU programming such as CUDA C/C++ are occasionally used.

Upon successful completion of the "Parallel Programming" course, undergraduate students will be able to:

• Understand the advanced organization of a computer and the relationship between architecture and parallel processing.

• Autonomously follow further developments in the field of parallel programming.

• Select the appropriate form of parallel hardware according to the needs of each application.

• Transform a parallel algorithm into the appropriate parallel program.

1st week:
Introduction to parallel processing and parallel computer architectures.

2nd week:
The memory hierarchy (main and cache memories) and its impact on performance.

3rd week:
Laboratory exercises on the impact of memory hierarchies on processing.

4th week:
Programming with SSE (Streaming SIMD Extensions) instructions.

5th week:
Laboratory exercises in programming with SSE instructions.

6th week:
Programming with POSIX Threads (pthreads).

7th week:
Programming with POSIX Threads (pthreads) (II) – synchronization methods.

8th week:
Laboratory exercises in programming with POSIX Threads.

9th week:
Thread programming with the OpenMP platform.

10th week:
Thread programming with the Intel Threading Building Blocks (TBB) platform.

11th week:
Tasks and pipelines in the Intel Threading Building Blocks (TBB) platform.

12th week:
Parallel programming on GPU – the CUDA model.

13th week:
Laboratory exercises with the CUDA model.