The course is included in the specialization part of the degree and its main goal is to give competences in the area of High Performance Computing; i.e. to the methodologies and techniques that allow the exploitation of hardware architectures with multiple processors and heterogeneous, in order to reduce the execution times of programs that need high computational resources. The course also takes an hands-on approach with laboratory and project activities.

Knowledge and understanding goals

• Multiprocessor and heterogeneous architectures (CPUs and GPUs)
• Techniques for problem decomposition
• Languages, libraries and tools needed for all the development cycle of a parallel program

Know-how goals

• Performance optimization of a sequential program
• Ability to use tools for measuring the performance of symmetric multiprocessors and clusters
• Multiprocessor programming using C/C++ OpenMP, MPI and CUDA from Nvidia

Soft-Skills:

• Critical reasoning and evaluation on alternative solutions and approaches for problem solving.

1. Parallel computing: hardware, software, applications and performance theory.
2. Programming shared-memory multiprocessors.
3. GPU computing.
4. Programming distributed-memory multiprocessors.

Main:

• P. Pacheco, " An Introduction to Parallel Programming", Morgan Kauffman, 2011
• CUDA C Programming Guide, NVIDIA corporation

Complementary:

• Norm Matloff, "Programming on Parallel Machines: GPU, Multicore, Clusters and More", University of California, Davis,

Students should have knowledge about computer architecture, computer networks, and operating systems, and good programming skills.