Credits: 9

Type: professional

Objectives:

To examine the principle techniques used to maximise the possibilities offered by parallel processing systems. Special emphasis will be laid upon the possibilities offered by PC networks interconnected by fast networks.

Knowledge:

• Hardware. SIMD and MIMD models. Local and shared memory. Processor interconnection schemes.
• Basic vectorial arithmetic: matrix-vector products and matrix-matrix products; band matrices.
• Parallel computing tools. Message passing libraries. PVM and MPI library specifications.
• Resolution of linear systems. Direct and iterative methods. Conjugate gradient. Preconditioners. Applications.

Subject-specific competences:

• To be able to implement parallel algorithms.
• To be able to numerically resolve some Physics problems using parallel computing.