Saltar al contingut Menu
Mapa
  • Inici
  • Informació
  • Contacte
  • Mapa

ALGORITMES I MODELS DE PROGRAMACIĶ PARALˇLELA (34325)

Crčdits Dept.
7.5 (6.0 ECTS) AC

Professors

Responsable:  (-)
Altres:(-)

Objectius Generals

In this course, students should gain an understanding of the concepts behind parallel programming models for current supercomputing architectures, based on a collection of distributed-memory nodes, each one with a shared-memory architecture. Nodes are built from current multicore chips and/or accelerators. The course will be very practical with simple parallelization assignments using different programming models (such as OpenMP, MPI, ...).

Objectius Específics

Coneixements

  1. Programming models: shared memory (OpenMP) and message passing (MPI).
  2. Trends in programming models for multicore and accelerator-based architectures.

Habilitats

  1. Ability to write parallel applications in mixed MPI/OpenMP.
  2. Identification of the most appropriate programming model for a given application.
  3. Simple modelling for parallel applications.

Competències

  1. Be able to parallelize a sequential application and tune it for a specific parallel architecture.
  2. Be able to work in group to discuss the use of different programming models and parallelization strategies for a given sequential application.

Continguts

Hores estimades de:

T P L Alt L Ext. Est A Ext.
Teoria Problemes Laboratori Altres activitats Laboratori extern Estudi Altres hores fora d'horari fixat

1. Basic concepts: process, data access and synchronization.
T      P      L      Alt    L Ext. Est    A Ext. Total 
2,0 0 0 0 0 2,0 0 4,0

2. Modeling a parallel application: from A to Z.
T      P      L      Alt    L Ext. Est    A Ext. Total 
4,0 4,0 0 0 0 6,0 0 14,0

3. Shared-memory programming models: OpenMP.
T      P      L      Alt    L Ext. Est    A Ext. Total 
4,0 4,0 12,0 0 15,0 4,0 0 39,0

4. Distributed-memory programming models: MPI.
T      P      L      Alt    L Ext. Est    A Ext. Total 
4,0 4,0 12,0 0 15,0 4,0 0 39,0

5. Issues in novel programming models for multicore architectures.
T      P      L      Alt    L Ext. Est    A Ext. Total 
8,0 4,0 0 0 10,0 10,0 0 32,0


Total per tipus T      P      L      Alt    L Ext. Est    A Ext. Total 
22,0 16,0 24,0 0 40,0 26,0 0 128,0
Hores addicionals dedicades a l'avaluació 0
Total hores de treball per l'estudiant 128,0

Metodologia docent

Classes building up concepts in a structured fashion and setting out the commitment required for their practical application. Classes focusing on a set of parallel programming models. Laboratory classes focusing on co-operative work in order to consolidate concepts, skills and competencies.

Mètode d'avaluació

The evaluation of the course will be based on the implementation and report presented for a set of practical assignments and a presentation and discussion of a bibliographical item.

Bibliografía bàsica

  • W. Gropp, W., Lusk, E. and Skjellum, A. Using MPI - Portable Parallel Programming with the Message Passing Interface, Mit Press, in Scientific and Engineering Computation Series. ISBN-13 978-0-262-57132-6, 1999.
  • Chapman, B. , Jost. G. and Van Der Pas, R. Using OpenMP - Portable Shared Memory Parallel Programming, Mit Press, in Scientific and Engineering Computation Series. ISBN-13 978-0-262-53302-7, 2007.

Bibliografía complementària

  • Other papers in international conferences and journals related to novel programming models proposals, , .

Enllaços web

  1. http://www.openmp.org/
    OpenMP specification.


  2. http://www-unix.mcs.anl.gov/mpi/
    MPI specification.


Capacitats prèvies

Understanding of parallel architectures, including both shared memory and distributed memory multiprocessor systems.


Compartir

 
logo FIB © Facultat d'Informàtica de Barcelona - Contacte - RSS
Aquest web utilitza cookies prōpies per oferir una millor experičncia i servei. En continuar amb la navegaciķ entenem que acceptes la nostra política de cookies.
Versiķ clāssica Versiķ mōbil