PRIMJENA PARALELNIH RAČUNALA
(Applications of Parallel Computers)
Syllabus
Lectures:
 Sources of parallelism in realworld problems.
 Parallel solution of linear systems and basic matrix
factorizations.
 Iterative solution of linear systems.
 Parallel sorting and searching.
 Parallel graph algorithms and graph partitioning.
 Parallel solution of nbody problems.
 Parallel eigenvalue computation.
 Sparse linear systems.
Tutorials:
 Introduction to programming on graphics processors with CUDA.
 Managing scientific data with HDF5.
Materials

Svojstveni problem
( pdf )

Jacobijev algoritam
( pdf )

Dekompozicija domene
( pdf )

Metoda konjugiranih gradijenata
( pdf )

Bitonički sort
( pdf )
Course information and additional materials:
DeGiorgi Forum.
TA's (Neven Krajina) support pages:
here.
CUDA student projects
 Dejan Peretin: Doubly Compensated Summation.
( zip )

Ivan Stojić: Jacobi iterations for linear systems.
( zip )