Naslov
Optimizing Sequential Programs for Parallel Execution

Autori
Martinović, Goran ; Krpić, Zdravko

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Proceedings of 27th International Conference "Science in Practice" SiP 2009 / Kvasnicza, Zoltan ; Elmer, Georg ; Megyeri, Peter - Pečuh : Institute for Information Technology and Electrical Engineering of the University of Pécs Pollack Mihály Faculty of Engineering, University of Pecs, 2009, 41-44

Skup
27th International Conference "Science in Practice" SiP 2009

Mjesto i datum
Pečuh, Mađarska, 26.02.2009. - 27.02.2009

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
parallel computing; computer cluster; heterogeneous systems; MPI; load balancing

Sažetak
In engineering program implementations there is always need for more computer resources. The most common are applications which require lot of processor power or memory space. Generally, programs which are too demanding, even when fully optimized, often need too much of execution time. Thereby, in many cases is often better (cheaper) to optimize existing algorithm then to write a new one. Various data structures, loops and iterations can be parallelized efficiently without rewriting the whole program code. One of the major parallel library these days is MPI, because it's simple, platform independent and robust. In this article some of the most common parallelizing methods will be presented, so as their advantages in terms of cost-effectiveness and simplicity. This research will also cover situations where complete new program code is a better choice.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo

POVEZANOST RADA