COMP10052-AAAF-10-2010-for-viewing

COMP10052-AAAF-10-2010-for-viewing

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: G O D   L A’ T A’’ Note that, in this case (with respect to the diagram above):     ALGO = A = L = G = O (i.e., each is an instance of the same piece of code); DATA = D ∪ A’ ∪ T ∪ A’’ (i.e., those instances each act on a parBBon of a single data set). DistribuBng EssenBally ­Centralized Systems (2) 4       In this case, our moBvaBon to design and engineer this applicaBon as a distributed system is efficiency, not effecBveness. In other words, we could just as well run the code in a single machine without parBBoning the data set: it would just take much longer to do so. The computaBon is essenBally a centralized one: we distribute it in order to improve its performance.     Recall that we can refer to this parallelizaBon strategy as a case of divide ­and ­ conquer. Recall examples such as:       Rendering mulBple frames of high ­quality animaBon; SimulaBng thousands of characters; Indexing the web (e.g., Google, using something...
View Full Document

This document was uploaded on 02/10/2014.

Ask a homework question - tutors are online