Differences

This shows you the differences between two versions of the page.

--- paga:index [2012/10/16 08:28] – moreno
+++ paga:index [2013/03/27 12:41] (current) – gdangelo
@@ Line 1: / Line 1: @@
 ======  PArallel Graph Algorithms (PAGA)  ======
-//University of Bologna. Department of Computer Science//
+//University of Bologna. Department of Computer Science and Engineering (DISI)//
 Welcome to the PAGA homepage.
@@ Line 6: / Line 6: @@
 The goal of this project is to design, implement and evaluate parallel algorithms for large-scale graph analysis. In particular, we are interested in scalable approaches for computing centrality metrics on distributed memory architectures. Centrality metrics such as betweenness centrality, clustering coefficient, closeness centrality and others, are widely used in social network analysis.
-The PAGA project will last from 2012/10/02 to 2013/07/02.
+The PAGA project is supported by an Italian Supercomputing Research Allocation (ISCRA) project HP10CR15I1 from 2012/10/02 to 2013/07/02. PAGA has been granted the use of [[http://www.hpc.cineca.it/hardware/ibm-bgq-fermi|FERMI]], an [[http://www.research.ibm.com/bluegene/index.html|IBM BlueGene/Q]] supercomputer installed at [[http://www.cineca.it/|CINECA]]. In June 2012, FERMI ranked at the 7th position in the [[http://www.top500.org/list/2012/06/100|Top500 list]].
 === Background ===
@@ Line 20: / Line 20: @@
 === Expected Outcomes ===
-Our primary goal is to understand if and how centrality metrics can be effectively computed on distributed memory architectures. A positive answer will be a significant contribution to the research community. Our secondary goal is to laid the foundations for the development of a software package for social network analysis on distributed memory architectures based on MPI. This package should run both on general-purpose, inexpensive clusters of PCs, and also on high performance parallel machines. The use of MPI will increase the (potential) user base for our package, and consequently the impact of this project within the scientific community.
+Our primary goal is to understand if and how centrality metrics can be effectively computed on distributed memory architectures. A positive answer will be a significant contribution to the research community. Our secondary goal is to lay down the foundations for the development of a software package for social network analysis on distributed memory architectures based on MPI. This package should run both on general-purpose, inexpensive clusters of PCs, and also on high performance parallel machines. The use of MPI will increase the (potential) user base for our package, and consequently the impact of this project within the scientific community.