Permalink
Browse files

Integration of the HDRF algorithm within PowerGraph.

HDRF is a novel stream-based graph partitioning algorithm that provides important performance improvements with respect to existing solutions in partitioning quality.
In particular, HDRF provides the smallest average replication factor with close to optimal load balance. These two characteristics put together allow HDRF to significantly reduce the time needed to perform computation on graphs and makes it the best choice for partitioning graph data.

The HDRF algorithm is extensively described in the following publication:

F. Petroni, L. Querzoni, K. Daudjee, S. Kamali and G. Iacoboni: 
"HDRF: Stream-Based Partitioning for Power-Law Graphs". 
CIKM, 2015.
http://www.dis.uniroma1.it/~midlab/articoli/PQDKI15CIKM.pdf
  • Loading branch information...
fabiopetroni committed Jul 29, 2015
1 parent eae2018 commit 50ec06b0a4879afc993a2cc189c8fca02e4c9a78
Showing with 1 addition and 1 deletion.
  1. +1 −1 src/graphlab/graph/distributed_graph.hpp
@@ -191,7 +191,7 @@ namespace graphlab {
* runtime memory consumption.
*
* \li \c "grid" Runs at rouphly the same speed of random. Randomly places
* edges on machines with a grid constraint.
* edges on machines with a grid constraint.
* This obtains quality partition, close to oblivious,
* but currently only works with perfect square number of machines.
*

0 comments on commit 50ec06b

Please sign in to comment.