Brussels / 2 & 3 February 2013


Signal/Collect: Processing Large Graphs in Seconds

You should attend this session if you are interested in running complex algorithms on graphs with billions of vertices and edges quickly. You will learn about the ideas behind the Signal/Collect programming model, some benchmarks of our distributed system and how to implement your own algorithms. During the live demo at the end I implement and run an example algorithm using our open source system.

We have some pretty impressive new results (PageRank on Yahoo Webgraph with 1.4 billion vertices and 6.6 billion edges in a bit more than 2 minutes). I have been working on Signal/Collect as part of my PhD for the past 2.5 years and am its main contributor. I presented the parallel version of the model/system at the International Semantic Web Conference (ISWC) 2010 in Shanghai.


Philip Stutz