This completes the core of the GraphLab tutorial. We have went through an overview of

How to start a GraphLab project
How to read a graph from disk/HDFS
How to write a vertex program
Dynamic Scheduling in a vertex program
How to save a graph to disk/HDFS

There are many more features which we are unable to introduce through the course of this tutorial which we hope you will be able to discover by exploring the documentation.

Some really useful tools that we would like to bring to your attention are:

Perform MapReduce over the vertices or the edges in the graph.
- graphlab::distributed_graph::map_reduce_vertices()
- graphlab::distributed_graph::map_reduce_edges()
Perform MapReduce over the vertices or edges in the graph, while being provided a context in the Map function, thus allowing finer grained control over signalling.
- graphlab::iengine::map_reduce_vertices()
- graphlab::iengine::map_reduce_edges()
Make a modification to all the vertices or edges in the graph
- graphlab::distributed_graph::transform_vertices()
- graphlab::distributed_graph::transform_edges()
Make a modification to all the vertices or edges in the graph, while being provided a context in the Map function, thus allowing finer grained control over signalling.
- graphlab::iengine::transform_vertices()
- graphlab::iengine::transform_edges()
Register a MapReduce operation which performs periodically while a GraphLab engine is running thus allowing for global state.

Interleaving these operations together with GraphLab vertex_programs allow for a huge amount of flexibility, allowing for a large number of algorithms to be implemented easily, and efficiently.