The answer to this post (find distinct paths in neo4j database) recommends using the Traversal API to identify distinct paths in a graph (please see original question) instead of a Cypher query. For a graph containing 10,000 nodes the API returned the results immediately; much faster than the Cypher query.
I would like to create the same results for a graph containing more than 100 mio. nodes. Based on a simple progress check I added to the code, identifying the distinct paths might well take several hours. The database files are in total ca. 500 MB. Memory does not seem to be a problem, but CPU is running at 12% (on an 8-core machine) => it looks like the process is CPU-constrained. Is there any way to make the process run in parallel?
I am using Neo4j 2.1.3, JRE 1.7.0_51-b13 (64-Bit), runnning on Windows 8.