I did some experiments and found that Neo4j performs better than TurboISO, which is the state of the art subgraph isomorphism algorithm.
I'm using Neo4j 3.5.11 community edition, and I tested 14 queries (match the subgraphs and return the vertices' IDs), the result is here:
Why Neo4j runs so fast?
It may be due to the index-free adjacency that Neo4j and other native graph databases use for relationship traversal, but it would help to know what queries you ran on Neo4j, and to get more detail on how TurboISO works, I'm not familiar with how they store data and traverse relationships.