The ezfindexsubtree cronjob uses a (hardcoded) default object limit of 50 objects per commit.

On large indexes, the commit operation can take a long time to complete, thus causing the full index process to take considerably longer than necessary - particularly for very large subtrees.

It should be possible to specify, in some way (such as a .ini setting), the number of objects to "index" before a commit operation is performed (limit parameter).