The documentation on the new initializeOrderedBulkOp command needs some further detail on how the bulk operations are actually processed. In particular, "ordered" operations have a performance impact depending on how the operations are interleaved (consecutive operations that are the same can be be more effectively batched).

A few blog posts go into more helpful detail:

• http://christiankvalheim.com/post/the_new_bulk_api/
• http://blog.mongohq.com/better-bulking-for-mongodb-2-6-and-beyond/.