Scheduling is a functionality that is never standardized and therefore it is hard to know what different companies are doing in their schedulers. In general, the point with CoMP is that the decisions made in different cells are not independent. You can for example schedule users in a way that limits the inter cell interference. Without CoMP, you might only make scheduling decisions that are good for the users in the own cell, without caring about intercell interference.
CoMP scheduling can be divided into two types - Centralized and Distributed based on the method of coordinating the eNBs. Usually the attached eNB acts as a master node in Centralized scheduling and other nodes in the eNB set serves as slave. Master ( or central) node performs the scheduling process. On the other hand, each eNB in the set performs the scheduling process in distributed scheduling approach according to the CSI feedback from their cooperating eNBs. Conventional schedulers do not coordinate with other eNBs.