In our system there is a Schedule table which has two parts, one is being extensively updated (fields LockedBy,LockExpirationTime,NextDueTime), and second is mostly static (fields like Enabled,Name,SchedulePartition). In a such configuration does it make sense to split it actually in two different tables, with one to one mapping?
The update stored proc competes for schedules. It modifies only those dynamic fields, but takes into consideration static fields (such as enabled), currently it uses rowlock with readpast. There are other tables, which rely on static part of the table, thus there are many read queries to the static part.