While it is true that for a given schedule the nodes associated
with that schedule are picked for randomization in a "last in first
out" basis, that in no way effects the starttime given for each node.
There is no tracking of or comparison of what each node is given as a
starttime. So the node selected first from the DB may have a later
starttime than a node selected second from the DB. The only limiting
factor is if your MAXSESSIONS or MAXSCHEDSESSIONS is less than the number
of nodes to be processed for a given schedule.
 There are no changes for this in V3.

Tom Brooks
ADSM Development