concurrent paths of execution don't have to run in separate threads if you're using persistence. in persistent mode, you go from one state to another. in case of a fork, this is one token before the fork when the tx starts to multiple tokens after the fork when the tx ends. in persistent mode, the number of threads that you use to calculate that tx is irrelevant. the important thing is that different transactions are isolated from each other. single threaded fork calculation is easier for persistence. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4111132#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...