I just ran the test with non-Identical insertions and performance was as I stated before. (The identical insertions were due to me trying to simplify the example to fit better in a forum post. My initial tests were more robust than the examples I posted. (Alerts were external objects having unique ids and db representation, db persistance and other non-rule times were excluded from my statistics. I just tried to make objects simple for the forum discussion of performance)
I currently run rules in batches via runAllRules because my previous performance tests had runAllRules running in a separate thread outperforming runUntilHalt. I have not repeated that test recently and it may no longer be true. I will try to check that again if time allows. Either way the unit tests use runAllRules because it allows for testing from a known state to another known state with detailed tracking of time involved required by the rule processing. -- View this message in context: http://drools.46999.n3.nabble.com/rules-users-Performance-scaling-with-fact-insertion-tp3727727p3733352.html Sent from the Drools: User forum mailing list archive at Nabble.com. _______________________________________________ rules-users mailing list [email protected] https://lists.jboss.org/mailman/listinfo/rules-users
