[rules-users] Drools slowness in first calls
Hi, We are using Drools for a heavy process and we notice some slowness during the first calls to kbases (containing about 1500 rules each) using StatefulKnowledgeSession. Even if the kbase is loaded, we need to make some fictive calls (fire the rules with some facts) until it begins to respond within a reasonable time. It's like we need to warm it up.. Any idea about what Drools does during this time? Is there a way to programmatically launch these processing without the need to have facts ? Thanks in advance and sorry for my english. NB : we use Drools 5.5. Regards, Gael -- View this message in context: http://drools.46999.n3.nabble.com/Drools-slowness-in-first-calls-tp4029887.html Sent from the Drools: User forum mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Drools slowness in first calls
On 5 Jun 2014, at 16:48, Gael gael.weil-jour...@prestataire.april.com wrote: Hi, We are using Drools for a heavy process and we notice some slowness during the first calls to kbases (containing about 1500 rules each) using StatefulKnowledgeSession. Even if the kbase is loaded, we need to make some fictive calls (fire the rules with some facts) until it begins to respond within a reasonable time. It's like we need to warm it up.. That’s sort of how the JVM works, it takes many thousands of iterations for JIT byte code to kick in. There is also some additional byte code generation we do on the fly to speed things up - you may also be seeing this. Any idea about what Drools does during this time? Is there a way to programmatically launch these processing without the need to have facts ? Not really, this is how the JVM works, likewise the additional byte code generation we do. In 6.x there is an additional one off cost too, in calculating all the node memories and segments. This also is done “on the fly”, during the first session. Mark Thanks in advance and sorry for my english. NB : we use Drools 5.5. Regards, Gael -- View this message in context: http://drools.46999.n3.nabble.com/Drools-slowness-in-first-calls-tp4029887.html Sent from the Drools: User forum mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Drools slowness in first calls
Google: JIT Java -W On 05/06/2014, Gael gael.weil-jour...@prestataire.april.com wrote: Hi, We are using Drools for a heavy process and we notice some slowness during the first calls to kbases (containing about 1500 rules each) using StatefulKnowledgeSession. Even if the kbase is loaded, we need to make some fictive calls (fire the rules with some facts) until it begins to respond within a reasonable time. It's like we need to warm it up.. Any idea about what Drools does during this time? Is there a way to programmatically launch these processing without the need to have facts ? Thanks in advance and sorry for my english. NB : we use Drools 5.5. Regards, Gael -- View this message in context: http://drools.46999.n3.nabble.com/Drools-slowness-in-first-calls-tp4029887.html Sent from the Drools: User forum mailing list archive at Nabble.com. ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users