Hi, I think there's a bit of confusion here and I agree this is caused by not exhaustive documentation (we are trying to improve it).
The permgen threshold option is used for constraint jitting and that one works exactly in the same way regardless of the dialects. However when using the java dialect the permgen occupation grows not because of constraint jitting but because we have to generate java classes to implement the consequences. To cut it short permgen threshold and dialects are 2 orthogonal things, but it's true that keeping the threshold to 0 and using the mvel dialect is the best combination to reduce to the minimum the permgen occupation, because the first prevents the generation of java classes for constraints jitting while the second has the same effect for what regards consequences. Nevertheless there's a performance drawback in doing this because not allowing constraint jitting means that they will be always evaluated in interpreted mode, while for the same reason mvel consequences are a bit slower than java ones. For this reason my personal advice is to look for the best compromise between permgen occupation and performances for your specific use case. I hope this helps, Mario -- View this message in context: http://drools.46999.n3.nabble.com/Drools-Perm-gen-grows-constantly-tp4026673p4026810.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