JRE 1.6.0_7 vs 1.6.0_18

[Orest Ivasiv]

Hi all,
I've done benchmark and test my application on various JRE (see subject). The most interesting thing that I've discovered real huge performance degradation in JRE 1.6.0_18.
E.g. There was selected complex use case and test various JRE's 1.6. And the most interesting thing it that this use case takes only ~1 min on JRE 1.6.0_7 and more that ~2 min on JRE 1.6.0_18.

Small note about this legacy application: it's applet as presentation tier and back-end on Tomcat (doesn't matter the JRE on server side, because I've got only this performance degradation on client side and checked different JRE combination both client/sever sides).

This testing was performed on the next environment: 1) client PC; 2) tomcat server; 3) db server - Oracle - three physical machine. No changes except version of JRE on client machine. I guess there are some huge changes in Java plug-in.

Could you please explain such BIG difference in time execution?

Thank you in advance,
Orest

[Kripal Singh]

try testing the app with different jre on 2 different client PC ( with fresh install ) . also make sure you are using same browser . probably nothing to do with your app but client OS, browser and jvm combination .

[Ernest Friedman-Hill]

Do you have any idea where the time goes in either version? i.e., what does the client do that takes one minute? Is it CPU limited, or I/O (network) limited? Have you done any profiling to find out? (Guessing doesn't count -- programmers are notoriously bad at guessing performance bottlenecks!)

[Kripal Singh]

it might make sense to attach applet process to perfmon or java profiling tool to see what's going on . also open java console and see what's going on . check sun bug reports or google you issue