Richard, That patch solves the performance issue, now it works fine without any locking problems. Thank you very much.
-----Original Message----- From: Richard Frovarp [mailto:[email protected]] Sent: Monday, November 30, 2009 11:47 PM To: Maxim Borkunov Cc: [email protected] Subject: Re: Flow script performance issue Maxim Borkunov wrote: > Here is my performance benchmark result. > > User/AVG time(s) > Test case 1 25 50 75 100 > 1 Analytical Research (XHTML) 0.112 2.052 4.228 6.25 7.665 > 2 Consortium Communicati..(XHTML) 0.111 1.967 3.944 5.815 8.544 > 3 FSS Inbound Template (PDF) 0.009 0.024 0.031 0.023 0.015 > 4 FSS STMT Inbound Templ (PDF) 0.012 0.023 0.026 0.021 0.022 > 5 LAS OBL Inbound Templ (PDF) 0.009 0.026 0.031 0.013 0.016 > 6 Home (XHTML) 0.228 2.382 5.209 8.401 10.338 > 8 LAS LOANS Inbound Templ (PDF) 0.01 0.019 0.031 0.014 0.019 > 9 Continuation(login page)(flow) 0.124 7.416 61.222 229.407 503.981 > 10 Login (flow) 0.181 11.313 85.959 314.342 670.145 > 11 Logout (flow) 0.066 10.799 58.368 156.67 376.282 > 12 Methodology and process (XHTML) 0.13 2.243 4.431 6.384 7.992 > 13 Project Plan (XHTML) 0.119 1.8 3.821 6.185 7.85 > 14 Publications for RB (XHTML) 0.105 1.84 4.671 8.754 18.135 > 15 Publications for SAAB (XHTML) 0.108 2.045 4.541 8.215 14.352 > 16 Reports for RB (XHTML) 0.116 1.97 4.227 6.269 7.491 > 17 Reports for SAAB (XHTML) 0.117 2.31 4.018 5.965 7.906 > 19 Search for alerts (Search) 0.206 2.926 10.993 19.228 47.919 > 20 Search for something (Search) 0.207 3.201 20.379 61.375 109.367 > 21 Search for money (Search) 0.268 4.765 27.578 89.302 165.854 > 22 Search for process (Search) 0.214 2.578 8.425 15.708 28.589 > 23 Search for research (Search) 0.214 2.616 7.736 11.817 19.786 > 24 Send message - my page (flow) 0.148 12.465 71.613 234.497 510.283 > 25 Deadlines (Custom generator) 0.111 2.003 4.115 6.478 8.812 > > Login page generation, logout process and every page with flow script usage > is taking too much time. > Thank you for your results. I did come across a problem with massive locking issues with many of the pages you list as having problems. Unfortunately it did not make it into 2.0.2. You could try making the change yourself and re-running your benchmarks to see if it has the desired effect. It involves removing the importPackages line from usecases.js. The commit can be found here: http://svn.apache.org/viewvc/lenya/trunk/src/modules-core/usecase/usecases.js?r1=730252&r2=730251&pathrev=730252 --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
