Hello, You might also try (using nightly build )using jsr223 éléments using groovy for example. In nightly build distrib, scripts passed as files are now compiled and cached which highly improves performances.
But of course it's not yet an official release. http://jmeter.apache.org/nightly.html Regards Philippe On Wednesday, July 18, 2012, sebb wrote: > On 18 July 2012 19:17, Anthony Johnson <ans...@gmail.com <javascript:;>> > wrote: > > Hey Jacky, > > If you are using the Beanshell Pre-processor and doing the 200 > > threads that you noted in another thread... you may be hitting a CPU > > bottleneck on your machine. > > > > Some options that might help: > > > > Set Reset Interpretor to False (so that the beanshell instance can be > re-used. > > Use less threads > > Use a faster machine or execute your test on more than 1 machine via > > Remote Testing. > > It may help to use the BeanShell init script to define suitable > methods, as that only has to be parsed once. > > > Hope this helps, > > > > Anthony > > > > On Wed, Jul 18, 2012 at 2:07 PM, unjc email > > <unjc.em...@gmail.com<javascript:;>> > wrote: > >> I need to use a preprocessor that allows script. I know > >> user-parameter is okay but I can't put logic in it. > >> > >> > >> On Wed, Jul 18, 2012 at 12:54 PM, Robin D. Wilson > >> <rwils...@gmail.com<javascript:;>> > wrote: > >>> Does it matter which preprocessor you use? I use the User Parameters > preprocessor (a lot), and I haven't noticed this. > >>> > >>> -- > >>> Robin D. Wilson > >>> Sr. Director of Web Development > >>> KingsIsle Entertainment, Inc. > >>> VOICE: 512-777-1861 > >>> www.KingsIsle.com > >>> > >>> > >>> -----Original Message----- > >>> From: unjc email [mailto:unjc.em...@gmail.com <javascript:;>] > >>> Sent: Wednesday, July 18, 2012 11:09 AM > >>> To: JMeter Users List > >>> Subject: Bad Throughput with Preprocessor > >>> > >>> Hello all, > >>> > >>> I want to test with dynamic-generated Url. I found there is a huge > >>> drop in throughput when I use preprocessor (even an empty > >>> preprocessor) for the HTTP sampler. The difference is as much as ten > >>> times in my test. Do any of you have any suggestion what to use if I > >>> have to implement some logics for the sampler? Any advice will be > >>> greatly appreciated. > >>> > >>> > >>> > >>> Thanks, > >>> Jacky > >>> > >>> --------------------------------------------------------------------- > >>> To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org<javascript:;> > >>> For additional commands, e-mail: user-h...@jmeter.apache.org<javascript:;> > >>> > >>> > >>> --------------------------------------------------------------------- > >>> To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org<javascript:;> > >>> For additional commands, e-mail: user-h...@jmeter.apache.org<javascript:;> > >>> > >> > >> --------------------------------------------------------------------- > >> To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org<javascript:;> > >> For additional commands, e-mail: user-h...@jmeter.apache.org<javascript:;> > >> > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org<javascript:;> > > For additional commands, e-mail: user-h...@jmeter.apache.org<javascript:;> > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org <javascript:;> > For additional commands, e-mail: user-h...@jmeter.apache.org<javascript:;> > > -- Cordialement. Philippe Mouawad.
