Another way to achieve this is to ensure that the test runs for long enough to ensure that all the logins have taken place.
This allows one to measure the performance once the test has stabilised. Test runs should generally be long enough to ensure that any start-up overhead is minor compared with the main part of the test. E.g. if it takes 1 minute for all the logins to complete, run the test for at least 10 minutes. On 29/02/2008, Kam <[EMAIL PROTECTED]> wrote: > I'm testing a web application which involves multiple thread groups; > one thread group for each type of role a user can hold in it. I need > the users in all thread groups to complete their logins first, before > any requests can start to be sent. So I need to test performance > independent of login performance. I understand that the Synchronizing > Timer is not supposed to work across Thread Groups. However when I > put a Synchronizing Timer, as a child of a Test Action sampler, after > each login for each of the individual thread groups, all threads stop > at the Synchronizing Timer until they all finish the logins. This > looks like the desired outcome I want, but I'm not sure it is an > accurate way or clean way to be capturing performance independent of > login performance. I put the total number of users (or threads) in a > "Number of Simulated Users to Group by" parameter. Does anyone > believe that using the Synchronizing Timer across Thread Groups in > this manner is problematic? > > > > > ____________________________________________________________________________________ > Be a better friend, newshound, and > know-it-all with Yahoo! Mobile. Try it now. > http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
