Thank you for you answer, Peter! What I wanted to find out is, among other things, is what constellations make absolutely no sense - because of network limitaions, memory limitations, JMeter hard-coded stuff etc. Sure I'm having a lot of trial & error upcoming, but I can gain by limiting it :-) NL
On 1/3/06, Peter Lin <[EMAIL PROTECTED]> wrote: > > this question is asked frequently. > > it all depends on the think time between requests. I've easily ran tests > with 250threads with no think time for static files in tomcat without any > problems. the best way is to try 50, 100, 150, 200 threads and see which > produces the max throughput. > > hope that helps > > peter > > > On 1/3/06, Nop Lists <[EMAIL PROTECTED]> wrote: > > > > Hello, > > I am setting a large scale stress test and want to use as few client > > JMeter-PC as possible. > > Does anyone have an idea - or better say a rough formula (depending e.g. > > on > > memory, CPU etc.) - what could be a maximal count of threads for a PC? > > Thanks for your ideas and stimulations > > NL > > > > > >
