Yes it can be the reason if you have some fixed times between the block. Try with 2*1 users and check the differences between gui and non-gui model, should be lower than for 2*40 user.
Best regards, Jakub 2009/10/19 Olofsson, Lars <[email protected]>: > Hi > Thruput is about the same, but that is probably becaus of of the long "Think > Time", betwen each block, about 3-6 seconds > We have 2*40 users simulated, so I guess it can generate 2*40 request faster > in non-gui, but then it has to wait for ThinkTime, until next burst. > > Br > Lars Olofsson > > -----Original Message----- > From: Jakub Białek [mailto:[email protected]] > Sent: Monday, October 19, 2009 8:52 AM > To: JMeter Users List > Subject: Re: Big difference between non-gui and gui result ? > > Hi, > > how about throughput? Does it grow up in non gui mode? If yes then > explanation is that: non gui jmeter can generate more requests in the same > time than gui jmeter so the target server load is higher and response time > become longer. > > Best regards, > Jakub > > 2009/10/19 Olofsson, Lars <[email protected]>: >> Clarification... >> >> All time seems to be the same for each Request/Response, botg in guin and >> non-gui. >> >> But all summary times for each block are 10-15 % higher in non-gui compared >> to gui-run ! >> >> I only checked/compare these summary times. >> >> Why does this happens ? >> I know there are processing time included in summary times, but why >> Should this be higher for non-gui ? >> >> Best regards >> Lars "newbie" Olofsson >> >> -----Original Message----- >> From: sebb [mailto:[email protected]] >> Sent: Thursday, October 08, 2009 11:47 PM >> To: JMeter Users List >> Subject: Re: Big difference between non-gui and gui result ? >> >> On 08/10/2009, Olofsson, Lars <[email protected]> wrote: >>> Hi >>> >>> I am new to jMeter and to this list ? >>> Is it normal that the average response difference a lot ( twice as h >>> high) between a non-gui and an gui run. >>> We have 5 sec average on gui and 10 sec on non-gui run. >>> 1 Testcase, 2*40 users, 2 jmeter-servers. >>> >>> I just wonder if this big difference is normal and can be explained >>> by the the fact that in non-gui the jmeter can send the request much >>> faster to the jmeter-servers. >> >> In client-server mode, JMeter sends the test plan to the servers which then >> execute the test independently, and send back the results. >> >> If the JMeter server system is seriously overloaded, then the sample times >> can be affected, but this is unlikely to be the case here. >> Non-GUI mode uses far fewer resources. >> >> The client JMeter is only responsible for reading and processing the sample >> responses, and cannot directly affect the sample timings of the servers. >> >>> Or we have another problem in our testcase, environment or another problem. >> >> Yes, that seems the most likely. >> >>> best regards >>> >>> Lars Olofsson >>> >> >> --------------------------------------------------------------------- >> 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] >> >> > > --------------------------------------------------------------------- > 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] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
