You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@jmeter.apache.org by kbutler <ke...@gmail.com> on 2010/08/20 21:56:04 UTC
Re: Jmeter providing bad results
Is it a real difference or a reported difference? i.e. do 1000 transactions
actually take as long under both test tools but are reported as being slower
under JMeter? or JMeter really takes twice as long to execute the tests?
--
View this message in context: http://jmeter.512774.n5.nabble.com/Jmeter-providing-bad-results-tp2640077p2642744.html
Sent from the JMeter - User mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
Re: Jmeter providing bad results
Posted by Deepak Goel <de...@gmail.com>.
Hey
Namaskara~Nalama~Guten Tag
Can you please post the entire inputs, results (number of concurrent users
or threads, think time, transactions/sec output, response time) for both the
products: Jmeter and Seige?
Thanks
Deepak
--
Keigu
Deepak
+91-9765089593
deicool@gmail.com
http://www.simtree.net
Skype: thumsupdeicool
Google talk: deicool
Blog: http://loveandfearless.wordpress.com
Facebook: http://www.facebook.com/deicool
"Contribute to the world, environment and more : http://www.gridrepublic.org
"
On Sat, Aug 21, 2010 at 1:52 AM, jesmith17 <Jo...@adknowledge.com>wrote:
>
> The tests seems to take 2x as long to execute under JMeter. When comparing
> successfull transactions against time, the tx/sec numbers are correct, but
> the reported output is approx half what other load tests report for the same
> process run from the same location.
>
> I wonder if the difference is in how java opens the connections or manages
> the threads.
>
>
>
> On Fri, 2010-08-20 at 12:56 -0700, kbutler [via JMeter] wrote:
> Is it a real difference or a reported difference? i.e. do 1000
> transactions actually take as long under both test tools but are reported as
> being slower under JMeter? or JMeter really takes twice as long to execute
> the tests?
>
> ________________________________
> View message @
> http://jmeter.512774.n5.nabble.com/Jmeter-providing-bad-results-tp2640077p2642744.html
> To unsubscribe from Jmeter providing bad results, click here<
> http://jmeter.512774.n5.nabble.com/template/NodeServlet.jtp?tpl=unsubscribe_by_code&node=2640077&code=Sm9zaFNtaXRoQGFka25vd2xlZGdlLmNvbXwyNjQwMDc3fC0zNTQ1OTQ3MTU=
> >.
>
>
>
> --
> View this message in context:
> http://jmeter.512774.n5.nabble.com/Jmeter-providing-bad-results-tp2640077p2642780.html
> Sent from the JMeter - User mailing list archive at Nabble.com.
>
Re: Jmeter providing bad results
Posted by kblearner <sa...@citi.com>.
I haven't used Seige, but going by what I've read about it:
Seige's a GNU/Linux based tool, wheareas JMeter is a pure Java based
desktop application. There's bound to be some difference between these two
engines. Refer the link below. It provides some good insight into Seige Vs
JMeter and when they might be used.
http://www.linux.com/archive/feature/143896 Linux Forum
This is what I do:
Step 1 : Measure the response times manually using a stopwatch (literally)
for transactions
Step 2: Take up a tool, record the same transactions, then replay and
check out the response times.
If the comparison proves to be much closer, I take that the tool is
reporting just fine. If the response times given by the tool are little
higher, I'm OK with it. Frankly, I don't mind reporting little higher
response times than actual to the app/tech team, since it would ony lead to
little more tuning, which is good.
I've tried comparing the response times for both JMeter and LoadRunner and
they seem pretty close too.
Cheers.
--
View this message in context: http://jmeter.512774.n5.nabble.com/Jmeter-providing-bad-results-tp2640077p2798046.html
Sent from the JMeter - User mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
Re: Jmeter providing bad results
Posted by Deepak Shetty <sh...@gmail.com>.
what is the correct value though? While the load test runs , access the page
in a browser from a different machine and check the response time (using
firebug or YSlow or PageSpeed) - Check only the time it takes to return the
page (not the CSS / javascript etc - assuming that you dont have fetch
embedded requests).
Also do you your tests have data assertions? i.e. you are being able to
detect errors reliably?
On Fri, Aug 20, 2010 at 1:22 PM, jesmith17 <Jo...@adknowledge.com>wrote:
>
> The tests seems to take 2x as long to execute under JMeter. When comparing
> successfull transactions against time, the tx/sec numbers are correct, but
> the reported output is approx half what other load tests report for the same
> process run from the same location.
>
> I wonder if the difference is in how java opens the connections or manages
> the threads.
>
>
>
> On Fri, 2010-08-20 at 12:56 -0700, kbutler [via JMeter] wrote:
> Is it a real difference or a reported difference? i.e. do 1000
> transactions actually take as long under both test tools but are reported as
> being slower under JMeter? or JMeter really takes twice as long to execute
> the tests?
>
> ________________________________
> View message @
> http://jmeter.512774.n5.nabble.com/Jmeter-providing-bad-results-tp2640077p2642744.html
> To unsubscribe from Jmeter providing bad results, click here<
> http://jmeter.512774.n5.nabble.com/template/NodeServlet.jtp?tpl=unsubscribe_by_code&node=2640077&code=Sm9zaFNtaXRoQGFka25vd2xlZGdlLmNvbXwyNjQwMDc3fC0zNTQ1OTQ3MTU=
> >.
>
>
>
> --
> View this message in context:
> http://jmeter.512774.n5.nabble.com/Jmeter-providing-bad-results-tp2640077p2642780.html
> Sent from the JMeter - User mailing list archive at Nabble.com.
>
Re: Jmeter providing bad results
Posted by jesmith17 <Jo...@adknowledge.com>.
The tests seems to take 2x as long to execute under JMeter. When comparing successfull transactions against time, the tx/sec numbers are correct, but the reported output is approx half what other load tests report for the same process run from the same location.
I wonder if the difference is in how java opens the connections or manages the threads.
On Fri, 2010-08-20 at 12:56 -0700, kbutler [via JMeter] wrote:
Is it a real difference or a reported difference? i.e. do 1000 transactions actually take as long under both test tools but are reported as being slower under JMeter? or JMeter really takes twice as long to execute the tests?
________________________________
View message @ http://jmeter.512774.n5.nabble.com/Jmeter-providing-bad-results-tp2640077p2642744.html
To unsubscribe from Jmeter providing bad results, click here<http://jmeter.512774.n5.nabble.com/template/NodeServlet.jtp?tpl=unsubscribe_by_code&node=2640077&code=Sm9zaFNtaXRoQGFka25vd2xlZGdlLmNvbXwyNjQwMDc3fC0zNTQ1OTQ3MTU=>.
--
View this message in context: http://jmeter.512774.n5.nabble.com/Jmeter-providing-bad-results-tp2640077p2642780.html
Sent from the JMeter - User mailing list archive at Nabble.com.