You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@jmeter.apache.org by Bruno Casarotti <br...@gmail.com> on 2013/04/09 00:01:21 UTC

How does the Response Time Graph and Thread Group works

Hi guys!
I'm new to the list, so I'm sorry if I made some mistake when sending this
email, but let me know to don't do it again ;)

So, I want to know how does it graph works?

Let's imagine that I have a single Thread Group with 4000 users, and my
Ramp-Up is 0 and also my loop count is setted for forever.

I also have 10 HTTP Requests to this Thread Group.

Question 1) When I run the test, will I run 4000 request per HTTP Request
at the same time? Am I running 40.000 threads at the same time in this
scenario?

Question 2)My Time Graph is showing to me the avarage response time of 4000
requests for each HTTP Request?

Question 3) I need to run 4000 requests for an HTTP link, and after that
more 4000 request for another one and so on, there is some way to do it at
once? I mean, 4000 per time.

Re: How does the Response Time Graph and Thread Group works

Posted by Deepak Shetty <sh...@gmail.com>.
The question is can your CPU(s) really run 4000 threads in parallel and
keep fidelity close to as if it were 4000 independent cpus running? In most
cases the answer is no , you dont get the same degree of parallelism (but
it is machine and test script dependent - scripts that have larger think
times and where the samples take a longer time on the server will work
better)




On Thu, Apr 11, 2013 at 12:39 PM, Bruno Casarotti <
brunoo.casarotti@gmail.com> wrote:

> I could run that, but because I have a "good" server here, but the results
> also isn't so real, I mean, the server will have a big I/O, which will
> reflect directly in the threads response time, anyway, if you have enough
> memory and a good system of SSD, you can run it without problems ;)
>
>
> 2013/4/11 Deepak Shetty <sh...@gmail.com>
>
> > Hi
> > that has been my experience too - however other users on this list have
> > reported running more threads successfully (i n the sense the results
> were
> > a  reasonable simulation)
> >
> > regards
> > deepak
> >
> >
> > On Thu, Apr 11, 2013 at 10:18 AM, Sergio Boso <sergio@bosoconsulting.it
> > >wrote:
> >
> > > Il 09/04/2013 18:35, Bruno Casarotti ha scritto:
> > >
> > >  Now I got it =) Thanks!
> > >>
> > >>
> > >> 2013/4/9 Deepak Shetty <sh...@gmail.com>
> > >>
> > >>  Hi
> > >>> no. Each thread (user) will request Sample1, Sample2 , Sample3
> > ,...Sample
> > >>> 30.
> > >>> All 4000 threads will act in parallel .
> > >>>
> > >>> However the sample may have different response time for different
> > threads
> > >>> (because thats the way server applications are) - all threads wont
> hit
> > >>> the
> > >>> same sample request at the same time
> > >>>
> > >>> regards
> > >>> deepak
> > >>>
> > >>>
> > >>>  I can add that, in my experience, 4000 threads can hardly be
> supported
> > > by a single server.
> > > Let alone the idea of a 0 milliseconds ramp up..
> > >
> > > Have a look at :
> > > http://www.bosoconsulting.it/**1/upload/flyer_wload_v1.0_en.**pdf<
> > http://www.bosoconsulting.it/1/upload/flyer_wload_v1.0_en.pdf>
> > >
> > > it may give you some helpful info.
> > >
> > >
> > > regards
> > >
> > > Sergio
> > > --
> > >
> > > Ing. Sergio Boso
> > >
> > >
> > >
> > >
> > >
> ------------------------------**------------------------------**---------
> > > To unsubscribe, e-mail: user-unsubscribe@jmeter.**apache.org<
> > user-unsubscribe@jmeter.apache.org>
> > > For additional commands, e-mail: user-help@jmeter.apache.org
> > >
> > >
> >
>

Re: How does the Response Time Graph and Thread Group works

Posted by Bruno Casarotti <br...@gmail.com>.
I could run that, but because I have a "good" server here, but the results
also isn't so real, I mean, the server will have a big I/O, which will
reflect directly in the threads response time, anyway, if you have enough
memory and a good system of SSD, you can run it without problems ;)


2013/4/11 Deepak Shetty <sh...@gmail.com>

> Hi
> that has been my experience too - however other users on this list have
> reported running more threads successfully (i n the sense the results were
> a  reasonable simulation)
>
> regards
> deepak
>
>
> On Thu, Apr 11, 2013 at 10:18 AM, Sergio Boso <sergio@bosoconsulting.it
> >wrote:
>
> > Il 09/04/2013 18:35, Bruno Casarotti ha scritto:
> >
> >  Now I got it =) Thanks!
> >>
> >>
> >> 2013/4/9 Deepak Shetty <sh...@gmail.com>
> >>
> >>  Hi
> >>> no. Each thread (user) will request Sample1, Sample2 , Sample3
> ,...Sample
> >>> 30.
> >>> All 4000 threads will act in parallel .
> >>>
> >>> However the sample may have different response time for different
> threads
> >>> (because thats the way server applications are) - all threads wont hit
> >>> the
> >>> same sample request at the same time
> >>>
> >>> regards
> >>> deepak
> >>>
> >>>
> >>>  I can add that, in my experience, 4000 threads can hardly be supported
> > by a single server.
> > Let alone the idea of a 0 milliseconds ramp up..
> >
> > Have a look at :
> > http://www.bosoconsulting.it/**1/upload/flyer_wload_v1.0_en.**pdf<
> http://www.bosoconsulting.it/1/upload/flyer_wload_v1.0_en.pdf>
> >
> > it may give you some helpful info.
> >
> >
> > regards
> >
> > Sergio
> > --
> >
> > Ing. Sergio Boso
> >
> >
> >
> >
> > ------------------------------**------------------------------**---------
> > To unsubscribe, e-mail: user-unsubscribe@jmeter.**apache.org<
> user-unsubscribe@jmeter.apache.org>
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
>

Re: How does the Response Time Graph and Thread Group works

Posted by Deepak Shetty <sh...@gmail.com>.
Hi
that has been my experience too - however other users on this list have
reported running more threads successfully (i n the sense the results were
a  reasonable simulation)

regards
deepak


On Thu, Apr 11, 2013 at 10:18 AM, Sergio Boso <se...@bosoconsulting.it>wrote:

> Il 09/04/2013 18:35, Bruno Casarotti ha scritto:
>
>  Now I got it =) Thanks!
>>
>>
>> 2013/4/9 Deepak Shetty <sh...@gmail.com>
>>
>>  Hi
>>> no. Each thread (user) will request Sample1, Sample2 , Sample3 ,...Sample
>>> 30.
>>> All 4000 threads will act in parallel .
>>>
>>> However the sample may have different response time for different threads
>>> (because thats the way server applications are) - all threads wont hit
>>> the
>>> same sample request at the same time
>>>
>>> regards
>>> deepak
>>>
>>>
>>>  I can add that, in my experience, 4000 threads can hardly be supported
> by a single server.
> Let alone the idea of a 0 milliseconds ramp up..
>
> Have a look at :
> http://www.bosoconsulting.it/**1/upload/flyer_wload_v1.0_en.**pdf<http://www.bosoconsulting.it/1/upload/flyer_wload_v1.0_en.pdf>
>
> it may give you some helpful info.
>
>
> regards
>
> Sergio
> --
>
> Ing. Sergio Boso
>
>
>
>
> ------------------------------**------------------------------**---------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.**apache.org<us...@jmeter.apache.org>
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>

Re: How does the Response Time Graph and Thread Group works

Posted by Sergio Boso <se...@bosoconsulting.it>.
Il 09/04/2013 18:35, Bruno Casarotti ha scritto:
> Now I got it =) Thanks!
>
>
> 2013/4/9 Deepak Shetty <sh...@gmail.com>
>
>> Hi
>> no. Each thread (user) will request Sample1, Sample2 , Sample3 ,...Sample
>> 30.
>> All 4000 threads will act in parallel .
>>
>> However the sample may have different response time for different threads
>> (because thats the way server applications are) - all threads wont hit the
>> same sample request at the same time
>>
>> regards
>> deepak
>>
>>
I can add that, in my experience, 4000 threads can hardly be supported by a single server.
Let alone the idea of a 0 milliseconds ramp up..

Have a look at :
http://www.bosoconsulting.it/1/upload/flyer_wload_v1.0_en.pdf

it may give you some helpful info.


regards

Sergio
-- 

Ing. Sergio Boso




---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org


Re: How does the Response Time Graph and Thread Group works

Posted by Bruno Casarotti <br...@gmail.com>.
Now I got it =) Thanks!


2013/4/9 Deepak Shetty <sh...@gmail.com>

> Hi
> no. Each thread (user) will request Sample1, Sample2 , Sample3 ,...Sample
> 30.
> All 4000 threads will act in parallel .
>
> However the sample may have different response time for different threads
> (because thats the way server applications are) - all threads wont hit the
> same sample request at the same time
>
> regards
> deepak
>
>
> On Tue, Apr 9, 2013 at 9:23 AM, Bruno Casarotti
> <br...@gmail.com>wrote:
>
> > Thanks for the reply, but I didn't understand yet.
> >
> > If I have a Thread Group, with 4000 users, and this thread group have
> like
> > 30 sampler http request, does it mean that the 4000 users will be divided
> > by the number of samplers?
> >
> >
> > 2013/4/8 Deepak Shetty <sh...@gmail.com>
> >
> > > 1) Threads (in your case 4000) run in parallel . Within a thread
> > everything
> > > (the 10 requests) run sequentially. So thread 1 to 4000 will all start
> at
> > > the same time (no ramp up) and  make HTTPrequest1 . Lets say Thread1
> > > returns the response then it will move on to the next request and each
> > > thread will do the same. however since your server response time will
> > vary
> > > , different threads might be on different requests or processing
> > responses.
> > >
> > > 2) " If many samples exist for the same timestamp, the mean value is
> > > displayed." --
> > >
> > >
> >
> http://jmeter.apache.org/usermanual/component_reference.html#Response_Time_Graph
> > > It depends on how many threads are executing the request at the same
> > time -
> > > its unlikely there are 4000
> > >
> > > 3) Use a synchronizing timer  --
> > >
> > >
> >
> http://jmeter.apache.org/usermanual/component_reference.html#Synchronizing_Timer
> > >
> > >
> > >
> > > On Mon, Apr 8, 2013 at 3:01 PM, Bruno Casarotti
> > > <br...@gmail.com>wrote:
> > >
> > > > Hi guys!
> > > > I'm new to the list, so I'm sorry if I made some mistake when sending
> > > this
> > > > email, but let me know to don't do it again ;)
> > > >
> > > > So, I want to know how does it graph works?
> > > >
> > > > Let's imagine that I have a single Thread Group with 4000 users, and
> my
> > > > Ramp-Up is 0 and also my loop count is setted for forever.
> > > >
> > > > I also have 10 HTTP Requests to this Thread Group.
> > > >
> > > > Question 1) When I run the test, will I run 4000 request per HTTP
> > Request
> > > > at the same time? Am I running 40.000 threads at the same time in
> this
> > > > scenario?
> > > >
> > > > Question 2)My Time Graph is showing to me the avarage response time
> of
> > > 4000
> > > > requests for each HTTP Request?
> > > >
> > > > Question 3) I need to run 4000 requests for an HTTP link, and after
> > that
> > > > more 4000 request for another one and so on, there is some way to do
> it
> > > at
> > > > once? I mean, 4000 per time.
> > > >
> > >
> >
>

Re: How does the Response Time Graph and Thread Group works

Posted by Deepak Shetty <sh...@gmail.com>.
Hi
no. Each thread (user) will request Sample1, Sample2 , Sample3 ,...Sample
30.
All 4000 threads will act in parallel .

However the sample may have different response time for different threads
(because thats the way server applications are) - all threads wont hit the
same sample request at the same time

regards
deepak


On Tue, Apr 9, 2013 at 9:23 AM, Bruno Casarotti
<br...@gmail.com>wrote:

> Thanks for the reply, but I didn't understand yet.
>
> If I have a Thread Group, with 4000 users, and this thread group have like
> 30 sampler http request, does it mean that the 4000 users will be divided
> by the number of samplers?
>
>
> 2013/4/8 Deepak Shetty <sh...@gmail.com>
>
> > 1) Threads (in your case 4000) run in parallel . Within a thread
> everything
> > (the 10 requests) run sequentially. So thread 1 to 4000 will all start at
> > the same time (no ramp up) and  make HTTPrequest1 . Lets say Thread1
> > returns the response then it will move on to the next request and each
> > thread will do the same. however since your server response time will
> vary
> > , different threads might be on different requests or processing
> responses.
> >
> > 2) " If many samples exist for the same timestamp, the mean value is
> > displayed." --
> >
> >
> http://jmeter.apache.org/usermanual/component_reference.html#Response_Time_Graph
> > It depends on how many threads are executing the request at the same
> time -
> > its unlikely there are 4000
> >
> > 3) Use a synchronizing timer  --
> >
> >
> http://jmeter.apache.org/usermanual/component_reference.html#Synchronizing_Timer
> >
> >
> >
> > On Mon, Apr 8, 2013 at 3:01 PM, Bruno Casarotti
> > <br...@gmail.com>wrote:
> >
> > > Hi guys!
> > > I'm new to the list, so I'm sorry if I made some mistake when sending
> > this
> > > email, but let me know to don't do it again ;)
> > >
> > > So, I want to know how does it graph works?
> > >
> > > Let's imagine that I have a single Thread Group with 4000 users, and my
> > > Ramp-Up is 0 and also my loop count is setted for forever.
> > >
> > > I also have 10 HTTP Requests to this Thread Group.
> > >
> > > Question 1) When I run the test, will I run 4000 request per HTTP
> Request
> > > at the same time? Am I running 40.000 threads at the same time in this
> > > scenario?
> > >
> > > Question 2)My Time Graph is showing to me the avarage response time of
> > 4000
> > > requests for each HTTP Request?
> > >
> > > Question 3) I need to run 4000 requests for an HTTP link, and after
> that
> > > more 4000 request for another one and so on, there is some way to do it
> > at
> > > once? I mean, 4000 per time.
> > >
> >
>

Re: How does the Response Time Graph and Thread Group works

Posted by Bruno Casarotti <br...@gmail.com>.
Thanks for the reply, but I didn't understand yet.

If I have a Thread Group, with 4000 users, and this thread group have like
30 sampler http request, does it mean that the 4000 users will be divided
by the number of samplers?


2013/4/8 Deepak Shetty <sh...@gmail.com>

> 1) Threads (in your case 4000) run in parallel . Within a thread everything
> (the 10 requests) run sequentially. So thread 1 to 4000 will all start at
> the same time (no ramp up) and  make HTTPrequest1 . Lets say Thread1
> returns the response then it will move on to the next request and each
> thread will do the same. however since your server response time will vary
> , different threads might be on different requests or processing responses.
>
> 2) " If many samples exist for the same timestamp, the mean value is
> displayed." --
>
> http://jmeter.apache.org/usermanual/component_reference.html#Response_Time_Graph
> It depends on how many threads are executing the request at the same time -
> its unlikely there are 4000
>
> 3) Use a synchronizing timer  --
>
> http://jmeter.apache.org/usermanual/component_reference.html#Synchronizing_Timer
>
>
>
> On Mon, Apr 8, 2013 at 3:01 PM, Bruno Casarotti
> <br...@gmail.com>wrote:
>
> > Hi guys!
> > I'm new to the list, so I'm sorry if I made some mistake when sending
> this
> > email, but let me know to don't do it again ;)
> >
> > So, I want to know how does it graph works?
> >
> > Let's imagine that I have a single Thread Group with 4000 users, and my
> > Ramp-Up is 0 and also my loop count is setted for forever.
> >
> > I also have 10 HTTP Requests to this Thread Group.
> >
> > Question 1) When I run the test, will I run 4000 request per HTTP Request
> > at the same time? Am I running 40.000 threads at the same time in this
> > scenario?
> >
> > Question 2)My Time Graph is showing to me the avarage response time of
> 4000
> > requests for each HTTP Request?
> >
> > Question 3) I need to run 4000 requests for an HTTP link, and after that
> > more 4000 request for another one and so on, there is some way to do it
> at
> > once? I mean, 4000 per time.
> >
>

Re: How does the Response Time Graph and Thread Group works

Posted by Deepak Shetty <sh...@gmail.com>.
1) Threads (in your case 4000) run in parallel . Within a thread everything
(the 10 requests) run sequentially. So thread 1 to 4000 will all start at
the same time (no ramp up) and  make HTTPrequest1 . Lets say Thread1
returns the response then it will move on to the next request and each
thread will do the same. however since your server response time will vary
, different threads might be on different requests or processing responses.

2) " If many samples exist for the same timestamp, the mean value is
displayed." --
http://jmeter.apache.org/usermanual/component_reference.html#Response_Time_Graph
It depends on how many threads are executing the request at the same time -
its unlikely there are 4000

3) Use a synchronizing timer  --
http://jmeter.apache.org/usermanual/component_reference.html#Synchronizing_Timer



On Mon, Apr 8, 2013 at 3:01 PM, Bruno Casarotti
<br...@gmail.com>wrote:

> Hi guys!
> I'm new to the list, so I'm sorry if I made some mistake when sending this
> email, but let me know to don't do it again ;)
>
> So, I want to know how does it graph works?
>
> Let's imagine that I have a single Thread Group with 4000 users, and my
> Ramp-Up is 0 and also my loop count is setted for forever.
>
> I also have 10 HTTP Requests to this Thread Group.
>
> Question 1) When I run the test, will I run 4000 request per HTTP Request
> at the same time? Am I running 40.000 threads at the same time in this
> scenario?
>
> Question 2)My Time Graph is showing to me the avarage response time of 4000
> requests for each HTTP Request?
>
> Question 3) I need to run 4000 requests for an HTTP link, and after that
> more 4000 request for another one and so on, there is some way to do it at
> once? I mean, 4000 per time.
>