loop controller and loop count differences

["Doroz, Tomasz" <To...@axadirect.pl>]

I have a simple scenario
- log in
- perform search1
- perform search2
- perform serach3

Now, is there a difference between those two approaches
1. 10 threads, 1 loop, login + loop controller (3 loops) for search ( cookies cleared each iteration )
2. 10 threads, 3 loops, once only controller for login + search  ( cookies not cleared each iteration )

I believe, in theory, the behavior should be the same, ie. 10 users logging in and running three searches, right?

So I ran both version and it seems that the second approach is much slower than the first one. The number of requests is the same in both cases, but the response times are much worse when the thread group loop count is higher than 1. The Active Threads over Time distribution is pretty much the same.

Any idea why the second version runs slower? Is there some mistake in my reasoning that both approaches should yield the same results?

td

Pomysl o srodowisku zanim wydrukujesz te wiadomosc.

AVANSSUR S.A. Oddzial w Polsce, ul. Chlodna 51, 00-867 Warszawa
tel. +48 22 599 90 00, fax +48 22 599 90 01, www.axadirect.pl
NIP 107-000-33-23, REGON 140408155
Wpisany do rejestru przedsiebiorcow prowadzonego przez Sad Rejonowy dla m.st. Warszawy, XII Wydzial Gospodarczy Krajowego Rejestru Sadowego, numer KRS 0000249486
AVANSSUR S.A. nalezy do Grupy AXA. Kapital zakladowy AVANSSUR S.A. wynosi 67 155 752,86 EUR

Re: loop controller and loop count differences

[sebb <se...@gmail.com>]

On 28 June 2012 15:09, Doroz, Tomasz <To...@axadirect.pl> wrote:
> I have a simple scenario
> - log in
> - perform search1
> - perform search2
> - perform serach3
>
> Now, is there a difference between those two approaches
> 1. 10 threads, 1 loop, login + loop controller (3 loops) for search ( cookies cleared each iteration )
> 2. 10 threads, 3 loops, once only controller for login + search  ( cookies not cleared each iteration )
>
> I believe, in theory, the behavior should be the same, ie. 10 users logging in and running three searches, right?
>
> So I ran both version and it seems that the second approach is much slower than the first one. The number of requests is the same in both cases, but the response times are much worse when the thread group loop count is higher than 1. The Active Threads over Time distribution is pretty much the same.
>
> Any idea why the second version runs slower? Is there some mistake in my reasoning that both approaches should yield the same results?

Is the data that is returned identical in both cases?
Are you sure that the requests are identical?

When you say that the second approach is slower, what are you measuring?
Individual samples, or some overall timing?

> td
>
> Pomysl o srodowisku zanim wydrukujesz te wiadomosc.
>
>
>
> AVANSSUR S.A. Oddzial w Polsce, ul. Chlodna 51, 00-867 Warszawa
>
> tel. +48 22 599 90 00, fax +48 22 599 90 01, www.axadirect.pl
>
> NIP 107-000-33-23, REGON 140408155
>
> Wpisany do rejestru przedsiebiorcow prowadzonego przez Sad Rejonowy dla m.st. Warszawy, XII Wydzial Gospodarczy Krajowego Rejestru Sadowego, numer KRS 0000249486
>
> AVANSSUR S.A. nalezy do Grupy AXA. Kapital zakladowy AVANSSUR S.A. wynosi 67 155 752,86 EUR
>
>

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