DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG 
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://nagoya.apache.org/bugzilla/show_bug.cgi?id=11523>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND 
INSERTED IN THE BUG DATABASE.

http://nagoya.apache.org/bugzilla/show_bug.cgi?id=11523

Enhancement: constant-throughput timer





------- Additional Comments From [EMAIL PROTECTED]  2002-11-30 23:47 -------
Hi.

I'm attaching the code for a new Timer I've created which allows you to put the
server under pre-specified load.

I've found this useful when the performance requirement to test against is in
the lines of "the system should be able to serve X pages per minute with
latencies lower than Y milliseconds." (Which is probably one of the best ways to
express performance requirements).

To use it:

- Drop the two attached files in
src/components/org/apache/jmeter/timers/ConstantThroughputTimer.java
src/components/org/apache/jmeter/timers/gui/ConstantThroughputTimerGui.java

- Add the following resources to
./src/core/org/apache/jmeter/resources/messages.properties (and the
corresponding translations, of course):
constant_throughput_timer_memo=Add a delay between sampling to attain constant
throughput
constant_throughput_timer_title=Constant Throughput Timer
constant_throughput_timer_throughput=Thread Throughput (in samples per minute):

- Create your test, adding a Constant Throughput Timer to the Thread Group.
(I've not tested the behaviour of this timer when placed at other levels, nor
when combined with other timers).

- Run your test. Sampling will happen at the specified rate -- or as fast as
possible if the system is not fast enough to attain that rate.

Note the rate is for the ensemble of all threads: adding more threads will cause
each of them to run slower.

One difficulty I've encountered when writing this is that there's no method to
signal a test component when the test is starting. I've assumed that a new timer
object is created (or cloned) for every test run. The timer seems to behave as
expected, so looks like this assumption is correct...

Salut,

Jordi.

--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to