maven-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Kristian Rosenvold <kristian.rosenv...@gmail.com>
Subject Re: Wrong number of tests run when running JUnit tests in parallel: any help?
Date Mon, 12 Apr 2010 12:37:30 GMT
As the blogpost mentions, junit+surefire by itself is unable to 
constrain the number of threads; you need CPC=true for any kind of 
thread limitation.

Both threadCount and perCoreThreadCount are thread constraints, so you 
need CPC for both of them.

And btw; if you like running betas you'll get a significant speed boost 
by switching to jdk 7 ;)


Kristian


Den 12.04.2010 14:05, skrev Paolo Castagna:
> Kristian Rosenvold wrote:
>> All your questions should be answered here; 
>> http://incodewetrustinc.blogspot.com/2010/01/run-your-junit-tests-concurrently-with.html

>
>
> Thanks for the link, very useful.
>
> I have installed the configurable-parallel-computer as well.
> Now I see configurableParallelComputerPresent=true. :-)
>
> But, I am still confused by the threadCount option...
> Does it have any effect without the configurable-parallel-computer?
> Does it have any effect if used in conjunction with perCoreThreadCount?
>
>> I generally do not advise the use of "both", "classes" is both easier 
>> to get running and usually faster.
>
> I am now using "classes", I trust you.
>
> For the project I am currently testing the "parallel" execution of
> tests, I am probably in this scenario, as you wrote: "For a fairly
> optimized unit-test set, expect little or no gain - maybe 15-20%."
>
> But, I'll use the approach for other projects.
>
> Paolo
>
>>
>>
>> Kristian
>>
>>
>>
>>
>>
>> Den 12.04.2010 12:58, skrev Paolo Castagna:
>>> Thank you Kristian,
>>> my mistake. I was trying to use parallel with an old project which
>>> uses JUnit3 tests via JUnit4, so no @Test annotations.
>>>
>>> Now, I am trying again with another project which is using JUnit4
>>> with tests annotated with @Test.
>>>
>>> But I do not see any speed-up improvement (on a Quad Core).
>>>
>>> Also, I see:
>>>
>>> [INFO] Concurrency config is {threadCount=4, parallel=both, 
>>> configurableParallelComputerPresent=false}
>>>
>>> What the "configurableParallelComputerPresent=false" means?
>>>
>>> I am not even sure if threadCount=4 means 4 threads per core
>>> or 4 threads in total.
>>>
>>> But, at least, this time the number of tests executed is correct.
>>>
>>> Thanks again,
>>> Paolo
>>>
>>> Kristian Rosenvold wrote:
>>>> Please note that the parallel provider actually requires correctly 
>>>> defined junit4/junit3 tests to run, annotate your methods with @Test.
>>>>
>>>> The "classic" Junit 4 provider would run a large number of tests 
>>>> that were incorrectly defined according to junit specifications. 
>>>> The concurrent provider uses the same
>>>> selection mechanism as the junit 4 provider, but junit itself adds 
>>>> an additional compliance check before running the test.
>>>>
>>>> I also recommend you use the latest release of Junit; 4.8.2, or 
>>>> 4.8.1 if 4.8.2 hasn't reached the repos yet.
>>>>
>>>> Kristian
>>>> Den 11.04.2010 23:58, skrev Paolo Castagna:
>>>>> Hi,
>>>>> I am trying to see if running JUnit tests in parallel makes any
>>>>> difference and/or significant speed-up.
>>>>>
>>>>> I have this in my pom.xml:
>>>>>
>>>>> <dependency>
>>>>> <groupId>junit</groupId>
>>>>> <artifactId>junit</artifactId>
>>>>> <version>4.7</version>
>>>>> <scope>test</scope>
>>>>> </dependency>
>>>>>
>>>>> [...]
>>>>>
>>>>> <plugin>
>>>>> <groupId>org.apache.maven.plugins</groupId>
>>>>> <artifactId>maven-surefire-plugin</artifactId>
>>>>> <version>2.5</version>
>>>>> <configuration>
>>>>> <includes>
>>>>> <include>[...]</include>
>>>>> </includes>
>>>>> <parallel>methods</parallel>
>>>>> <threadCount>4</threadCount>
>>>>> </configuration>
>>>>> </plugin>
>>>>>
>>>>> When I run the tests without <parallel> and <threadCount>
I see:
>>>>>
>>>>> Tests run: 9491, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 
>>>>> ...
>>>>>
>>>>> While when I run the tests with <parallel> and <threadCount>
as 
>>>>> showed
>>>>> above:
>>>>>
>>>>> Tests run: 40, Failures: 0, Errors: 0, Skipped: 0
>>>>>
>>>>> Is this a bug?
>>>>>
>>>>> Am I doing something wrong here?
>>>>>
>>>>> Thanks,
>>>>> Paolo
>>>>>
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
>>>>> For additional commands, e-mail: users-help@maven.apache.org
>>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
>>>> For additional commands, e-mail: users-help@maven.apache.org
>>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
>>> For additional commands, e-mail: users-help@maven.apache.org
>>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
>> For additional commands, e-mail: users-help@maven.apache.org
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
> For additional commands, e-mail: users-help@maven.apache.org
>


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
For additional commands, e-mail: users-help@maven.apache.org


Mime
View raw message