Re: [eigen] Testing 2.0

[ Thread Index | Date Index | More lists.tuxfamily.org/eigen Archives ]


Good!

Well, when I released it, we were already 4 hours into February 12 in
your timezone, so I'll use the timezones excuses for that early
release :) Seriously though, it was more convenient for me to do it a
bit early.

I'm all for the dashboard, but we have a problem, it's hard to limit
how much space it eats, so recently Gael had to nuke it again. That's
why we haven't been able to keep the results long enough. Also, the
dashboard stores a ton of information but all I need when making a
release is to know that all tests built without warnings and passed. I
wish we could store this minimal information for a long time while
frequently erasing the big stuff (coverage data, etc).

Benoit

2010/2/12 Hauke Heibel <hauke.heibel@xxxxxxxxxxxxxx>:
> You were too fast for me to test but my dashboard is green:
> http://eigen.tuxfamily.org/CDash/buildSummary.php?buildid=1514
>
> IMHO testing via the testsuite should be what we should aim at when
> making releases.
>
> - Hauke
>
> On Thu, Feb 11, 2010 at 9:43 PM, Benoit Jacob <jacob.benoit.1@xxxxxxxxx> wrote:
>> Everybody: we're releasing Eigen 2.0.12 tomorrow.
>>
>> I don't remember if GCC 3.3 has been tested (sorry Jitse) ? What about
>> MSVC 2005, 2008? And ICC?
>>
>> To test:
>>
>> cmake -DEIGEN_BUILD_TESTS=ON /path/to/eigen2.0
>> make
>>
>> I'm also happy to announce support for one more platform: QCC/QNX,
>> after bug #92.
>>
>> Benoit
>>
>> 2010/2/7 Benoit Jacob <jacob.benoit.1@xxxxxxxxx>:
>>> 2010/2/7 Jitse Niesen <jitse@xxxxxxxxxxxxxxxxx>:
>>>> On Sun, 7 Feb 2010, Benoit Jacob wrote:
>>>>
>>>>> ah yes, the lu tests fail typically on the exact rank computation, and
>>>>> the reason why it fails is that eigen 2.0's algorithm for that is
>>>>> really not very accurate.
>>>>>
>>>>> This has been fixed in the development branch: the exact rank
>>>>> computation now works very consistently, so you can run the lu tests
>>>>> with 1000 repetitions.
>>>>
>>>> Here the lu_2 test fails regularly in the development branch. I tried it
>>>> with random seeds from 1 to 10 (and the default of 10 repeats) and it fails
>>>> with 4, 5 and 6 in line 102. It's good to know that it's supposed to work
>>>> consistently, so now I can treat it as a bug and hunt it down when I have
>>>> time.
>>>
>>> Hm indeed it's a bug, then!
>>>
>>> But here, all lu tests work... so I can't reproduce. I tried
>>> EIGEN_SEED=1, too...
>>>
>>> Benoit
>>>
>>>
>>>
>>>>
>>>> Cheers,
>>>> Jitse
>>>>
>>>>
>>>>
>>>
>>
>>
>>
>
>
>



Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/