On Monday, August 20, 2018 at 8:44:48 AM UTC+10, Travis Scrimshaw wrote:
>
> Hi Julian,
>
> On Saturday, August 18, 2018 at 5:59:55 AM UTC+2, Travis Scrimshaw wrote:
>>>
>>> Thank you for working towards giving us extra tools to help us prevent 
>>> regressions. It is definitely something we can put to good use. However, 
>>> the total time it takes to test a file is somewhat of a bad metric
>>>
>> Where does it show the total time for a file for you? Currently, this is 
>> always for a block of doctest, i.e., typically all the tests defined by a 
>> single docstring. When a docstring changes the old and the new docstring 
>> are not going to show up as the same benchmark as the included hash value 
>> has then changed.
>>
>
> From your post above, you said, 'all "long time" doctests,' which means to 
> me all tests in the file. So if what you're saying is true, I have no idea 
> how to interpret this data then. When these docstrings change, how does it 
> keep track? What if something like just a variable name changes? How do we 
> really know it is comparing apples to apples?
>

Oh, I see. The name of the image tells you (approximately) which method's 
tests were being run. Okay, this is useful.

I can reproduce the slowdown (thankfully I happen to already have an 
8.3.beta3 version). Hmm...the code slowness in the Lie algebras is a 
symptom of an issue. It seems https://trac.sagemath.org/ticket/25504 caused 
the regression as the _get_action_() in matrix_space.py has a significant 
time increase (called via echelonize()).

Best,
Travis


-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Reply via email to