[SciPy-dev] ticket #945?

Thouis (Ray) Jones thouis@broad.mit....
Thu Nov 26 14:55:01 CST 2009


The speed comparisons are here:
 http://mail.scipy.org/pipermail/scipy-dev/2009-May/012001.html

Most are about the same in speed.  histogram() is much slower, because
it tracks numpy.histogram instead of implementing its own.  Some could
be sped up, mostly those using _select in the modified
measurements.py.

histogram() is not backwards compatible (at least, it doesn't produce
exactly the same values).  Also, any code relying on non-lists being
returned when the functions are called with a list of labels (the
original motivation for this work) would break, but I expect that is a
limited number.

Ray


On Thu, Nov 26, 2009 at 14:57, Zachary Pincus <zachary.pincus@yale.edu> wrote:
> If it's as fast, and backward-compatible, this certainly sounds like a
> great idea. Especially if it opens up the way for more easily adding
> new measurements.
>
> Zach
>
>
>
> On Nov 26, 2009, at 2:39 PM, Thouis (Ray) Jones wrote:
>
>> I'm soliciting thoughts on http://projects.scipy.org/scipy/ticket/945
>>
>> We're using these changes in CellProfiler, and would be interested in
>> having them move to the trunk if they're useful.  Otherwise, we'll
>> need to make our own version so we can track the released version of
>> scipy (at some point).
>>
>> Ray Jones
>> _______________________________________________
>> SciPy-Dev mailing list
>> SciPy-Dev@scipy.org
>> http://mail.scipy.org/mailman/listinfo/scipy-dev
>
> _______________________________________________
> SciPy-Dev mailing list
> SciPy-Dev@scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-dev
>


More information about the SciPy-Dev mailing list