[Numpy-discussion] fast putmask implementation
efiring at hawaii.edu
Thu Aug 16 22:39:02 EDT 2007
David Cournapeau wrote:
> On 8/17/07, Eric Firing <efiring at hawaii.edu> wrote:
>> In looking at maskedarray performance, I found that the filled()
>> function or method is a bottleneck. I think it can be sped up by using
>> putmask instead of indexed assignment, but I found that putmask itself
>> is slower than it needs to be. So I followed David Cournapeau's example
>> of fastclip and made a similar fastputmask. The diff relative to
>> current svn (3967) is attached.
> Great ! putmask was actually the function I wanted to improve after
> clip, because it is the second bottleneck for matplotlib imagesc :) I
> would not be suprised if now imagesc has descent speed compared to
>> I hope someone will take a look and either tell me what I need to fix or
>> commit it as-is.
> It looks like there are a lot of spurious diff in you patch (space vs
> tab, or endline problems ?). Could you regenerate a patch without
> them, since half of the patch is "garbage" ? It would be much easier
> to see the changes you actually made.
Agreed. This is because my editor deletes spurious whitespace that was
already in the file. If I ruled the world, the spurious whitespace and
hard tabs would never be there in the first place. (If I were younger I
might use smileys in places like this, but they just don't come
naturally to me.) As far as I can see there is no way of using svn diff
to deal with this automatically, so in the attached revision I have
manually removed chunks resulting solely from whitespace. Some of the
remaining chunks unfortunately have a mixture of whitespace and
substantive differences. And manually removing chunks is risky.
Is there a better way to handle this problem? A better way to make
diffs? Or any possibility of routinely cleaning the junk out of the svn
source files? (Yes, I know--what is junk to me probably results from
what others consider good behavior of the editor.)
> Numpy-discussion mailing list
> Numpy-discussion at scipy.org
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 8716 bytes
Desc: not available
More information about the NumPy-Discussion