[Octave-bug-tracker] [bug #58445] significant slow down in stable version for cellfun invocations which use function handles

[Rik]

Follow-up Comment #14, bug#58445 (group octave):

Somehow we left this regression alone for 3 years, but the slowdown is still
there.

I wrote another small test before re-discovering this bug report.  The code is
attached.



N = 1e5;

C = {magic(2), magic(3), magic(4), magic(5)};

tic;
for i = 1:N
  n = cellfun (@numel, C);
end
printf ('Time for cellfun (@numel, C) : %f\n', toc);

tic;
for i = 1:N
  n = cellfun ("numel", C);
end
printf ('Time for cellfun ("numel", C) : %f\n', toc);


Results are


Time for cellfun (@numel, C) : 0.705530
Time for cellfun ("numel", C) : 0.202134


Still a significant slowdown for "accelerated" function names.


(file #55632)

    _______________________________________________________

Additional Item Attachment:

File name: bm_cellfun_handle.m            Size:0 KB
    <https://file.savannah.gnu.org/file/bm_cellfun_handle.m?file_id=55632>


    AGPL NOTICE

These attachments are served by Savane. You can download the corresponding
source code of Savane at
https://git.savannah.nongnu.org/cgit/administration/savane.git/snapshot/savane-fab6a498be17545b049f2ffbe43e58c5a946615e.tar.gz


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?58445>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/