Feedback on whether filter speed is newsworthy

I realized today that a changeset I had contributed last year (octave: d2cd9ead4c84) in response to this report (GNU Octave - Bugs: bug #61674, deconv much slower than conv [Savannah]) went into the Octave 8 branch and not Octave 7. It was months before Octave 7 was released but soon after the Octave 7 release candidate was made, and I forgot all about it soon afterwards.

Anyway, I noticed the speed difference today with fresh eyes when running some old code, and would like your thoughts on whether it is worth adding to NEWS for Octave 8. If yes, then phrasing it is tricky and I would like your feedback on these (or any other) ways, if warranted:

Version 1:

- `filter` is now 5X-6X faster, which also speeds up `deconv`, `fftfilt`
+and `arma_rnd` by the same amount.

Version 2:

- The functions `deconv`, `filter`, `fftfilt`, and `arma_rnd` are
all 5X-6X faster.

If it’s not newsworthy, then no change required.

Tagging @dasergatskov and @mmuetzel as having been involved in that old bug report discussion.

1 Like

Great effort @arungiridhar should be mentioned! I prefer version 1 :innocent:

1 Like

Thanks. I made it β€œ5X” and pushed it here: octave: 6203e303c5ac

2 Likes