https://bugs.freedesktop.org/show_bug.cgi?id=79761

Michael Meeks <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |NEEDINFO

--- Comment #11 from Michael Meeks <[email protected]> ---
> moreover I'm curious to know - if it's possible - why these performance
> sometimes regresses during master developement...

At a low level - it's hard to say why this specific regression happened until
we look into it. From a high level, judging the performance impact of any given
change on the rest of the 8 million lines of code is ~impossible. Speeding one
thing up, can make another thing slower quite easily; adding an unrelated
feature can cause a significant performance regression etc.

To try to catch this we have automated performance regression tests which
generate callgrind data. These don't capture auto-correct performance though.

Marina ... thanks for running the profile; I'm afraid my instructions were
wrong ... [ sorry ! ].

The callgrind results will be saved in a callgrind.11541.out file - can you
gzip that and attach it ? - hopefully its still there. Since that has the real
data to fix this with !.

-- 
You are receiving this mail because:
You are the assignee for the bug.
_______________________________________________
Libreoffice-bugs mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs

Reply via email to