Profiling data ...

[Michael Meeks]

Hi guys,

	Coming up to release time, I thought it'd be worth sharing Matus' nice
performance regression data as a pretty picture or two (attached).

	Having the data is really great; it let me (at least) see the horrible
docx import regression with fdo#38513 - discover it was down to
performance fragility not some obviously wrong thing we could revert,
and to fix it. It'd be great to have a review of that for -4-3:

	https://gerrit.libreoffice.org/#/c/9818/

	Beyond that it seems that we're in a reasonably good place for
performance in 4.3 (though this really measures master), the chart below
the top one shows just first and second start - it seems we transiently
gained a 5% first-start win, and then lost it again =) while 2nd start
got 5% slower over the measured time - mostly trickling away - perhaps
worth someone digging into the trace there for sillies.

	HTH,

		Michael.

-- 
 michael.meeks at collabora.com  <><, Pseudo Engineer, itinerant idiot
-------------- next part --------------
A non-text attachment was scrubbed...
Name: performance-2014-06-18.ods
Type: application/vnd.oasis.opendocument.spreadsheet
Size: 197426 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/libreoffice/attachments/20140618/93979b84/attachment-0001.ods>